Initial commit - cleaned repository

This commit is contained in:
jungwoo choi
2025-09-28 20:41:57 +09:00
commit e3c28f796a
188 changed files with 28102 additions and 0 deletions

37
.env.example Normal file
View File

@ -0,0 +1,37 @@
# Environment Configuration Example
# Copy this file to .env and update with your values
ENV=development
# Port Configuration
CONSOLE_BACKEND_PORT=8011
CONSOLE_FRONTEND_PORT=3000
USERS_SERVICE_PORT=8001
MONGODB_PORT=27017
REDIS_PORT=6379
# Database Configuration
MONGODB_URL=mongodb://mongodb:27017
MONGODB_DATABASE=site11_db
USERS_DB_NAME=users_db
# Redis Configuration
REDIS_URL=redis://redis:6379
# JWT Configuration
JWT_SECRET_KEY=change-this-secret-key-in-production
JWT_ALGORITHM=HS256
ACCESS_TOKEN_EXPIRE_MINUTES=30
# Service URLs (Internal)
USERS_SERVICE_URL=http://users-backend:8000
# Frontend Configuration
VITE_API_URL=http://localhost:8011
# Kafka Configuration (Future)
# KAFKA_BOOTSTRAP_SERVERS=kafka:9092
# KAFKA_GROUP_ID=site11-group
# Docker Configuration
COMPOSE_PROJECT_NAME=site11

85
.gitignore vendored Normal file
View File

@ -0,0 +1,85 @@
# Python
__pycache__/
*.py[cod]
*$py.class
*.so
.Python
env/
venv/
.venv
pip-log.txt
pip-delete-this-directory.txt
.pytest_cache/
*.egg-info/
dist/
build/
# Node
node_modules/
npm-debug.log*
yarn-debug.log*
yarn-error.log*
.pnpm-debug.log*
dist/
build/
# Environment
.env
.env.local
.env.*.local
*.local
# IDE
.vscode/
.idea/
*.swp
*.swo
*~
.DS_Store
# Docker
*.log
docker-compose.override.yml
# Database
data/
*.db
*.sqlite
# Testing
coverage/
.coverage
htmlcov/
.tox/
.hypothesis/
# Temporary
tmp/
temp/
*.tmp
*.temp
*.bak
# Secrets
*.pem
*.key
*.crt
secrets/data/
# Backups and logs
backups/
*.archive
*.log
# Images and media
*.png
*.jpg
*.jpeg
*.gif
# Node modules
node_modules/
**/node_modules/
# Large data files
data/

275
CLAUDE.md Normal file
View File

@ -0,0 +1,275 @@
# CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
## Development Principles
**IMPORTANT**:
1. 모든 개발은 Docker 환경에서만 진행
2. Docker 빌드는 백그라운드로 실행하고 완료까지 대기
3. 로컬 환경 설정 금지 (venv, npm install 등)
## File Naming Convention
**IMPORTANT**: 모든 문서 파일은 대문자.md 형식으로 생성
- 예: README.md, CHANGELOG.md, TODO.md, ARCHITECTURE.md
## Context Recovery Guide
**IMPORTANT**: 새 세션 시작 시 반드시 확인할 파일들:
1. `docs/PROGRESS.md` - 현재 진행 상황과 다음 단계
2. `docs/PLAN.md` - 전체 구현 계획
3. `docker ps` - 실행 중인 서비스 확인
## Quick Status Check
```bash
# 진행 상황 빠른 확인
cat docs/PROGRESS.md | grep "Next Action"
```
## System Architecture
### Architecture Pattern: Microservices with Central Console
**Core Concept**: Console acts as the orchestrator and dashboard, while individual microservices handle domain-specific logic.
### Architecture Approaches
#### 1. API Gateway Pattern (Recommended)
Console serves as an API Gateway that:
- Routes requests to appropriate microservices
- Handles authentication/authorization centrally
- Aggregates responses from multiple services
- Provides service discovery and health monitoring
#### 2. Service Mesh Pattern
- Each service communicates directly via service mesh (Istio/Linkerd)
- Console focuses on monitoring and management
- Better for complex inter-service communication
#### 3. Event-Driven Architecture
- Services communicate via message broker (RabbitMQ/Kafka)
- Console subscribes to events for real-time monitoring
- Loose coupling between services
### Technology Stack
- **Backend**: FastAPI (Python 3.11) + Motor (MongoDB async driver)
- **Frontend**: React 18 + Vite + TypeScript + Material-UI v7
- **Database**: MongoDB 7.0
- **Cache**: Redis 7
- **Reverse Proxy**: Nginx
- **Container**: Docker & Docker Compose
- **Version Control**: git
- **API Documentation**: OpenAPI/Swagger
- **Service Communication**: REST + gRPC (for internal services)
### Service Configuration
- **Nginx**: Port 80 (reverse proxy)
- **Console Backend**: Port 8000 (API Gateway)
- **Console Frontend**: Port 3000
- **Services**:
- Images: 8001-8002
- OAuth: 8003-8004
- Applications: 8005-8006
- Users: 8007-8008
- Data: 8009-8010
- Statistics: 8011-8012
- **MongoDB**: Port 27017 (internal)
- **Redis**: Port 6379 (internal)
## Microservices Implementation Strategy
### Console as API Gateway (Recommended Approach)
#### Responsibilities
**Console Backend**:
- Service discovery and routing
- Authentication & authorization (JWT/OAuth2)
- Request/response transformation
- Rate limiting & throttling
- Circuit breaking for fault tolerance
- Centralized logging & monitoring
- API composition for complex operations
**Individual Microservices**:
- Domain-specific business logic
- Own database/collection management
- Event publishing for async operations
- Health endpoints for monitoring
- OpenAPI documentation
### Service Communication Patterns
#### 1. Synchronous Communication (REST)
```python
# Console backend routing example
@app.get("/api/users/{user_id}")
async def get_user(user_id: str):
# Route to users service
response = await http_client.get(f"http://users-service:8007/users/{user_id}")
return response.json()
```
#### 2. Asynchronous Communication (Event-driven)
```python
# Service publishes event
await redis_client.publish("user.created", user_data)
# Console subscribes to events
async def handle_user_created(data):
# Update dashboard metrics
await update_statistics(data)
```
#### 3. Service Registry Pattern
```yaml
# services-registry.yaml
services:
users:
backend: "http://users-backend:8007"
frontend: "http://users-frontend:8008"
health: "/health"
oauth:
backend: "http://oauth-backend:8003"
frontend: "http://oauth-frontend:8004"
health: "/health"
```
### Development Workflow
#### Commands
```bash
# Start all services
docker-compose up -d
# Start specific service
docker-compose up -d console users
# View logs
docker-compose logs -f [service-name]
# Rebuild service
docker-compose build [service-name]
docker-compose up -d [service-name]
# Run tests
docker-compose exec [service-name] pytest
# Database migrations
docker-compose exec [service-name] alembic upgrade head
```
#### Service Development Guidelines
1. Each service should be independently deployable
2. Services share nothing except API contracts
3. Use correlation IDs for distributed tracing
4. Implement health checks and readiness probes
5. Version APIs appropriately (e.g., /api/v1/)
## Project Structure
```
site11/
├── docker-compose.yml
├── nginx/
│ └── nginx.conf
├── console/
│ ├── backend/
│ │ ├── Dockerfile
│ │ └── requirements.txt
│ └── frontend/
│ ├── Dockerfile
│ ├── package.json
│ ├── vite.config.ts
│ ├── tsconfig.json
│ └── src/
│ ├── App.tsx
│ ├── main.tsx
│ ├── layouts/
│ │ └── AdminLayout.tsx
│ └── pages/
│ ├── Login.tsx
│ └── Register.tsx
├── services/
│ ├── images/
│ │ ├── backend/
│ │ └── frontend/
│ ├── oatuh/
│ │ ├── backend/
│ │ └── frontend/
│ ├── applications/
│ │ ├── backend/
│ │ └── frontend/
│ ├── users/
│ │ ├── backend/
│ │ └── frontend/
│ ├── data/
│ │ ├── backend/
│ │ └── frontend/
│ └── statistics/
│ ├── backend/
│ └── frontend/
├── docs/
│ └── PLAN.md
└── CLAUDE.md
```
## Service-Specific Architecture
### Console Service
- **Purpose**: Central orchestrator and dashboard
- **Key Features**:
- Service health monitoring dashboard
- Unified authentication portal
- API Gateway for all services
- Real-time metrics aggregation
- Service configuration management
### Microservice Template
Each service follows this structure:
```
service-name/
├── backend/
│ ├── Dockerfile
│ ├── requirements.txt
│ ├── app/
│ │ ├── main.py # FastAPI app
│ │ ├── models.py # Pydantic models
│ │ ├── routes.py # API endpoints
│ │ ├── database.py # MongoDB connection
│ │ └── services.py # Business logic
│ └── tests/
└── frontend/
├── Dockerfile
├── package.json
└── src/
├── api/ # Service API client
├── components/ # React components
└── hooks/ # Custom hooks
```
## Inter-Service Communication
### API Gateway Routes (Console)
```python
# Console backend routing configuration
SERVICES = {
"users": {"url": "http://users-backend:8007", "prefix": "/api/users"},
"oauth": {"url": "http://oauth-backend:8003", "prefix": "/api/auth"},
"images": {"url": "http://images-backend:8001", "prefix": "/api/images"},
"applications": {"url": "http://apps-backend:8005", "prefix": "/api/apps"},
"data": {"url": "http://data-backend:8009", "prefix": "/api/data"},
"statistics": {"url": "http://stats-backend:8011", "prefix": "/api/stats"}
}
```
### Service Discovery
Services register themselves with Console on startup and send periodic heartbeats.
## Database Strategy
- **Shared MongoDB instance** with separate databases per service
- **Redis** for caching and pub/sub messaging
- Each service owns its data and exposes it via APIs
## Security Considerations
- JWT tokens issued by OAuth service
- Console validates tokens and forwards to services
- Internal service communication uses service tokens
- Rate limiting at API Gateway level

185
PLAN.md Normal file
View File

@ -0,0 +1,185 @@
# 프로젝트 개발 계획
## 프로젝트 목표
마이크로서비스 아키텍처 기반의 확장 가능한 웹 애플리케이션 구축
## 아키텍처 원칙
1. **Console as API Gateway**: 모든 외부 요청은 Console을 통해 라우팅
2. **Docker-Only Development**: 모든 개발과 실행은 Docker 컨테이너 내에서
3. **Event-Driven Architecture**: Kafka를 통한 서비스 간 비동기 통신
4. **Service Isolation**: 각 서비스는 독립적으로 배포 가능
## 완료된 단계 (✅)
### Phase 1: 기반 구축
- [x] Step 1: 기본 프로젝트 구조 및 Docker 설정
- [x] Step 2: Users 마이크로서비스 구현
- [x] Step 3: MongoDB 통합
- [x] Step 4: Redis 캐싱 시스템
- [x] Step 5: Frontend 스켈레톤 (React + Vite)
- [x] Step 6: JWT 인증 시스템
- [x] Step 6.5: Images 서비스 통합
- [x] Step 7: Kafka 이벤트 시스템
- [x] Step 8: OAuth 2.0 인증 시스템 및 프로필 기능
## 진행 예정 단계
### Phase 2: 이벤트 기반 시스템 확장
#### Step 9: 고급 이벤트 처리
- [ ] 이벤트 소비자 구현
- Console에서 user-events 토픽 구독
- 알림 서비스 이벤트 처리
- [ ] Dead Letter Queue 구현
- [ ] 이벤트 재시도 메커니즘
- [ ] 이벤트 스키마 레지스트리
#### Step 10: 태스크 큐 시스템
- [ ] Kafka 기반 백그라운드 작업 처리
- [ ] 이미지 프로세싱 작업 큐
- [ ] 이메일 전송 큐
- [ ] 배치 작업 스케줄러
### Phase 3: 고급 기능
#### Step 11: 실시간 기능
- [ ] WebSocket 통합 (Console)
- [ ] 실시간 알림 시스템
- [ ] 온라인 사용자 상태 추적
- [ ] 실시간 데이터 동기화
#### Step 12: 파일 시스템
- [ ] 파일 업로드 서비스
- [ ] S3 호환 객체 스토리지 (MinIO)
- [ ] 파일 메타데이터 관리
- [ ] 썸네일 생성 서비스
#### Step 13: 검색 시스템
- [ ] Elasticsearch 통합
- [ ] 전문 검색 기능
- [ ] 자동완성 기능
- [ ] 검색 분석 및 최적화
### Phase 4: 프로덕션 준비
#### Step 14: 모니터링 및 로깅
- [ ] Prometheus 메트릭 수집
- [ ] Grafana 대시보드
- [ ] ELK Stack 로깅
- [ ] 분산 추적 (Jaeger)
#### Step 15: 보안 강화
- [ ] Rate Limiting
- [ ] API Key 관리
- [ ] OAuth2 통합
- [ ] 데이터 암호화
#### Step 16: 테스트 및 CI/CD
- [ ] 단위 테스트 작성
- [ ] 통합 테스트
- [ ] E2E 테스트
- [ ] GitHub Actions CI/CD
#### Step 17: 성능 최적화
- [ ] 데이터베이스 인덱싱
- [ ] 쿼리 최적화
- [ ] 캐싱 전략 개선
- [ ] CDN 통합
## 서비스 구성
### 현재 서비스
1. **Console** (API Gateway)
- Frontend: React SPA
- Backend: FastAPI, JWT 인증
2. **Users Service**
- User CRUD
- MongoDB 저장소
- Kafka 이벤트 발행
- 프로필 정보 관리 (사진, bio, location 등)
3. **Images Service**
- 이미지 프록시
- 캐싱 시스템
- WebP 변환
4. **OAuth Service**
- OAuth 2.0 인증 서버
- 애플리케이션 등록 및 관리
- 토큰 발급 및 검증
- SSO 지원 (Google, GitHub, SAML)
- 스코프 기반 권한 관리
### 계획된 서비스
5. **Notification Service**
- 이메일/SMS 전송
- 푸시 알림
- 알림 히스토리
6. **Analytics Service**
- 사용자 행동 분석
- 비즈니스 메트릭
- 리포트 생성
7. **Payment Service**
- 결제 처리
- 구독 관리
- 청구서 생성
## 기술 스택 로드맵
### 현재 사용 중
- FastAPI, React, TypeScript
- MongoDB, Redis
- Apache Kafka
- Docker, Docker Compose
### 도입 예정
- Elasticsearch (검색)
- MinIO (객체 스토리지)
- Prometheus/Grafana (모니터링)
- Jaeger (분산 추적)
- Nginx (리버스 프록시)
## 개발 일정
### 2025 Q1
- Phase 2 완료 (이벤트 시스템)
- Phase 3 시작 (고급 기능)
### 2025 Q2
- Phase 3 완료
- Phase 4 시작 (프로덕션 준비)
### 2025 Q3
- Phase 4 완료
- 프로덕션 배포
## 성공 지표
1. **기술적 지표**
- 서비스 응답 시간 < 200ms
- 시스템 가용성 > 99.9%
- 초당 처리 가능 요청 > 1000
2. **개발 지표**
- 테스트 커버리지 > 80%
- 빌드 시간 < 5분
- 배포 시간 < 10분
3. **확장성 지표**
- 수평 확장 가능
- 서비스 독립 배포
- 무중단 업데이트
## 리스크 및 대응 방안
1. **복잡도 증가**
- 대응: 점진적 구현, 문서화 강화
2. **성능 병목**
- 대응: 프로파일링, 캐싱 전략
3. **데이터 일관성**
- 대응: 이벤트 소싱, SAGA 패턴
4. **보안 취약점**
- 대응: 정기 보안 감사, 자동화된 스캔

147
PROGRESS.md Normal file
View File

@ -0,0 +1,147 @@
# 프로젝트 진행 상황
## 완료된 단계
### Step 1: 기본 프로젝트 구조 생성 ✅
- Docker Compose 설정
- Console 서비스 (API Gateway) 기본 구현
- 프로젝트 문서 (CLAUDE.md, PLAN.md) 작성
### Step 2: Users 마이크로서비스 구현 ✅
- Users 서비스 CRUD API
- MongoDB 연동 (Beanie ODM)
- 서비스 간 통신 설정
### Step 3: MongoDB 통합 ✅
- MongoDB 컨테이너 설정
- Beanie ODM 설정
- Users 모델 및 데이터베이스 연결
- PyMongo 버전 호환성 문제 해결
### Step 4: Redis 통합 ✅
- Redis 컨테이너 설정
- 캐싱 시스템 준비
- 향후 세션 관리 및 캐싱 구현 예정
### Step 5: Frontend 스켈레톤 ✅
- React + Vite + TypeScript 설정
- Material-UI 통합
- Console Frontend 기본 구조
- npm ci → npm install 문제 해결
### Step 6: 환경 변수 및 인증 ✅
- .env 파일 설정
- JWT 인증 시스템 구현
- Console이 인증 처리 담당
- 포트 충돌 해결 (8000 → 8011)
### Step 6.5: Images 서비스 통합 ✅
- site00의 image-service 마이그레이션
- 프록시 및 캐싱 기능 유지
- WebP 변환 기능 포함
- Console에서 Images 서비스로 라우팅
### Step 7: Kafka 이벤트 시스템 ✅
- Kafka 및 Zookeeper 컨테이너 추가
- 공유 Kafka 라이브러리 생성 (Producer/Consumer)
- 이벤트 타입 정의 (USER_CREATED, USER_UPDATED, USER_DELETED 등)
- Users 서비스에 이벤트 발행 기능 추가
- aiokafka 통합
### Step 8: OAuth 2.0 인증 시스템 및 프로필 기능 ✅
- OAuth 2.0 서비스 구현
- Authorization Code, Client Credentials, Refresh Token 플로우
- 애플리케이션 등록 및 관리 (CRUD)
- 토큰 introspection 및 revocation
- PKCE 지원
- SSO 설정 지원
- Google, GitHub, SAML 프로바이더 설정
- 도메인 기반 접근 제어
- 스코프 시스템 구현
- 실용적인 스코프 카테고리 (기본 인증, 사용자, 앱, 조직, API)
- picture 스코프 추가 (프로필 사진 접근)
- Users 서비스 프로필 기능 확장
- 프로필 사진 및 썸네일 필드
- bio, location, website 필드
- 이메일 인증 및 계정 활성화 상태
- Docker 통합 완료 (포트 8003)
## 현재 실행 중인 서비스
- **Console Frontend**: http://localhost:3000
- **Console Backend**: http://localhost:8011
- **Users Service**: http://localhost:8001
- **Images Service**: http://localhost:8002
- **OAuth Service**: http://localhost:8003
- **MongoDB**: localhost:27017
- **Redis**: localhost:6379
- **Kafka**: localhost:9092
- **Zookeeper**: localhost:2181
## 다음 단계 (예정)
### Step 9: 고급 이벤트 처리
- 이벤트 소비자 구현
- 이벤트 기반 워크플로우
- 에러 처리 및 재시도 로직
### Step 10: 태스크 큐 시스템
- Kafka 기반 백그라운드 작업 처리
- 이미지 프로세싱 작업 큐
- 이메일 전송 큐
### Step 11: 고급 기능
- 실시간 알림 (WebSocket)
- 파일 업로드 시스템
- 검색 기능 (Elasticsearch)
### Step 12: 프로덕션 준비
- 로깅 시스템 (ELK Stack)
- 모니터링 (Prometheus/Grafana)
- CI/CD 파이프라인
- 테스트 자동화
## 기술 스택
- **Backend**: FastAPI (Python)
- **Frontend**: React + TypeScript + Vite + Material-UI
- **Database**: MongoDB
- **Cache**: Redis
- **Message Queue**: Apache Kafka
- **Container**: Docker & Docker Compose
- **Authentication**: JWT
## 주요 환경 변수
```env
COMPOSE_PROJECT_NAME=site11
CONSOLE_BACKEND_PORT=8011
USERS_BACKEND_PORT=8001
OAUTH_BACKEND_PORT=8003
JWT_SECRET_KEY=your-secret-key-change-in-production-12345
KAFKA_BOOTSTRAP_SERVERS=kafka:9092
```
## 문제 해결 기록
1. **PyMongo 호환성**: motor와 pymongo 버전 충돌 → pymongo==4.6.1로 고정
2. **npm ci 실패**: package-lock.json 부재 → npm install로 변경
3. **포트 충돌**: 8000 포트 사용 중 → Console을 8011로 변경
4. **WebP 변환 문제**: 검정색 이미지 출력 → convert_to_webp 임시 비활성화
5. **httpx 미사용 import**: OAuth 서비스 초기 구현시 불필요한 import → 제거
6. **Kafka producer 파라미터 오류**: max_in_flight_requests_per_connection → 제거
7. **Users 서비스 포트 누락**: docker-compose.yml에 포트 매핑 추가
## 컨텍스트 복구 정보
Claude가 재시작되면 이 정보를 참조:
- 작업 디렉토리: `/Users/jungwoochoi/Desktop/prototype/site11`
- Git 저장소: 각 단계마다 커밋됨
- Docker 개발 원칙: 모든 개발은 Docker 내에서 진행
- 문서 형식: 모든 마크다운 파일은 대문자 (CLAUDE.md, PLAN.md, PROGRESS.md)
- Console 서비스가 중앙 API Gateway 역할
- Kafka를 메인 이벤트 시스템으로 사용
- Redis는 캐싱 전용
- OAuth 2.0 서비스로 인증/인가 관리
- 모든 서비스는 독립적으로 배포 가능

29
README.md Normal file
View File

@ -0,0 +1,29 @@
# Site11 - Microservices Architecture
## Overview
Microservices platform with Console as API Gateway orchestrating multiple domain services.
## Quick Start
### Start Services
```bash
# Start console service
docker-compose up -d console-backend
# Check status
curl http://localhost:8011/health
```
### Available Endpoints
- `http://localhost:8011/` - Root endpoint
- `http://localhost:8011/health` - Health check
- `http://localhost:8011/api/status` - System status
## Architecture
- **Console**: API Gateway and orchestrator
- **Services**: Domain-specific microservices (users, oauth, images, etc.)
- **Database**: MongoDB for persistence
- **Cache**: Redis for caching and pub/sub
## Development
See `docs/PLAN.md` for implementation roadmap and `docs/PROGRESS.md` for current status.

View File

@ -0,0 +1,13 @@
FROM python:3.11-slim
WORKDIR /app
# Install dependencies
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
# Copy application
COPY . .
# Run the application
CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "8000", "--reload"]

View File

@ -0,0 +1,218 @@
"""
Article Generation Module
Claude API를 사용한 기사 생성 로직
"""
from typing import Dict, Any, List, Optional
from datetime import datetime
import json
import uuid
import logging
from anthropic import AsyncAnthropic
from pydantic import BaseModel, Field
logger = logging.getLogger(__name__)
# Data Models
class NewsSource(BaseModel):
"""뉴스 소스 정보"""
title: str
url: str
published_date: Optional[str] = None
source_site: str = "Unknown"
class EventInfo(BaseModel):
"""이벤트 정보"""
name: str
date: Optional[str] = None
location: Optional[str] = None
class Entities(BaseModel):
"""추출된 엔티티"""
people: List[str] = Field(default_factory=list)
organizations: List[str] = Field(default_factory=list)
groups: List[str] = Field(default_factory=list)
countries: List[str] = Field(default_factory=list)
events: List[EventInfo] = Field(default_factory=list)
keywords: List[str] = Field(default_factory=list)
class SubTopic(BaseModel):
"""기사 소주제"""
title: str
content: List[str]
class GeneratedArticle(BaseModel):
"""생성된 기사"""
news_id: str = Field(default_factory=lambda: str(uuid.uuid4()))
title: str
summary: str
subtopics: List[SubTopic]
categories: List[str]
entities: Entities
sources: List[NewsSource] = Field(default_factory=list)
created_at: datetime = Field(default_factory=datetime.now)
generation_metadata: Dict[str, Any] = Field(default_factory=dict)
async def generate_article_with_claude(
news_data: Dict[str, Any],
style: str = "professional",
claude_api_key: str = None
) -> GeneratedArticle:
"""Claude API를 사용하여 기사 생성"""
if not claude_api_key:
import os
claude_api_key = os.getenv("CLAUDE_API_KEY")
# Initialize Claude client
claude_client = AsyncAnthropic(api_key=claude_api_key)
# Collect source information
sources_info = []
# Prepare the prompt
system_prompt = """당신은 전문적인 한국 언론사의 수석 기자입니다.
제공된 데이터를 기반으로 깊이 있고 통찰력 있는 기사를 작성해야 합니다.
기사는 다음 요구사항을 충족해야 합니다:
1. 소주제는 최소 2개, 최대 6개로 구성해야 합니다
2. 각 소주제는 최소 1개, 최대 10개의 문단으로 구성해야 합니다
3. 전문적이고 객관적인 어조를 유지해야 합니다
4. 사실에 기반한 분석과 통찰을 제공해야 합니다
5. 한국 독자를 대상으로 작성되어야 합니다
6. 이벤트 정보는 가능한 일시와 장소를 포함해야 합니다
7. 핵심 키워드를 최대 10개까지 추출해야 합니다
반드시 다음 JSON 형식으로 응답하세요:
{
"title": "기사 제목",
"summary": "한 줄 요약 (100자 이내)",
"subtopics": [
{
"title": "소주제 제목",
"content": ["문단1", "문단2", ...] // 1-10개 문단
}
], // 2-6개 소주제
"categories": ["카테고리1", "카테고리2"],
"entities": {
"people": ["인물1", "인물2"],
"organizations": ["기관1", "기관2"],
"groups": ["단체1", "단체2"],
"countries": ["나라1", "나라2"],
"events": [
{
"name": "이벤트명",
"date": "2025년 1월 15일", // 선택사항
"location": "서울 코엑스" // 선택사항
}
],
"keywords": ["키워드1", "키워드2", ...] // 최대 10개
}
}"""
# Prepare news content for Claude and collect sources
news_content = []
for item in news_data.get("news_items", []):
# Add RSS source info
rss_title = item.get('rss_title', '')
rss_link = item.get('rss_link', '')
rss_published = item.get('rss_published', '')
if rss_title and rss_link:
sources_info.append(NewsSource(
title=rss_title,
url=rss_link,
published_date=rss_published,
source_site="RSS Feed"
))
item_text = f"제목: {rss_title}\n"
for result in item.get("google_results", []):
# Add Google search result sources
if "title" in result and "link" in result:
sources_info.append(NewsSource(
title=result.get('title', ''),
url=result.get('link', ''),
published_date=None,
source_site="Google Search"
))
if "full_content" in result and result["full_content"]:
content = result["full_content"]
if isinstance(content, dict):
item_text += f"출처: {content.get('url', '')}\n"
item_text += f"내용: {content.get('content', '')[:1000]}...\n\n"
else:
item_text += f"내용: {str(content)[:1000]}...\n\n"
news_content.append(item_text)
combined_content = "\n".join(news_content[:10]) # Limit to prevent token overflow
user_prompt = f"""다음 뉴스 데이터를 기반으로 종합적인 기사를 작성하세요:
키워드: {news_data.get('keyword', '')}
수집된 뉴스 수: {len(news_data.get('news_items', []))}
뉴스 내용:
{combined_content}
스타일: {style}
- professional: 전통적인 뉴스 기사 스타일
- analytical: 분석적이고 심층적인 스타일
- investigative: 탐사보도 스타일
위의 데이터를 종합하여 통찰력 있는 기사를 JSON 형식으로 작성해주세요."""
try:
# Call Claude API
response = await claude_client.messages.create(
model="claude-3-5-sonnet-20241022",
max_tokens=4000,
temperature=0.7,
system=system_prompt,
messages=[
{"role": "user", "content": user_prompt}
]
)
# Parse response
content = response.content[0].text
# Extract JSON from response
json_start = content.find('{')
json_end = content.rfind('}') + 1
if json_start != -1 and json_end > json_start:
json_str = content[json_start:json_end]
article_data = json.loads(json_str)
else:
raise ValueError("No valid JSON found in response")
# Create article object
article = GeneratedArticle(
title=article_data.get("title", ""),
summary=article_data.get("summary", ""),
subtopics=[
SubTopic(
title=st.get("title", ""),
content=st.get("content", [])
) for st in article_data.get("subtopics", [])
],
categories=article_data.get("categories", []),
entities=Entities(**article_data.get("entities", {})),
sources=sources_info,
generation_metadata={
"style": style,
"keyword": news_data.get('keyword', ''),
"model": "claude-3-5-sonnet-20241022",
"timestamp": datetime.now().isoformat()
}
)
logger.info(f"Successfully generated article: {article.title}")
return article
except json.JSONDecodeError as e:
logger.error(f"Failed to parse Claude response as JSON: {e}")
raise
except Exception as e:
logger.error(f"Error generating article with Claude: {e}")
raise

View File

@ -0,0 +1,746 @@
"""
AI Writer Service
Claude API를 사용한 전문적인 뉴스 기사 생성 서비스
"""
from fastapi import FastAPI, HTTPException, BackgroundTasks
from fastapi.middleware.cors import CORSMiddleware
from typing import List, Dict, Any, Optional
from datetime import datetime
from pydantic import BaseModel, Field
import httpx
import asyncio
import logging
import json
import uuid
from anthropic import AsyncAnthropic
import os
# Configure logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
app = FastAPI(
title="AI Writer Service",
description="Claude API를 사용한 전문적인 뉴스 기사 생성 서비스",
version="1.0.0"
)
# CORS 설정
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
# Configuration
NEWS_AGGREGATOR_URL = os.getenv("NEWS_AGGREGATOR_URL", "http://news-aggregator-backend:8000")
CLAUDE_API_KEY = os.getenv("CLAUDE_API_KEY", "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")
MONGODB_URL = os.getenv("MONGODB_URL", "mongodb://mongodb:27017")
DB_NAME = os.getenv("DB_NAME", "ai_writer_db")
# Claude client
claude_client = AsyncAnthropic(api_key=CLAUDE_API_KEY)
# HTTP Client
http_client = httpx.AsyncClient(timeout=120.0)
# Queue Manager
from app.queue_manager import RedisQueueManager
from app.queue_models import NewsJobData, JobResult, JobStatus, QueueStats
queue_manager = RedisQueueManager(
redis_url=os.getenv("REDIS_URL", "redis://redis:6379")
)
# MongoDB client (optional for storing generated articles)
from motor.motor_asyncio import AsyncIOMotorClient
mongo_client = None
db = None
# Data Models
class NewsSource(BaseModel):
"""참고한 뉴스 소스 정보"""
title: str = Field(..., description="뉴스 제목")
url: str = Field(..., description="뉴스 URL")
published_date: Optional[str] = Field(None, description="발행일")
source_site: Optional[str] = Field(None, description="출처 사이트")
class SubTopic(BaseModel):
"""기사 소주제"""
title: str = Field(..., description="소주제 제목")
content: List[str] = Field(..., description="소주제 내용 (문단 리스트)", min_items=1, max_items=10)
class Event(BaseModel):
"""이벤트 정보"""
name: str = Field(..., description="이벤트명")
date: Optional[str] = Field(None, description="일시")
location: Optional[str] = Field(None, description="장소")
class NewsEntities(BaseModel):
"""뉴스에 포함된 개체들"""
people: List[str] = Field(default_factory=list, description="뉴스에 포함된 인물")
organizations: List[str] = Field(default_factory=list, description="뉴스에 포함된 기관")
groups: List[str] = Field(default_factory=list, description="뉴스에 포함된 단체")
countries: List[str] = Field(default_factory=list, description="뉴스에 포함된 나라")
events: List[Event] = Field(default_factory=list, description="뉴스에 포함된 일정/이벤트 (일시와 장소 포함)")
keywords: List[str] = Field(default_factory=list, description="핵심 키워드 (최대 10개)", max_items=10)
class GeneratedArticle(BaseModel):
"""생성된 기사"""
news_id: str = Field(..., description="뉴스 아이디")
title: str = Field(..., description="뉴스 제목")
created_at: str = Field(..., description="생성년월일시분초")
summary: str = Field(..., description="한 줄 요약")
subtopics: List[SubTopic] = Field(..., description="소주제 리스트", min_items=2, max_items=6)
categories: List[str] = Field(..., description="카테고리 리스트")
entities: NewsEntities = Field(..., description="뉴스에 포함된 개체들")
source_keyword: Optional[str] = Field(None, description="원본 검색 키워드")
source_count: Optional[int] = Field(None, description="참조한 소스 수")
sources: List[NewsSource] = Field(default_factory=list, description="참고한 뉴스 소스 목록")
class ArticleGenerationRequest(BaseModel):
"""기사 생성 요청"""
keyword: str = Field(..., description="검색 키워드")
limit: int = Field(5, description="처리할 RSS 항목 수", ge=1, le=20)
google_results_per_title: int = Field(3, description="각 제목당 구글 검색 결과 수", ge=1, le=10)
lang: str = Field("ko", description="언어 코드")
country: str = Field("KR", description="국가 코드")
style: str = Field("professional", description="기사 스타일 (professional/analytical/investigative)")
class PerItemGenerationRequest(BaseModel):
"""개별 아이템별 기사 생성 요청"""
keyword: str = Field(..., description="검색 키워드")
limit: Optional[int] = Field(None, description="처리할 RSS 항목 수 (None이면 전체)")
google_results_per_title: int = Field(3, description="각 제목당 구글 검색 결과 수", ge=1, le=10)
lang: str = Field("ko", description="언어 코드")
country: str = Field("KR", description="국가 코드")
style: str = Field("professional", description="기사 스타일 (professional/analytical/investigative)")
skip_existing: bool = Field(True, description="이미 생성된 기사는 건너뛰기")
@app.on_event("startup")
async def startup():
"""서비스 시작"""
global mongo_client, db
try:
mongo_client = AsyncIOMotorClient(MONGODB_URL)
db = mongo_client[DB_NAME]
logger.info("AI Writer Service starting...")
logger.info(f"Connected to MongoDB: {MONGODB_URL}")
# Redis 큐 연결
await queue_manager.connect()
logger.info("Connected to Redis queue")
except Exception as e:
logger.error(f"Failed to connect to services: {e}")
@app.on_event("shutdown")
async def shutdown():
"""서비스 종료"""
await http_client.aclose()
if mongo_client:
mongo_client.close()
await queue_manager.disconnect()
logger.info("AI Writer Service stopped")
@app.get("/")
async def root():
return {
"service": "AI Writer Service",
"version": "1.0.0",
"description": "Claude API를 사용한 전문적인 뉴스 기사 생성 서비스",
"endpoints": {
"generate_article": "POST /api/generate",
"generate_per_item": "POST /api/generate/per-item",
"generate_from_aggregated": "POST /api/generate/from-aggregated",
"get_article": "GET /api/articles/{article_id}",
"list_articles": "GET /api/articles",
"health": "GET /health"
}
}
@app.get("/health")
async def health_check():
"""헬스 체크"""
try:
# Check News Aggregator service
aggregator_response = await http_client.get(f"{NEWS_AGGREGATOR_URL}/health")
aggregator_healthy = aggregator_response.status_code == 200
# Check MongoDB
mongo_healthy = False
if db is not None:
await db.command("ping")
mongo_healthy = True
return {
"status": "healthy" if (aggregator_healthy and mongo_healthy) else "degraded",
"services": {
"news_aggregator": "healthy" if aggregator_healthy else "unhealthy",
"mongodb": "healthy" if mongo_healthy else "unhealthy",
"claude_api": "configured"
},
"timestamp": datetime.now().isoformat()
}
except Exception as e:
return {
"status": "unhealthy",
"error": str(e),
"timestamp": datetime.now().isoformat()
}
async def generate_article_with_claude(news_data: Dict[str, Any], style: str = "professional") -> GeneratedArticle:
"""Claude API를 사용하여 기사 생성"""
# Collect source information
sources_info = []
# Prepare the prompt
system_prompt = """당신은 전문적인 한국 언론사의 수석 기자입니다.
제공된 데이터를 기반으로 깊이 있고 통찰력 있는 기사를 작성해야 합니다.
기사는 다음 요구사항을 충족해야 합니다:
1. 소주제는 최소 2개, 최대 6개로 구성해야 합니다
2. 각 소주제는 최소 1개, 최대 10개의 문단으로 구성해야 합니다
3. 전문적이고 객관적인 어조를 유지해야 합니다
4. 사실에 기반한 분석과 통찰을 제공해야 합니다
5. 한국 독자를 대상으로 작성되어야 합니다
6. 이벤트 정보는 가능한 일시와 장소를 포함해야 합니다
7. 핵심 키워드를 최대 10개까지 추출해야 합니다
반드시 다음 JSON 형식으로 응답하세요:
{
"title": "기사 제목",
"summary": "한 줄 요약 (100자 이내)",
"subtopics": [
{
"title": "소주제 제목",
"content": ["문단1", "문단2", ...] // 1-10개 문단
}
], // 2-6개 소주제
"categories": ["카테고리1", "카테고리2"],
"entities": {
"people": ["인물1", "인물2"],
"organizations": ["기관1", "기관2"],
"groups": ["단체1", "단체2"],
"countries": ["나라1", "나라2"],
"events": [
{
"name": "이벤트명",
"date": "2025년 1월 15일", // 선택사항
"location": "서울 코엑스" // 선택사항
}
],
"keywords": ["키워드1", "키워드2", ...] // 최대 10개
}
}"""
# Prepare news content for Claude and collect sources
news_content = []
for item in news_data.get("news_items", []):
# Add RSS source info
rss_title = item.get('rss_title', '')
rss_link = item.get('rss_link', '')
rss_published = item.get('rss_published', '')
if rss_title and rss_link:
sources_info.append(NewsSource(
title=rss_title,
url=rss_link,
published_date=rss_published,
source_site="RSS Feed"
))
item_text = f"제목: {rss_title}\n"
for result in item.get("google_results", []):
# Add Google search result sources
if "title" in result and "link" in result:
sources_info.append(NewsSource(
title=result.get('title', ''),
url=result.get('link', ''),
published_date=None,
source_site="Google Search"
))
if "full_content" in result and result["full_content"]:
content = result["full_content"]
if isinstance(content, dict):
item_text += f"출처: {content.get('url', '')}\n"
item_text += f"내용: {content.get('content', '')[:1000]}...\n\n"
else:
item_text += f"내용: {str(content)[:1000]}...\n\n"
news_content.append(item_text)
combined_content = "\n".join(news_content[:10]) # Limit to prevent token overflow
user_prompt = f"""다음 뉴스 데이터를 기반으로 종합적인 기사를 작성하세요:
키워드: {news_data.get('keyword', '')}
수집된 뉴스 수: {len(news_data.get('news_items', []))}
뉴스 내용:
{combined_content}
스타일: {style}
- professional: 전통적인 뉴스 기사 스타일
- analytical: 분석적이고 심층적인 스타일
- investigative: 탐사보도 스타일
위의 데이터를 종합하여 통찰력 있는 기사를 JSON 형식으로 작성해주세요."""
try:
# Call Claude API
response = await claude_client.messages.create(
model="claude-3-5-sonnet-20241022", # Latest Claude model
max_tokens=4000,
temperature=0.7,
system=system_prompt,
messages=[
{"role": "user", "content": user_prompt}
]
)
# Parse Claude's response
content = response.content[0].text
# Extract JSON from response
import re
json_match = re.search(r'\{.*\}', content, re.DOTALL)
if json_match:
article_data = json.loads(json_match.group())
else:
# If no JSON found, try to parse the entire content
article_data = json.loads(content)
# Create GeneratedArticle object
entities_data = article_data.get("entities", {})
events_data = entities_data.get("events", [])
# Parse events - handle both old string format and new object format
parsed_events = []
for event in events_data:
if isinstance(event, str):
# Old format: just event name as string
parsed_events.append(Event(name=event))
elif isinstance(event, dict):
# New format: event object with name, date, location
parsed_events.append(Event(
name=event.get("name", ""),
date=event.get("date"),
location=event.get("location")
))
article = GeneratedArticle(
news_id=str(uuid.uuid4()),
title=article_data.get("title", "제목 없음"),
created_at=datetime.now().isoformat(),
summary=article_data.get("summary", ""),
subtopics=[
SubTopic(
title=st.get("title", ""),
content=st.get("content", [])
) for st in article_data.get("subtopics", [])
],
categories=article_data.get("categories", []),
entities=NewsEntities(
people=entities_data.get("people", []),
organizations=entities_data.get("organizations", []),
groups=entities_data.get("groups", []),
countries=entities_data.get("countries", []),
events=parsed_events,
keywords=entities_data.get("keywords", [])
),
source_keyword=news_data.get("keyword"),
source_count=len(news_data.get("news_items", [])),
sources=sources_info
)
return article
except Exception as e:
logger.error(f"Error generating article with Claude: {e}")
raise HTTPException(status_code=500, detail=f"Failed to generate article: {str(e)}")
@app.post("/api/generate")
async def generate_article(request: ArticleGenerationRequest):
"""
뉴스 수집부터 기사 생성까지 전체 파이프라인 실행
RSS → Google Search → AI 기사 생성
단일 종합 기사 생성 (기존 방식)
"""
try:
# Step 1: Get aggregated news from News Aggregator service
logger.info(f"Fetching aggregated news for keyword: {request.keyword}")
aggregator_response = await http_client.get(
f"{NEWS_AGGREGATOR_URL}/api/aggregate",
params={
"q": request.keyword,
"limit": request.limit,
"google_results_per_title": request.google_results_per_title,
"lang": request.lang,
"country": request.country
}
)
aggregator_response.raise_for_status()
news_data = aggregator_response.json()
if not news_data.get("news_items"):
raise HTTPException(status_code=404, detail="No news items found for the given keyword")
# Step 2: Generate article using Claude
logger.info(f"Generating article with Claude for {len(news_data['news_items'])} news items")
article = await generate_article_with_claude(news_data, request.style)
# Step 3: Store article in MongoDB (optional)
if db is not None:
try:
article_dict = article.dict()
await db.articles.insert_one(article_dict)
logger.info(f"Article saved with ID: {article.news_id}")
except Exception as e:
logger.error(f"Failed to save article to MongoDB: {e}")
return article
except httpx.HTTPStatusError as e:
logger.error(f"HTTP error from aggregator service: {e}")
raise HTTPException(status_code=e.response.status_code, detail=str(e))
except Exception as e:
logger.error(f"Error in generate_article: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.post("/api/generate/from-aggregated", response_model=GeneratedArticle)
async def generate_from_aggregated_data(news_data: Dict[str, Any], style: str = "professional"):
"""
이미 수집된 뉴스 데이터로부터 직접 기사 생성
(News Aggregator 결과를 직접 입력받아 처리)
"""
try:
if not news_data.get("news_items"):
raise HTTPException(status_code=400, detail="No news items in provided data")
# Generate article using Claude
logger.info(f"Generating article from {len(news_data['news_items'])} news items")
article = await generate_article_with_claude(news_data, style)
# Store article in MongoDB
if db is not None:
try:
article_dict = article.dict()
await db.articles.insert_one(article_dict)
logger.info(f"Article saved with ID: {article.news_id}")
except Exception as e:
logger.error(f"Failed to save article to MongoDB: {e}")
return article
except Exception as e:
logger.error(f"Error in generate_from_aggregated_data: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.get("/api/articles/{article_id}", response_model=GeneratedArticle)
async def get_article(article_id: str):
"""저장된 기사 조회"""
if db is None:
raise HTTPException(status_code=503, detail="Database not available")
article = await db.articles.find_one({"news_id": article_id})
if not article:
raise HTTPException(status_code=404, detail="Article not found")
# Convert MongoDB document to GeneratedArticle
article.pop("_id", None)
return GeneratedArticle(**article)
@app.get("/api/articles")
async def list_articles(
skip: int = 0,
limit: int = 10,
keyword: Optional[str] = None,
category: Optional[str] = None
):
"""저장된 기사 목록 조회"""
if db is None:
raise HTTPException(status_code=503, detail="Database not available")
query = {}
if keyword:
query["source_keyword"] = {"$regex": keyword, "$options": "i"}
if category:
query["categories"] = category
cursor = db.articles.find(query).skip(skip).limit(limit).sort("created_at", -1)
articles = []
async for article in cursor:
article.pop("_id", None)
articles.append(article)
total = await db.articles.count_documents(query)
return {
"articles": articles,
"total": total,
"skip": skip,
"limit": limit
}
@app.post("/api/generate/batch")
async def generate_batch_articles(keywords: List[str], style: str = "professional"):
"""여러 키워드에 대한 기사 일괄 생성"""
results = []
errors = []
for keyword in keywords[:5]: # Limit to 5 keywords to prevent overload
try:
request = ArticleGenerationRequest(
keyword=keyword,
style=style
)
article = await generate_article(request)
results.append({
"keyword": keyword,
"status": "success",
"article_id": article.news_id,
"title": article.title
})
except Exception as e:
errors.append({
"keyword": keyword,
"status": "error",
"error": str(e)
})
return {
"success": results,
"errors": errors,
"total_processed": len(results) + len(errors)
}
@app.post("/api/generate/per-item")
async def generate_articles_per_rss_item(request: PerItemGenerationRequest):
"""
RSS 피드의 각 아이템별로 개별 기사 생성
각 RSS 아이템이 독립적인 기사가 됨
중복 생성 방지 기능 포함
"""
try:
# Step 1: Get aggregated news from News Aggregator service
logger.info(f"Fetching aggregated news for keyword: {request.keyword}")
# limit이 None이면 모든 항목 처리 (최대 100개로 제한)
actual_limit = request.limit if request.limit is not None else 100
aggregator_response = await http_client.get(
f"{NEWS_AGGREGATOR_URL}/api/aggregate",
params={
"q": request.keyword,
"limit": actual_limit,
"google_results_per_title": request.google_results_per_title,
"lang": request.lang,
"country": request.country
}
)
aggregator_response.raise_for_status()
news_data = aggregator_response.json()
if not news_data.get("news_items"):
raise HTTPException(status_code=404, detail="No news items found for the given keyword")
# Step 2: Check for existing articles if skip_existing is True
existing_titles = set()
skipped_count = 0
if request.skip_existing and db is not None:
# RSS 제목으로 중복 체크 (최근 24시간 내)
from datetime import datetime, timedelta
cutoff_time = (datetime.now() - timedelta(hours=24)).isoformat()
existing_cursor = db.articles.find(
{
"source_keyword": request.keyword,
"created_at": {"$gte": cutoff_time}
},
{"sources": 1}
)
async for doc in existing_cursor:
for source in doc.get("sources", []):
if source.get("source_site") == "RSS Feed":
existing_titles.add(source.get("title", ""))
# Step 3: Generate individual article for each RSS item
generated_articles = []
for item in news_data["news_items"]:
try:
rss_title = item.get('rss_title', '')
# Skip if already exists
if request.skip_existing and rss_title in existing_titles:
logger.info(f"Skipping already generated article: {rss_title}")
skipped_count += 1
continue
logger.info(f"Generating article for RSS item: {rss_title or 'Unknown'}")
# Create individual news_data for this item
individual_news_data = {
"keyword": news_data.get("keyword"),
"news_items": [item] # Single item only
}
# Generate article for this single item
article = await generate_article_with_claude(individual_news_data, request.style)
# Store in MongoDB
if db is not None:
try:
article_dict = article.dict()
await db.articles.insert_one(article_dict)
logger.info(f"Article saved with ID: {article.news_id}")
except Exception as e:
logger.error(f"Failed to save article to MongoDB: {e}")
generated_articles.append(article)
except Exception as e:
logger.error(f"Failed to generate article for item: {e}")
# Continue with next item even if one fails
continue
if not generated_articles and skipped_count == 0:
raise HTTPException(status_code=500, detail="Failed to generate any articles")
# Return all generated articles
return {
"total_generated": len(generated_articles),
"total_items": len(news_data["news_items"]),
"skipped_duplicates": skipped_count,
"articles": generated_articles
}
except httpx.HTTPStatusError as e:
logger.error(f"HTTP error from aggregator service: {e}")
raise HTTPException(status_code=e.response.status_code, detail=str(e))
except Exception as e:
logger.error(f"Error in generate_articles_per_rss_item: {e}")
raise HTTPException(status_code=500, detail=str(e))
# Queue Management Endpoints
@app.post("/api/queue/enqueue")
async def enqueue_items(request: PerItemGenerationRequest):
"""
RSS 아이템들을 큐에 추가 (비동기 처리)
Consumer 워커가 백그라운드에서 처리
"""
try:
# Step 1: Get aggregated news from News Aggregator service
logger.info(f"Fetching aggregated news for enqueue: {request.keyword}")
actual_limit = request.limit if request.limit is not None else 100
aggregator_response = await http_client.get(
f"{NEWS_AGGREGATOR_URL}/api/aggregate",
params={
"q": request.keyword,
"limit": actual_limit,
"google_results_per_title": request.google_results_per_title,
"lang": request.lang,
"country": request.country
}
)
aggregator_response.raise_for_status()
news_data = aggregator_response.json()
if not news_data.get("news_items"):
raise HTTPException(status_code=404, detail="No news items found for the given keyword")
# Step 2: Check for existing articles if skip_existing is True
existing_titles = set()
skipped_count = 0
if request.skip_existing and db is not None:
from datetime import datetime, timedelta
cutoff_time = (datetime.now() - timedelta(hours=24)).isoformat()
existing_cursor = db.articles.find(
{
"source_keyword": request.keyword,
"created_at": {"$gte": cutoff_time}
},
{"sources": 1}
)
async for doc in existing_cursor:
for source in doc.get("sources", []):
if source.get("source_site") == "RSS Feed":
existing_titles.add(source.get("title", ""))
# Step 3: Enqueue items for processing
enqueued_jobs = []
for item in news_data["news_items"]:
rss_title = item.get('rss_title', '')
# Skip if already exists
if request.skip_existing and rss_title in existing_titles:
logger.info(f"Skipping already generated article: {rss_title}")
skipped_count += 1
continue
# Create job data
job_data = NewsJobData(
job_id=str(uuid.uuid4()),
keyword=request.keyword,
rss_title=rss_title,
rss_link=item.get('rss_link'),
rss_published=item.get('rss_published'),
google_results=item.get('google_results', []),
style=request.style,
created_at=datetime.now()
)
# Enqueue job
job_id = await queue_manager.enqueue(job_data)
enqueued_jobs.append({
"job_id": job_id,
"title": rss_title[:100]
})
logger.info(f"Enqueued job {job_id} for: {rss_title}")
return {
"total_enqueued": len(enqueued_jobs),
"total_items": len(news_data["news_items"]),
"skipped_duplicates": skipped_count,
"jobs": enqueued_jobs,
"message": f"{len(enqueued_jobs)} jobs added to queue for processing"
}
except httpx.HTTPStatusError as e:
logger.error(f"HTTP error from aggregator service: {e}")
raise HTTPException(status_code=e.response.status_code, detail=str(e))
except Exception as e:
logger.error(f"Error in enqueue_items: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.get("/api/queue/stats", response_model=QueueStats)
async def get_queue_stats():
"""큐 상태 및 통계 조회"""
try:
stats = await queue_manager.get_stats()
return stats
except Exception as e:
logger.error(f"Error getting queue stats: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.delete("/api/queue/clear")
async def clear_queue():
"""큐 초기화 (관리자용)"""
try:
await queue_manager.clear_queue()
return {"message": "Queue cleared successfully"}
except Exception as e:
logger.error(f"Error clearing queue: {e}")
raise HTTPException(status_code=500, detail=str(e))

View File

@ -0,0 +1,250 @@
"""
Redis Queue Manager for AI Writer Service
Redis를 사용한 작업 큐 관리
"""
import redis.asyncio as redis
import json
import uuid
from typing import Optional, List, Dict, Any
from datetime import datetime, timedelta
import logging
from queue_models import NewsJobData, JobResult, JobStatus, QueueStats
logger = logging.getLogger(__name__)
class RedisQueueManager:
"""Redis 기반 작업 큐 매니저"""
def __init__(self, redis_url: str = "redis://redis:6379"):
self.redis_url = redis_url
self.redis_client: Optional[redis.Redis] = None
# Redis 키 정의
self.QUEUE_KEY = "ai_writer:queue:pending"
self.PROCESSING_KEY = "ai_writer:queue:processing"
self.COMPLETED_KEY = "ai_writer:queue:completed"
self.FAILED_KEY = "ai_writer:queue:failed"
self.STATS_KEY = "ai_writer:stats"
self.WORKERS_KEY = "ai_writer:workers"
self.LOCK_PREFIX = "ai_writer:lock:"
async def connect(self):
"""Redis 연결"""
if not self.redis_client:
self.redis_client = await redis.from_url(
self.redis_url,
encoding="utf-8",
decode_responses=True
)
logger.info("Connected to Redis queue")
async def disconnect(self):
"""Redis 연결 해제"""
if self.redis_client:
await self.redis_client.close()
self.redis_client = None
logger.info("Disconnected from Redis queue")
async def enqueue(self, job_data: NewsJobData) -> str:
"""작업을 큐에 추가"""
try:
if not job_data.job_id:
job_data.job_id = str(uuid.uuid4())
# JSON으로 직렬화
job_json = job_data.json()
# 우선순위에 따라 큐에 추가
if job_data.priority > 0:
# 높은 우선순위는 앞쪽에
await self.redis_client.lpush(self.QUEUE_KEY, job_json)
else:
# 일반 우선순위는 뒤쪽에
await self.redis_client.rpush(self.QUEUE_KEY, job_json)
# 통계 업데이트
await self.redis_client.hincrby(self.STATS_KEY, "total_jobs", 1)
await self.redis_client.hincrby(self.STATS_KEY, "pending_jobs", 1)
logger.info(f"Job {job_data.job_id} enqueued")
return job_data.job_id
except Exception as e:
logger.error(f"Failed to enqueue job: {e}")
raise
async def dequeue(self, timeout: int = 0) -> Optional[NewsJobData]:
"""큐에서 작업 가져오기 (블로킹 가능)"""
try:
# 대기 중인 작업을 가져와서 처리 중 목록으로 이동
if timeout > 0:
result = await self.redis_client.blmove(
self.QUEUE_KEY,
self.PROCESSING_KEY,
timeout,
"LEFT",
"RIGHT"
)
else:
result = await self.redis_client.lmove(
self.QUEUE_KEY,
self.PROCESSING_KEY,
"LEFT",
"RIGHT"
)
if result:
# 통계 업데이트
await self.redis_client.hincrby(self.STATS_KEY, "pending_jobs", -1)
await self.redis_client.hincrby(self.STATS_KEY, "processing_jobs", 1)
return NewsJobData.parse_raw(result)
return None
except Exception as e:
logger.error(f"Failed to dequeue job: {e}")
return None
async def mark_completed(self, job_id: str, article_id: str):
"""작업을 완료로 표시"""
try:
# 처리 중 목록에서 작업 찾기
processing_jobs = await self.redis_client.lrange(self.PROCESSING_KEY, 0, -1)
for job_json in processing_jobs:
job = NewsJobData.parse_raw(job_json)
if job.job_id == job_id:
# 처리 중 목록에서 제거
await self.redis_client.lrem(self.PROCESSING_KEY, 1, job_json)
# 완료 결과 생성
result = JobResult(
job_id=job_id,
status=JobStatus.COMPLETED,
article_id=article_id,
completed_at=datetime.now()
)
# 완료 목록에 추가 (최대 1000개 유지)
await self.redis_client.lpush(self.COMPLETED_KEY, result.json())
await self.redis_client.ltrim(self.COMPLETED_KEY, 0, 999)
# 통계 업데이트
await self.redis_client.hincrby(self.STATS_KEY, "processing_jobs", -1)
await self.redis_client.hincrby(self.STATS_KEY, "completed_jobs", 1)
logger.info(f"Job {job_id} marked as completed")
break
except Exception as e:
logger.error(f"Failed to mark job as completed: {e}")
async def mark_failed(self, job_id: str, error_message: str):
"""작업을 실패로 표시"""
try:
# 처리 중 목록에서 작업 찾기
processing_jobs = await self.redis_client.lrange(self.PROCESSING_KEY, 0, -1)
for job_json in processing_jobs:
job = NewsJobData.parse_raw(job_json)
if job.job_id == job_id:
# 처리 중 목록에서 제거
await self.redis_client.lrem(self.PROCESSING_KEY, 1, job_json)
# 재시도 확인
if job.retry_count < job.max_retries:
job.retry_count += 1
# 다시 큐에 추가
await self.redis_client.rpush(self.QUEUE_KEY, job.json())
await self.redis_client.hincrby(self.STATS_KEY, "pending_jobs", 1)
logger.info(f"Job {job_id} requeued (retry {job.retry_count}/{job.max_retries})")
else:
# 실패 결과 생성
result = JobResult(
job_id=job_id,
status=JobStatus.FAILED,
error_message=error_message,
completed_at=datetime.now()
)
# 실패 목록에 추가
await self.redis_client.lpush(self.FAILED_KEY, result.json())
await self.redis_client.ltrim(self.FAILED_KEY, 0, 999)
# 통계 업데이트
await self.redis_client.hincrby(self.STATS_KEY, "failed_jobs", 1)
logger.error(f"Job {job_id} marked as failed: {error_message}")
await self.redis_client.hincrby(self.STATS_KEY, "processing_jobs", -1)
break
except Exception as e:
logger.error(f"Failed to mark job as failed: {e}")
async def get_stats(self) -> QueueStats:
"""큐 통계 조회"""
try:
stats_data = await self.redis_client.hgetall(self.STATS_KEY)
# 활성 워커 수 계산
workers = await self.redis_client.smembers(self.WORKERS_KEY)
active_workers = 0
for worker_id in workers:
# 워커가 최근 1분 이내에 활동했는지 확인
last_ping = await self.redis_client.get(f"{self.WORKERS_KEY}:{worker_id}")
if last_ping:
last_ping_time = datetime.fromisoformat(last_ping)
if datetime.now() - last_ping_time < timedelta(minutes=1):
active_workers += 1
return QueueStats(
pending_jobs=int(stats_data.get("pending_jobs", 0)),
processing_jobs=int(stats_data.get("processing_jobs", 0)),
completed_jobs=int(stats_data.get("completed_jobs", 0)),
failed_jobs=int(stats_data.get("failed_jobs", 0)),
total_jobs=int(stats_data.get("total_jobs", 0)),
workers_active=active_workers
)
except Exception as e:
logger.error(f"Failed to get stats: {e}")
return QueueStats(
pending_jobs=0,
processing_jobs=0,
completed_jobs=0,
failed_jobs=0,
total_jobs=0,
workers_active=0
)
async def register_worker(self, worker_id: str):
"""워커 등록"""
await self.redis_client.sadd(self.WORKERS_KEY, worker_id)
await self.redis_client.set(
f"{self.WORKERS_KEY}:{worker_id}",
datetime.now().isoformat(),
ex=300 # 5분 후 자동 만료
)
async def ping_worker(self, worker_id: str):
"""워커 활동 업데이트"""
await self.redis_client.set(
f"{self.WORKERS_KEY}:{worker_id}",
datetime.now().isoformat(),
ex=300
)
async def unregister_worker(self, worker_id: str):
"""워커 등록 해제"""
await self.redis_client.srem(self.WORKERS_KEY, worker_id)
await self.redis_client.delete(f"{self.WORKERS_KEY}:{worker_id}")
async def clear_queue(self):
"""큐 초기화 (테스트용)"""
await self.redis_client.delete(self.QUEUE_KEY)
await self.redis_client.delete(self.PROCESSING_KEY)
await self.redis_client.delete(self.COMPLETED_KEY)
await self.redis_client.delete(self.FAILED_KEY)
await self.redis_client.delete(self.STATS_KEY)
logger.info("Queue cleared")

View File

@ -0,0 +1,49 @@
"""
Queue Models for AI Writer Service
Redis 큐에서 사용할 데이터 모델 정의
"""
from pydantic import BaseModel, Field
from typing import Optional, List, Dict, Any
from datetime import datetime
from enum import Enum
class JobStatus(str, Enum):
"""작업 상태"""
PENDING = "pending"
PROCESSING = "processing"
COMPLETED = "completed"
FAILED = "failed"
SKIPPED = "skipped"
class NewsJobData(BaseModel):
"""큐에 들어갈 뉴스 작업 데이터"""
job_id: str = Field(..., description="작업 고유 ID")
keyword: str = Field(..., description="원본 검색 키워드")
rss_title: str = Field(..., description="RSS 제목")
rss_link: Optional[str] = Field(None, description="RSS 링크")
rss_published: Optional[str] = Field(None, description="RSS 발행일")
google_results: List[Dict[str, Any]] = Field(default_factory=list, description="구글 검색 결과")
style: str = Field("professional", description="기사 스타일")
created_at: datetime = Field(default_factory=datetime.now, description="작업 생성 시간")
priority: int = Field(0, description="우선순위 (높을수록 우선)")
retry_count: int = Field(0, description="재시도 횟수")
max_retries: int = Field(3, description="최대 재시도 횟수")
class JobResult(BaseModel):
"""작업 결과"""
job_id: str = Field(..., description="작업 고유 ID")
status: JobStatus = Field(..., description="작업 상태")
article_id: Optional[str] = Field(None, description="생성된 기사 ID")
error_message: Optional[str] = Field(None, description="에러 메시지")
processing_time: Optional[float] = Field(None, description="처리 시간(초)")
completed_at: Optional[datetime] = Field(None, description="완료 시간")
class QueueStats(BaseModel):
"""큐 통계"""
pending_jobs: int = Field(..., description="대기 중인 작업 수")
processing_jobs: int = Field(..., description="처리 중인 작업 수")
completed_jobs: int = Field(..., description="완료된 작업 수")
failed_jobs: int = Field(..., description="실패한 작업 수")
total_jobs: int = Field(..., description="전체 작업 수")
workers_active: int = Field(..., description="활성 워커 수")
average_processing_time: Optional[float] = Field(None, description="평균 처리 시간(초)")

View File

@ -0,0 +1,201 @@
"""
AI Writer Consumer Worker
큐에서 작업을 가져와 기사를 생성하는 백그라운드 워커
"""
import asyncio
import logging
import signal
import sys
import uuid
from datetime import datetime
from typing import Optional
import os
from motor.motor_asyncio import AsyncIOMotorClient
from anthropic import AsyncAnthropic
from queue_manager import RedisQueueManager
from queue_models import NewsJobData, JobStatus
from article_generator import generate_article_with_claude
# Configure logging
logging.basicConfig(
level=logging.INFO,
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger(__name__)
class AIWriterWorker:
"""AI Writer 백그라운드 워커"""
def __init__(self, worker_id: Optional[str] = None):
self.worker_id = worker_id or str(uuid.uuid4())
self.queue_manager = RedisQueueManager(
redis_url=os.getenv("REDIS_URL", "redis://redis:6379")
)
# MongoDB 설정
self.mongodb_url = os.getenv("MONGODB_URL", "mongodb://mongodb:27017")
self.db_name = os.getenv("DB_NAME", "ai_writer_db")
self.mongo_client = None
self.db = None
# Claude 클라이언트
self.claude_api_key = os.getenv("CLAUDE_API_KEY")
self.claude_client = AsyncAnthropic(api_key=self.claude_api_key)
# 실행 상태
self.running = False
self.tasks = []
async def start(self, num_workers: int = 1):
"""워커 시작"""
logger.info(f"Starting AI Writer Worker {self.worker_id} with {num_workers} concurrent workers")
try:
# Redis 연결
await self.queue_manager.connect()
await self.queue_manager.register_worker(self.worker_id)
# MongoDB 연결
self.mongo_client = AsyncIOMotorClient(self.mongodb_url)
self.db = self.mongo_client[self.db_name]
logger.info("Connected to MongoDB")
self.running = True
# 여러 워커 태스크 생성
for i in range(num_workers):
task = asyncio.create_task(self._process_jobs(f"{self.worker_id}-{i}"))
self.tasks.append(task)
# 워커 핑 태스크
ping_task = asyncio.create_task(self._ping_worker())
self.tasks.append(ping_task)
# 모든 태스크 대기
await asyncio.gather(*self.tasks)
except Exception as e:
logger.error(f"Worker error: {e}")
finally:
await self.stop()
async def stop(self):
"""워커 정지"""
logger.info(f"Stopping AI Writer Worker {self.worker_id}")
self.running = False
# 태스크 취소
for task in self.tasks:
task.cancel()
# 워커 등록 해제
await self.queue_manager.unregister_worker(self.worker_id)
# 연결 해제
await self.queue_manager.disconnect()
if self.mongo_client:
self.mongo_client.close()
logger.info(f"Worker {self.worker_id} stopped")
async def _process_jobs(self, sub_worker_id: str):
"""작업 처리 루프"""
logger.info(f"Sub-worker {sub_worker_id} started")
while self.running:
try:
# 큐에서 작업 가져오기 (5초 타임아웃)
job = await self.queue_manager.dequeue(timeout=5)
if job:
logger.info(f"[{sub_worker_id}] Processing job {job.job_id}: {job.rss_title[:50]}")
start_time = datetime.now()
try:
# 기사 생성
article = await self._generate_article(job)
# MongoDB에 저장
if article and self.db is not None:
article_dict = article.dict()
await self.db.articles.insert_one(article_dict)
# 처리 시간 계산
processing_time = (datetime.now() - start_time).total_seconds()
# 완료 표시
await self.queue_manager.mark_completed(
job.job_id,
article.news_id
)
logger.info(f"[{sub_worker_id}] Job {job.job_id} completed in {processing_time:.2f}s")
else:
raise Exception("Failed to generate article")
except Exception as e:
logger.error(f"[{sub_worker_id}] Job {job.job_id} failed: {e}")
await self.queue_manager.mark_failed(job.job_id, str(e))
except asyncio.CancelledError:
break
except Exception as e:
logger.error(f"[{sub_worker_id}] Worker error: {e}")
await asyncio.sleep(1)
logger.info(f"Sub-worker {sub_worker_id} stopped")
async def _generate_article(self, job: NewsJobData):
"""기사 생성"""
# 작업 데이터를 기존 형식으로 변환
news_data = {
"keyword": job.keyword,
"news_items": [{
"rss_title": job.rss_title,
"rss_link": job.rss_link,
"rss_published": job.rss_published,
"google_results": job.google_results
}]
}
# 기사 생성 (기존 함수 재사용)
return await generate_article_with_claude(news_data, job.style)
async def _ping_worker(self):
"""워커 활동 신호 전송"""
while self.running:
try:
await self.queue_manager.ping_worker(self.worker_id)
await asyncio.sleep(30) # 30초마다 핑
except asyncio.CancelledError:
break
except Exception as e:
logger.error(f"Ping error: {e}")
def signal_handler(signum, frame):
"""시그널 핸들러"""
logger.info(f"Received signal {signum}")
sys.exit(0)
async def main():
"""메인 함수"""
# 시그널 핸들러 등록
signal.signal(signal.SIGINT, signal_handler)
signal.signal(signal.SIGTERM, signal_handler)
# 워커 수 설정 (환경변수 또는 기본값)
num_workers = int(os.getenv("WORKER_COUNT", "3"))
# 워커 시작
worker = AIWriterWorker()
try:
await worker.start(num_workers=num_workers)
except KeyboardInterrupt:
logger.info("Keyboard interrupt received")
finally:
await worker.stop()
if __name__ == "__main__":
asyncio.run(main())

View File

@ -0,0 +1,62 @@
{
"news_id": "49bdf2f3-4dbc-47eb-8c49-5d9536f41d87",
"title": "유럽 전기차 시장의 새로운 전환점: 현대차·기아의 소형 전기차 전략과 글로벌 경쟁 구도",
"created_at": "2025-09-13T00:29:13.376541",
"summary": "현대차와 기아가 IAA 2025에서 소형 전기차 콘셉트 모델을 공개하며 유럽 시장 공략을 가속화, 배터리 협력과 가격 경쟁력으로 승부수",
"subtopics": [
{
"title": "현대차·기아의 유럽 소형 전기차 시장 공략",
"content": [
"현대자동차와 기아가 IAA 2025에서 콘셉트 쓰리와 EV2를 공개하며 유럽 소형 전기차 시장 공략에 박차를 가하고 있다. 이는 유럽의 급성장하는 소형 전기차 수요에 대응하기 위한 전략적 움직임으로 평가된다.",
"특히 두 모델은 실용성과 경제성을 모두 갖춘 제품으로, 유럽 소비자들의 니즈를 정확히 겨냥했다는 평가를 받고 있다. 현대차그룹은 이를 통해 유럽 시장에서의 입지를 더욱 강화할 것으로 전망된다.",
"현지 전문가들은 현대차그룹의 이번 전략이 유럽 전기차 시장의 '골든타임'을 잡기 위한 시의적절한 움직임이라고 분석하고 있다."
]
},
{
"title": "배터리 공급망 전략의 중요성 부각",
"content": [
"전기차 시장에서 배터리 공급망 확보가 핵심 경쟁력으로 부상하고 있다. IAA 모빌리티에서 폴스타가 SK온을 배터리 파트너로 공개적으로 언급한 것이 주목받고 있다.",
"배터리 제조사 선정에 대한 정보가 제한적인 가운데, 안정적인 배터리 공급망 구축이 전기차 제조사들의 성패를 좌우할 것으로 예상된다.",
"특히 소형 전기차의 경우 가격 경쟁력이 중요한 만큼, 효율적인 배터리 수급 전략이 시장 점유율 확대의 관건이 될 전망이다."
]
},
{
"title": "글로벌 전기차 시장의 경쟁 구도 변화",
"content": [
"유럽 전기차 시장에서 소형 모델을 중심으로 한 경쟁이 본격화되면서, 제조사들의 전략적 포지셔닝이 더욱 중요해지고 있다.",
"현대차그룹은 품질과 기술력을 바탕으로 한 프리미엄 이미지와 함께, 합리적인 가격대의 소형 전기차 라인업으로 시장 공략을 가속화하고 있다.",
"이러한 변화는 글로벌 자동차 산업의 패러다임 전환을 반영하며, 향후 전기차 시장의 주도권 경쟁이 더욱 치열해질 것으로 예상된다."
]
}
],
"categories": [
"자동차",
"경제",
"환경",
"기술"
],
"entities": {
"people": [],
"organizations": [
"현대자동차",
"기아",
"SK온",
"폴스타"
],
"groups": [
"유럽 자동차 제조사",
"배터리 제조업체"
],
"countries": [
"대한민국",
"독일",
"유럽연합"
],
"events": [
"IAA 2025",
"IAA 모빌리티"
]
},
"source_keyword": "전기차",
"source_count": 3
}

View File

@ -0,0 +1,66 @@
{
"news_id": "8a51bead-4558-4351-a5b2-b5e5ba1b3d38",
"title": "현대차·기아, 유럽 전기차 시장서 소형 모델로 새 돌파구 모색",
"created_at": "2025-09-13T00:29:35.661926",
"summary": "IAA 모빌리티 2025에서 현대차·기아가 소형 전기차 콘셉트카를 공개하며 유럽 시장 공략 가속화. 배터리 공급망 확보와 가격 경쟁력이 성공 관건",
"subtopics": [
{
"title": "유럽 소형 전기차 시장 공략 본격화",
"content": [
"현대차와 기아가 IAA 모빌리티 2025에서 각각 콘셉트 쓰리와 EV2를 공개하며 유럽 소형 전기차 시장 공략에 시동을 걸었다. 이는 유럽의 높은 환경 규제와 도심 이동성 수요에 대응하기 위한 전략적 움직임으로 해석된다.",
"특히 두 모델은 기존 전기차 대비 컴팩트한 사이즈와 효율적인 배터리 시스템을 갖추고 있어, 유럽 소비자들의 실용적 수요를 겨냥했다는 평가를 받고 있다.",
"업계 전문가들은 현대차그룹의 이번 행보가 테슬라와 중국 업체들이 주도하고 있는 유럽 전기차 시장에서 새로운 돌파구를 마련할 수 있을 것으로 전망하고 있다."
]
},
{
"title": "배터리 공급망 확보 과제",
"content": [
"전기차 성공의 핵심 요소인 배터리 수급에서 SK온이 주요 공급 파트너로 부상했다. 폴스타가 SK온을 배터리 공급사로 공개적으로 언급한 것이 이를 방증한다.",
"그러나 업계에서는 배터리 제조사들의 정보 공개가 제한적이어서 실제 공급망 구조를 파악하기 어려운 상황이다. 이는 글로벌 배터리 수급 경쟁이 치열해지고 있음을 시사한다.",
"안정적인 배터리 공급망 확보는 향후 소형 전기차의 가격 경쟁력과 직결되는 만큼, 현대차그룹의 추가적인 파트너십 구축이 예상된다."
]
},
{
"title": "가격 경쟁력 확보 전략",
"content": [
"소형 전기차 시장에서의 성공을 위해서는 합리적인 가격대 책정이 필수적이다. 현대차그룹은 규모의 경제를 통한 원가 절감을 목표로 하고 있다.",
"특히 유럽 시장에서는 테슬라와 중국 업체들의 공격적인 가격 정책에 대응해야 하는 상황이다. 현대차그룹은 프리미엄 품질을 유지하면서도 경쟁력 있는 가격대를 제시하는 것을 목표로 하고 있다.",
"전문가들은 배터리 기술 혁신과 생산 효율화를 통해 가격 경쟁력을 확보하는 것이 향후 성공의 핵심이 될 것으로 전망하고 있다."
]
}
],
"categories": [
"자동차",
"경제",
"산업",
"기술"
],
"entities": {
"people": [
"김성수",
"조용하",
"박종면"
],
"organizations": [
"현대자동차",
"기아",
"SK온",
"폴스타"
],
"groups": [
"유럽 자동차 제조사",
"중국 전기차 업체"
],
"countries": [
"대한민국",
"독일",
"중국"
],
"events": [
"IAA 모빌리티 2025",
"전기차 배터리 공급 계약"
]
},
"source_keyword": "전기차",
"source_count": 3
}

View File

@ -0,0 +1,62 @@
{
"news_id": "2c4cb595-9542-45ee-b4b9-2135c46950e3",
"title": "현대차·기아, 유럽 전기차 시장서 소형 모델로 승부수...배터리 협력 강화 주목",
"created_at": "2025-09-13T00:28:51.371773",
"summary": "현대차·기아가 유럽 전기차 시장에서 콘셉트 쓰리와 EV2로 소형 전기차 시장 공략 나서, 배터리 협력사 선정 등 경쟁력 강화 움직임 본격화",
"subtopics": [
{
"title": "유럽 소형 전기차 시장 공략 본격화",
"content": [
"현대자동차그룹이 유럽 전기차 시장 공략을 위해 소형 전기차 라인업 확대에 나섰다. IAA 모빌리티 2025에서 공개된 현대차의 콘셉트 쓰리와 기아의 EV2는 유럽 시장 맞춤형 전략의 핵심으로 평가받고 있다.",
"특히 소형 전기차 시장은 유럽에서 급성장이 예상되는 세그먼트로, 현대차그룹은 합리적인 가격대와 실용성을 앞세워 시장 선점을 노리고 있다.",
"현대차그룹의 이번 전략은 유럽의 환경 규제 강화와 소비자들의 실용적인 전기차 수요 증가에 대응하는 동시에, 중국 전기차 업체들의 유럽 진출에 대한 선제적 대응으로 해석된다."
]
},
{
"title": "배터리 협력 관계 재편 움직임",
"content": [
"전기차 경쟁력의 핵심인 배터리 수급과 관련해 업계의 이목이 집중되고 있다. IAA 모빌리티에서 폴스타가 SK온을 배터리 공급사로 지목한 것이 주목받고 있다.",
"글로벌 자동차 업체들의 배터리 조달 전략이 다변화되는 가운데, 한국 배터리 업체들과의 협력 강화 움직임이 감지되고 있다.",
"특히 현대차그룹은 안정적인 배터리 수급을 위해 다양한 배터리 제조사들과의 협력 관계를 검토 중인 것으로 알려졌다."
]
},
{
"title": "글로벌 전기차 시장 경쟁 심화",
"content": [
"전기차 시장에서 브랜드 간 경쟁이 치열해지는 가운데, 현대차그룹은 차별화된 제품 라인업과 기술력으로 시장 지위 강화에 나서고 있다.",
"특히 유럽 시장에서는 테슬라, 폭스바겐 그룹, 중국 업체들과의 경쟁이 불가피한 상황이며, 현대차그룹은 품질과 기술력을 앞세워 경쟁력 확보에 주력하고 있다.",
"시장 전문가들은 현대차그룹의 소형 전기차 전략이 향후 글로벌 시장에서의 입지 강화에 중요한 전환점이 될 것으로 전망하고 있다."
]
}
],
"categories": [
"자동차",
"경제",
"산업"
],
"entities": {
"people": [
"김성수",
"박영효"
],
"organizations": [
"현대자동차",
"기아",
"SK온",
"폴스타"
],
"groups": [
"현대차그룹",
"폭스바겐 그룹"
],
"countries": [
"대한민국",
"독일"
],
"events": [
"IAA 모빌리티 2025"
]
},
"source_keyword": "전기차",
"source_count": 3
}

View File

@ -0,0 +1,63 @@
{
"news_id": "ee154fb8-a913-4aa9-9fc9-fa421fd2d7c0",
"title": "2025년 기술 혁신의 분기점: AI·양자컴퓨팅이 그리는 새로운 미래",
"created_at": "2025-09-13T00:32:14.008706",
"summary": "2025년, AI와 양자컴퓨팅의 상용화가 가져올 산업 전반의 혁신적 변화와 사회적 영향을 심층 분석한 전망",
"subtopics": [
{
"title": "생성형 AI가 재편하는 산업 생태계",
"content": [
"2025년은 생성형 AI가 산업 전반에 본격적으로 도입되는 원년이 될 전망이다. 특히 의료 진단, 신약 개발, 교육 커리큘럼 설계 등 전문 분야에서 AI의 역할이 획기적으로 확대될 것으로 예측된다.",
"기업들의 업무 프로세스도 근본적인 변화를 맞이할 것으로 보인다. 창의적 작업 영역에서도 AI의 활용이 일상화되며, 인간-AI 협업 모델이 새로운 표준으로 자리잡을 것으로 전망된다.",
"다만 AI 도입에 따른 노동시장 재편과 윤리적 문제에 대한 사회적 합의가 시급한 과제로 대두될 것으로 예상된다. 특히 AI 의존도 증가에 따른 데이터 보안과 알고리즘 편향성 문제는 중요한 해결 과제가 될 것이다."
]
},
{
"title": "양자컴퓨팅의 상용화와 산업혁신",
"content": [
"양자컴퓨팅 기술이 실용화 단계에 진입하면서, 금융권의 리스크 분석과 암호화폐 보안 시스템에 획기적인 변화가 예상된다. 특히 복잡한 금융 모델링과 시장 예측에서 양자컴퓨터의 활용이 크게 증가할 전망이다.",
"제약 산업에서는 신약 개발 프로세스가 대폭 단축될 것으로 기대된다. 양자컴퓨터를 활용한 분자 시뮬레이션이 가능해지면서, 신약 개발 비용 절감과 효율성 증대가 실현될 것이다.",
"물류 및 공급망 관리 분야에서도 양자컴퓨팅의 영향력이 확대될 전망이다. 복잡한 경로 최적화와 재고 관리에 양자 알고리즘을 적용함으로써, 물류 비용 절감과 효율성 향상이 가능해질 것으로 예측된다."
]
},
{
"title": "기술 혁신에 따른 사회경제적 변화",
"content": [
"AI와 양자컴퓨팅의 발전은 노동시장의 구조적 변화를 가속화할 것으로 전망된다. 단순 반복 업무는 자동화되는 반면, AI 시스템 관리와 양자컴퓨팅 전문가 같은 새로운 직종의 수요가 급증할 것으로 예상된다.",
"교육 시스템도 큰 변화를 맞이할 것으로 보인다. AI 기반 맞춤형 학습과 양자컴퓨팅 원리에 대한 이해가 새로운 필수 교육과정으로 자리잡을 것으로 전망된다.",
"이러한 기술 혁신은 국가 간 기술 격차를 더욱 심화시킬 가능성이 있다. 선진국과 개발도상국 간의 디지털 격차 해소가 국제사회의 주요 과제로 대두될 것으로 예측된다."
]
}
],
"categories": [
"기술",
"산업",
"미래전망",
"경제"
],
"entities": {
"people": [],
"organizations": [
"금융권",
"제약회사",
"물류기업"
],
"groups": [
"AI 개발자",
"양자컴퓨팅 전문가",
"교육기관"
],
"countries": [
"한국",
"미국",
"중국"
],
"events": [
"AI 상용화",
"양자컴퓨터 실용화",
"디지털 전환"
]
},
"source_keyword": "2025년 기술 트렌드",
"source_count": 2
}

View File

@ -0,0 +1,62 @@
{
"news_id": "3109c578-9b08-4cd0-a9d6-3d92b97e64d4",
"title": "2025년 기술 혁신의 물결, AI·양자컴퓨팅이 이끄는 새로운 패러다임",
"created_at": "2025-09-13T00:31:52.782760",
"summary": "2025년, 생성형 AI와 양자컴퓨팅의 상용화로 산업 전반에 혁신적 변화가 예상되며, 인간-AI 협업이 일상화될 전망",
"subtopics": [
{
"title": "생성형 AI가 주도하는 창의적 혁신",
"content": [
"2025년은 생성형 AI 기술이 전례 없는 수준으로 발전하여 창의적 영역에서도 획기적인 변화가 예상된다. 기존에 인간의 고유 영역으로 여겨졌던 예술 창작, 콘텐츠 제작, 디자인 분야에서 AI가 핵심 협력자로 자리잡을 전망이다.",
"특히 의료 분야에서는 AI가 질병 진단과 치료 계획 수립에 적극적으로 활용될 것으로 예측된다. AI는 방대한 의료 데이터를 분석하여 개인 맞춤형 치료법을 제시하고, 의료진의 의사결정을 효과적으로 지원할 것으로 기대된다.",
"교육 분야에서도 AI 기반의 맞춤형 학습 시스템이 보편화될 전망이다. 학습자의 이해도와 진도에 따라 최적화된 커리큘럼을 제공하고, 실시간으로 학습 성과를 분석하여 개선점을 제시하는 등 교육의 질적 향상이 기대된다."
]
},
{
"title": "양자컴퓨팅의 산업 혁신 주도",
"content": [
"2025년은 양자컴퓨팅이 실용화 단계에 진입하는 원년이 될 것으로 전망된다. 특히 금융 산업에서는 복잡한 위험 분석과 포트폴리오 최적화에 양자컴퓨팅을 활용하여 투자 전략의 정확도를 높일 것으로 예상된다.",
"제약 산업에서는 양자컴퓨터를 활용한 신약 개발이 가속화될 전망이다. 분자 구조 시뮬레이션과 신약 후보 물질 스크리닝 과정에서 양자컴퓨팅의 강점이 발휘될 것으로 기대된다.",
"물류 분야에서도 양자컴퓨팅을 통한 최적화가 실현될 전망이다. 복잡한 공급망 관리와 배송 경로 최적화에 양자컴퓨팅을 도입함으로써 물류 비용 절감과 효율성 향상이 가능해질 것으로 예측된다."
]
},
{
"title": "인간-기계 협업의 새로운 패러다임",
"content": [
"2025년에는 AI와 인간의 협업이 일상화되면서 업무 방식의 근본적인 변화가 예상된다. 단순 반복적인 업무는 AI가 담당하고, 인간은 전략적 의사결정과 창의적 문제 해결에 집중하는 방식으로 업무 분담이 이루어질 것이다.",
"이러한 변화는 노동시장의 구조적 변화로 이어질 전망이다. AI와 협업할 수 있는 디지털 역량이 필수적인 직무 역량으로 부상하며, 새로운 형태의 직업이 등장할 것으로 예측된다.",
"하지만 이러한 변화 속에서도 윤리적 판단과 감성적 소통과 같은 인간 고유의 가치는 더욱 중요해질 것으로 전망된다. 기술 발전이 가져올 혜택을 최대화하면서도 인간 중심의 가치를 지켜나가는 균형이 중요한 과제로 대두될 것이다."
]
}
],
"categories": [
"기술",
"미래전망",
"산업동향"
],
"entities": {
"people": [],
"organizations": [
"AI 기업들",
"제약회사들",
"물류기업들"
],
"groups": [
"의료진",
"교육자",
"기술전문가"
],
"countries": [
"한국",
"미국",
"중국"
],
"events": [
"2025년 기술혁신",
"양자컴퓨팅 상용화",
"AI 혁명"
]
},
"source_keyword": "2025년 기술 트렌드",
"source_count": 2
}

View File

@ -0,0 +1,73 @@
{
"news_id": "ea9f3734-6a93-4ca7-8ebe-b85612e2fd0a",
"title": "정부, 내년 AI 산업에 10조원 투자...한국 경제 체질 대전환 나선다",
"created_at": "2025-09-13T01:09:43.892704",
"summary": "정부가 2025년 인공지능 산업 육성을 위해 10조원 규모의 대규모 투자를 단행하며 디지털 경제 전환 가속화에 나선다",
"subtopics": [
{
"title": "정부의 AI 산업 육성 청사진",
"content": [
"정부가 2025년 인공지능(AI) 산업 육성을 위해 10조원 규모의 투자를 단행한다. 이는 한국 경제의 디지털 전환을 가속화하고 글로벌 AI 강국으로 도약하기 위한 전략적 결정이다.",
"투자의 주요 방향은 AI 기술 개발, 인프라 구축, 전문인력 양성 등으로, 특히 반도체와 같은 핵심 산업과의 시너지 창출에 중점을 둘 예정이다."
]
},
{
"title": "민관 협력 체계 구축",
"content": [
"정부는 AI 산업 육성을 위해 대기업, 스타트업, 연구기관 등과의 협력 체계를 강화한다. 소버린AI를 비롯한 국내 AI 기업들과의 협력을 통해 실질적인 세계 2위 AI 강국 도약을 목표로 하고 있다.",
"특히 AI 전문가 공모와 전담 조직 신설 등을 통해 체계적인 산업 육성 기반을 마련할 계획이다."
]
},
{
"title": "글로벌 경쟁력 강화 전략",
"content": [
"정부는 국내 AI 기업들의 글로벌 경쟁력 강화를 위해 기술 개발 지원, 해외 시장 진출 지원, 규제 개선 등 다각적인 지원책을 마련한다.",
"특히 AI 산업의 핵심 인프라인 반도체 분야에서 SK하이닉스의 HBM4 개발 완료 등 가시적인 성과가 나타나고 있어, 이를 기반으로 한 시너지 효과가 기대된다."
]
}
],
"categories": [
"경제",
"기술",
"산업정책"
],
"entities": {
"people": [
"하정우 소버린AI 대표"
],
"organizations": [
"소버린AI",
"SK하이닉스",
"과학기술정보통신부"
],
"groups": [
"AI 기업",
"스타트업"
],
"countries": [
"대한민국",
"미국"
],
"events": [
{
"name": "2025년 AI 산업 육성 계획 발표",
"date": "2025년",
"location": "대한민국"
}
],
"keywords": [
"인공지능",
"AI 산업",
"디지털 전환",
"10조원 투자",
"반도체",
"HBM4",
"글로벌 경쟁력",
"민관협력",
"전문인력 양성",
"기술개발"
]
},
"source_keyword": "인공지능",
"source_count": 5
}

View File

@ -0,0 +1,9 @@
fastapi==0.104.1
uvicorn[standard]==0.24.0
httpx==0.25.2
pydantic==2.5.0
motor==3.1.1
pymongo==4.3.3
anthropic==0.39.0
python-multipart==0.0.6
redis[hiredis]==5.0.1

View File

@ -0,0 +1,168 @@
#!/usr/bin/env python3
"""
AI Writer Service Test
Claude API를 사용한 전문적인 뉴스 기사 생성 테스트
"""
import asyncio
import httpx
import json
from datetime import datetime
# Service URL
SERVICE_URL = "http://localhost:8019"
async def test_article_generation():
"""인공지능 키워드로 기사 생성 테스트"""
async with httpx.AsyncClient(timeout=120.0) as client:
print("\n" + "="*70)
print(" AI Writer Service - 전문 기사 생성 테스트 ")
print("="*70)
print("\n📰 '인공지능' 키워드로 전문 기사 생성 중...")
print("-" * 50)
# Generate article
response = await client.post(
f"{SERVICE_URL}/api/generate",
json={
"keyword": "인공지능",
"limit": 5,
"google_results_per_title": 3,
"lang": "ko",
"country": "KR",
"style": "professional"
}
)
if response.status_code == 200:
article = response.json()
print(f"\n✅ 기사 생성 완료!")
print(f"\n📌 기사 ID: {article['news_id']}")
print(f"📅 생성 시간: {article['created_at']}")
print(f"\n📰 제목: {article['title']}")
print(f"📝 요약: {article['summary']}")
print(f"\n🔍 카테고리: {', '.join(article['categories'])}")
# Print subtopics
print(f"\n📚 소주제 ({len(article['subtopics'])}개):")
for i, subtopic in enumerate(article['subtopics'], 1):
print(f"\n [{i}] {subtopic['title']}")
print(f" 문단 수: {len(subtopic['content'])}")
for j, paragraph in enumerate(subtopic['content'][:1], 1): # Show first paragraph only
print(f" 미리보기: {paragraph[:150]}...")
# Print entities
entities = article['entities']
print(f"\n🏷️ 추출된 개체:")
if entities['people']:
print(f" 👤 인물: {', '.join(entities['people'])}")
if entities['organizations']:
print(f" 🏢 기관: {', '.join(entities['organizations'])}")
if entities['groups']:
print(f" 👥 단체: {', '.join(entities['groups'])}")
if entities['countries']:
print(f" 🌍 국가: {', '.join(entities['countries'])}")
if entities.get('events'):
events = entities['events']
if events:
print(f" 📅 이벤트 ({len(events)}개):")
for evt in events[:3]: # 처음 3개만 표시
if isinstance(evt, dict):
evt_str = f" - {evt.get('name', '')}"
if evt.get('date'):
evt_str += f" [{evt['date']}]"
if evt.get('location'):
evt_str += f" @{evt['location']}"
print(evt_str)
else:
# 이전 형식 (문자열) 지원
print(f" - {evt}")
if entities.get('keywords'):
keywords = entities['keywords']
if keywords:
print(f" 🔑 키워드: {', '.join(keywords[:5])}" +
("..." if len(keywords) > 5 else ""))
print(f"\n📊 참조 소스: {article.get('source_count', 0)}")
# Save full article to file
with open('generated_article.json', 'w', encoding='utf-8') as f:
json.dump(article, f, ensure_ascii=False, indent=2)
print(f"\n💾 전체 기사가 'generated_article.json'에 저장되었습니다.")
else:
print(f"❌ 오류: {response.status_code}")
print(f" 상세: {response.text}")
async def test_health_check():
"""서비스 상태 확인"""
async with httpx.AsyncClient() as client:
print("\n" + "="*60)
print("서비스 Health Check")
print("="*60)
response = await client.get(f"{SERVICE_URL}/health")
if response.status_code == 200:
data = response.json()
print(f"✓ AI Writer 서비스 상태: {data.get('status', 'unknown')}")
if 'services' in data:
print(f" - News Aggregator: {data['services'].get('news_aggregator', 'unknown')}")
print(f" - MongoDB: {data['services'].get('mongodb', 'unknown')}")
print(f" - Claude API: {data['services'].get('claude_api', 'unknown')}")
if 'error' in data:
print(f" - Error: {data['error']}")
else:
print(f"✗ Health check 실패: {response.status_code}")
async def test_batch_generation():
"""여러 키워드 일괄 처리 테스트"""
async with httpx.AsyncClient(timeout=180.0) as client:
print("\n" + "="*60)
print("일괄 기사 생성 테스트")
print("="*60)
keywords = ["AI 혁신", "디지털 전환", "스마트시티"]
print(f"\n키워드: {', '.join(keywords)}")
response = await client.post(
f"{SERVICE_URL}/api/generate/batch",
json=keywords,
params={"style": "analytical"}
)
if response.status_code == 200:
data = response.json()
print(f"\n✅ 처리 완료: {data['total_processed']}")
if data['success']:
print("\n성공한 기사:")
for item in data['success']:
print(f" - {item['keyword']}: {item['title'][:50]}...")
if data['errors']:
print("\n실패한 항목:")
for item in data['errors']:
print(f" - {item['keyword']}: {item['error']}")
else:
print(f"❌ 오류: {response.status_code}")
async def main():
"""메인 테스트 실행"""
print("\n" + "="*70)
print(" AI Writer Service Test Suite ")
print(" RSS → Google Search → Claude AI 기사 생성 ")
print("="*70)
# Run tests
await test_health_check()
await test_article_generation()
# await test_batch_generation() # Optional: batch test
print("\n" + "="*70)
print(" 테스트 완료 ")
print("="*70)
if __name__ == "__main__":
asyncio.run(main())

View File

@ -0,0 +1,240 @@
#!/usr/bin/env python3
"""
AI Writer Service - 프롬프트 기반 기사 생성 테스트
다양한 스타일과 키워드로 기사를 생성하는 테스트
"""
import asyncio
import httpx
import json
from datetime import datetime
# Service URL
SERVICE_URL = "http://localhost:8019"
async def test_different_styles():
"""다양한 스타일로 기사 생성 테스트"""
test_cases = [
{
"keyword": "전기차",
"style": "professional",
"description": "전통적인 뉴스 기사 스타일"
},
{
"keyword": "전기차",
"style": "analytical",
"description": "분석적이고 심층적인 스타일"
},
{
"keyword": "전기차",
"style": "investigative",
"description": "탐사보도 스타일"
}
]
async with httpx.AsyncClient(timeout=180.0) as client:
for test_case in test_cases:
print("\n" + "="*70)
print(f" {test_case['description']} 테스트")
print("="*70)
print(f"키워드: {test_case['keyword']}")
print(f"스타일: {test_case['style']}")
print("-" * 50)
try:
response = await client.post(
f"{SERVICE_URL}/api/generate",
json={
"keyword": test_case["keyword"],
"limit": 3, # RSS 항목 수 줄여서 빠른 테스트
"google_results_per_title": 2,
"lang": "ko",
"country": "KR",
"style": test_case["style"]
}
)
if response.status_code == 200:
article = response.json()
print(f"\n✅ 기사 생성 성공!")
print(f"📰 제목: {article['title']}")
print(f"📝 요약: {article['summary']}")
print(f"🔍 카테고리: {', '.join(article['categories'])}")
print(f"📚 소주제 수: {len(article['subtopics'])}")
# 키워드 출력
if 'entities' in article and 'keywords' in article['entities']:
keywords = article['entities']['keywords']
print(f"🔑 키워드 ({len(keywords)}개): {', '.join(keywords[:5])}" +
("..." if len(keywords) > 5 else ""))
# 이벤트 정보 출력
if 'entities' in article and 'events' in article['entities']:
events = article['entities']['events']
if events:
print(f"📅 이벤트 ({len(events)}개):")
for evt in events[:2]: # 처음 2개만 표시
if isinstance(evt, dict):
evt_str = f" - {evt.get('name', '')}"
if evt.get('date'):
evt_str += f" [{evt['date']}]"
if evt.get('location'):
evt_str += f" @{evt['location']}"
print(evt_str)
# 첫 번째 소주제의 첫 문단만 출력
if article['subtopics']:
first_topic = article['subtopics'][0]
print(f"\n첫 번째 소주제: {first_topic['title']}")
if first_topic['content']:
print(f"미리보기: {first_topic['content'][0][:200]}...")
# 파일로 저장
filename = f"article_{test_case['keyword']}_{test_case['style']}.json"
with open(filename, 'w', encoding='utf-8') as f:
json.dump(article, f, ensure_ascii=False, indent=2)
print(f"\n💾 '{filename}'에 저장됨")
else:
print(f"❌ 오류: {response.status_code}")
print(f"상세: {response.text}")
except Exception as e:
print(f"❌ 테스트 실패: {e}")
# 다음 테스트 전 잠시 대기
await asyncio.sleep(2)
async def test_different_keywords():
"""다양한 키워드로 기사 생성 테스트"""
keywords = ["블록체인", "메타버스", "우주개발", "기후변화", "K-POP"]
async with httpx.AsyncClient(timeout=180.0) as client:
print("\n" + "="*70)
print(" 다양한 키워드 테스트")
print("="*70)
for keyword in keywords:
print(f"\n🔍 키워드: {keyword}")
print("-" * 30)
try:
response = await client.post(
f"{SERVICE_URL}/api/generate",
json={
"keyword": keyword,
"limit": 2, # 빠른 테스트를 위해 줄임
"google_results_per_title": 2,
"lang": "ko",
"country": "KR",
"style": "professional"
}
)
if response.status_code == 200:
article = response.json()
print(f"✅ 성공: {article['title'][:50]}...")
print(f" 카테고리: {', '.join(article['categories'][:3])}")
else:
print(f"❌ 실패: {response.status_code}")
except Exception as e:
print(f"❌ 오류: {e}")
await asyncio.sleep(1)
async def test_custom_prompt():
"""커스텀 프롬프트 테스트 - 직접 aggregated 데이터 제공"""
# 미리 수집된 데이터를 시뮬레이션
custom_news_data = {
"keyword": "2025년 기술 트렌드",
"news_items": [
{
"rss_title": "AI와 로봇이 바꾸는 2025년 일상",
"google_results": [
{
"title": "전문가들이 예측하는 2025년 AI 혁명",
"snippet": "2025년 AI 기술이 일상생활 전반을 혁신할 전망...",
"full_content": {
"url": "https://example.com/ai-2025",
"content": "2025년에는 AI가 의료, 교육, 업무 등 모든 분야에서 인간과 협업하는 시대가 열릴 것으로 전망된다. 특히 생성형 AI의 발전으로 창의적 작업에서도 AI의 역할이 크게 확대될 것이다."
}
}
]
},
{
"rss_title": "양자컴퓨터 상용화 임박",
"google_results": [
{
"title": "IBM, 2025년 1000큐비트 양자컴퓨터 출시 예정",
"snippet": "IBM이 2025년 상용 양자컴퓨터 출시를 앞두고...",
"full_content": {
"url": "https://example.com/quantum-2025",
"content": "양자컴퓨팅이 드디어 실용화 단계에 접어들었다. 2025년에는 금융, 제약, 물류 등 다양한 산업에서 양자컴퓨터를 활용한 혁신이 시작될 전망이다."
}
}
]
}
]
}
async with httpx.AsyncClient(timeout=180.0) as client:
print("\n" + "="*70)
print(" 커스텀 데이터로 기사 생성")
print("="*70)
for style in ["professional", "analytical"]:
print(f"\n스타일: {style}")
print("-" * 30)
try:
response = await client.post(
f"{SERVICE_URL}/api/generate/from-aggregated",
json=custom_news_data,
params={"style": style}
)
if response.status_code == 200:
article = response.json()
print(f"✅ 제목: {article['title']}")
print(f" 요약: {article['summary']}")
# 스타일별로 저장
filename = f"custom_article_{style}.json"
with open(filename, 'w', encoding='utf-8') as f:
json.dump(article, f, ensure_ascii=False, indent=2)
print(f" 💾 '{filename}'에 저장됨")
else:
print(f"❌ 실패: {response.text}")
except Exception as e:
print(f"❌ 오류: {e}")
await asyncio.sleep(2)
async def main():
"""메인 테스트 실행"""
print("\n" + "="*70)
print(" AI Writer 프롬프트 기반 기사 생성 테스트")
print("="*70)
# 1. 다양한 스타일 테스트
print("\n[1] 스타일별 기사 생성 테스트")
await test_different_styles()
# 2. 다양한 키워드 테스트
print("\n[2] 키워드별 기사 생성 테스트")
await test_different_keywords()
# 3. 커스텀 데이터 테스트
print("\n[3] 커스텀 데이터 기사 생성 테스트")
await test_custom_prompt()
print("\n" + "="*70)
print(" 모든 테스트 완료!")
print("="*70)
if __name__ == "__main__":
asyncio.run(main())

View File

@ -0,0 +1,19 @@
FROM python:3.11-slim
WORKDIR /app
# Copy requirements
COPY backend/requirements.txt .
# Install dependencies
RUN pip install --no-cache-dir -r requirements.txt
# Copy application code
COPY backend/app /app
# Set environment variables
ENV PYTHONUNBUFFERED=1
ENV WORKER_COUNT=3
# Run worker
CMD ["python", "worker.py"]

View File

@ -0,0 +1,153 @@
# Google Search Service
키워드를 구글에서 검색한 결과를 수신하는 서비스입니다.
## 주요 기능
### 1. 다중 검색 방법 지원
- **Google Custom Search API**: 공식 구글 API (권장)
- **SerpAPI**: 대체 검색 API
- **웹 스크래핑**: 폴백 옵션 (제한적)
### 2. 검색 옵션
- 최대 20개 검색 결과 지원
- 언어별/국가별 검색
- 날짜 기준 필터링 및 정렬
- 전체 콘텐츠 가져오기
## API 엔드포인트
### 기본 검색
```
GET /api/search?q=키워드&num=20&lang=ko&country=kr
```
**파라미터:**
- `q`: 검색 키워드 (필수)
- `num`: 결과 개수 (1-20, 기본값: 10)
- `lang`: 언어 코드 (ko, en 등)
- `country`: 국가 코드 (kr, us 등)
- `date_restrict`: 날짜 제한
- `d7`: 일주일 이내
- `m1`: 한달 이내
- `m3`: 3개월 이내
- `y1`: 1년 이내
- `sort_by_date`: 최신순 정렬 (true/false)
### 전체 콘텐츠 검색
```
GET /api/search/full?q=키워드&num=5
```
각 검색 결과 페이지의 전체 내용을 가져옵니다 (시간이 오래 걸릴 수 있음).
### 실시간 트렌딩
```
GET /api/trending?country=kr
```
## 사용 예제
### 1. 한국어 검색 (최신순)
```bash
curl "http://localhost:8016/api/search?q=인공지능&num=20&lang=ko&country=kr&sort_by_date=true"
```
### 2. 영어 검색 (미국)
```bash
curl "http://localhost:8016/api/search?q=artificial%20intelligence&num=10&lang=en&country=us"
```
### 3. 최근 일주일 내 결과만
```bash
curl "http://localhost:8016/api/search?q=뉴스&date_restrict=d7&lang=ko"
```
### 4. 전체 콘텐츠 가져오기
```bash
curl "http://localhost:8016/api/search/full?q=python%20tutorial&num=3"
```
## 환경 설정
### 필수 API 키 설정
1. **Google Custom Search API**
- [Google Cloud Console](https://console.cloud.google.com/apis/credentials)에서 API 키 발급
- [Programmable Search Engine](https://programmablesearchengine.google.com/)에서 검색 엔진 ID 생성
2. **SerpAPI (선택사항)**
- [SerpAPI](https://serpapi.com/)에서 API 키 발급
### .env 파일 설정
```env
# Google Custom Search API
GOOGLE_API_KEY=your_api_key_here
GOOGLE_SEARCH_ENGINE_ID=your_search_engine_id_here
# SerpAPI (선택사항)
SERPAPI_KEY=your_serpapi_key_here
# Redis 캐시
REDIS_HOST=redis
REDIS_PORT=6379
REDIS_DB=2
# 기본 설정
DEFAULT_LANGUAGE=ko
DEFAULT_COUNTRY=kr
CACHE_TTL=3600
```
## Docker 실행
```bash
# 빌드 및 실행
docker-compose build google-search-backend
docker-compose up -d google-search-backend
# 로그 확인
docker-compose logs -f google-search-backend
```
## 제한 사항
### Google Custom Search API
- 무료 계정: 일일 100회 쿼리 제한
- 검색당 최대 100개 결과
- snippet 길이는 서버에서 제한 (변경 불가)
### 해결 방법
- 20개 이상 결과 필요 시: 페이지네이션 사용
- 긴 내용 필요 시: `/api/search/full` 엔드포인트 사용
- API 제한 도달 시: SerpAPI 또는 웹 스크래핑으로 자동 폴백
## 캐시 관리
Redis를 사용하여 검색 결과를 캐싱합니다:
- 기본 TTL: 3600초 (1시간)
- 캐시 초기화: `POST /api/clear-cache`
## 헬스 체크
```bash
curl http://localhost:8016/health
```
## 문제 해결
### 1. 한글 검색 안될 때
URL 인코딩 사용:
```bash
# "인공지능" → %EC%9D%B8%EA%B3%B5%EC%A7%80%EB%8A%A5
curl "http://localhost:8016/api/search?q=%EC%9D%B8%EA%B3%B5%EC%A7%80%EB%8A%A5"
```
### 2. API 제한 에러
- Google API 일일 제한 확인
- SerpAPI 키 설정으로 대체
- 웹 스크래핑 자동 폴백 활용
### 3. 느린 응답 시간
- Redis 캐시 활성화 확인
- 결과 개수 줄이기
- 전체 콘텐츠 대신 기본 검색 사용

View File

@ -0,0 +1,21 @@
# Google Custom Search API Configuration
# Get your API key from: https://console.cloud.google.com/apis/credentials
GOOGLE_API_KEY=
# Get your Search Engine ID from: https://programmablesearchengine.google.com/
GOOGLE_SEARCH_ENGINE_ID=
# Alternative: SerpAPI Configuration
# Get your API key from: https://serpapi.com/
SERPAPI_KEY=
# Redis Configuration
REDIS_HOST=redis
REDIS_PORT=6379
REDIS_DB=2
# Search Settings
DEFAULT_LANGUAGE=ko
DEFAULT_COUNTRY=kr
CACHE_TTL=3600
MAX_RESULTS=10

View File

@ -0,0 +1,10 @@
FROM python:3.11-slim
WORKDIR /app
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
COPY . .
CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "8000", "--reload"]

View File

@ -0,0 +1,30 @@
from pydantic_settings import BaseSettings
from typing import Optional
class Settings(BaseSettings):
# Google Custom Search API 설정
google_api_key: Optional[str] = None
google_search_engine_id: Optional[str] = None
# SerpAPI 설정 (대안)
serpapi_key: Optional[str] = None
# Redis 캐싱 설정
redis_host: str = "redis"
redis_port: int = 6379
redis_db: int = 2
cache_ttl: int = 3600 # 1시간
# 검색 설정
max_results: int = 10
default_language: str = "ko"
default_country: str = "kr"
# 서비스 설정
service_name: str = "Google Search Service"
debug: bool = True
class Config:
env_file = ".env"
settings = Settings()

View File

@ -0,0 +1,188 @@
from fastapi import FastAPI, Query, HTTPException
from fastapi.middleware.cors import CORSMiddleware
from typing import Optional
from datetime import datetime
from contextlib import asynccontextmanager
from .search_service import GoogleSearchService
from .config import settings
@asynccontextmanager
async def lifespan(app: FastAPI):
# 시작 시
print("Google Search Service starting...")
yield
# 종료 시
print("Google Search Service stopping...")
app = FastAPI(
title="Google Search Service",
description="구글 검색 결과를 수신하는 서비스",
version="1.0.0",
lifespan=lifespan
)
# CORS 설정
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
# 검색 서비스 초기화
search_service = GoogleSearchService()
@app.get("/")
async def root():
return {
"service": "Google Search Service",
"version": "1.0.0",
"timestamp": datetime.now().isoformat(),
"endpoints": {
"search": "/api/search?q=keyword",
"custom_search": "/api/search/custom?q=keyword",
"serpapi_search": "/api/search/serpapi?q=keyword",
"scraping_search": "/api/search/scraping?q=keyword",
"trending": "/api/trending",
"health": "/health"
}
}
@app.get("/health")
async def health_check():
return {
"status": "healthy",
"service": "google-search",
"timestamp": datetime.now().isoformat()
}
@app.get("/api/search")
async def search(
q: str = Query(..., description="검색 키워드"),
num: int = Query(10, description="결과 개수", ge=1, le=20),
lang: Optional[str] = Query(None, description="언어 코드 (ko, en 등)"),
country: Optional[str] = Query(None, description="국가 코드 (kr, us 등)"),
date_restrict: Optional[str] = Query(None, description="날짜 제한 (d7=일주일, m1=한달, m3=3개월, y1=1년)"),
sort_by_date: bool = Query(False, description="최신순 정렬")
):
"""
자동으로 최적의 방법을 선택하여 구글 검색
1. Google Custom Search API (설정된 경우)
2. SerpAPI (설정된 경우)
3. 웹 스크래핑 (폴백)
"""
# Google Custom Search API 시도
if settings.google_api_key and settings.google_search_engine_id:
result = await search_service.search_with_custom_api(q, num, lang, country, date_restrict, sort_by_date)
if "error" not in result or not result["error"]:
result["method"] = "google_custom_search"
return result
# SerpAPI 시도
if settings.serpapi_key:
result = await search_service.search_with_serpapi(q, num, lang, country)
if "error" not in result or not result["error"]:
result["method"] = "serpapi"
return result
# 웹 스크래핑 폴백
result = await search_service.search_with_scraping(q, num, lang)
result["method"] = "web_scraping"
result["warning"] = "API 키가 설정되지 않아 웹 스크래핑을 사용합니다. 제한적이고 불안정할 수 있습니다."
return result
@app.get("/api/search/custom")
async def search_custom(
q: str = Query(..., description="검색 키워드"),
num: int = Query(10, description="결과 개수", ge=1, le=10),
lang: Optional[str] = Query(None, description="언어 코드"),
country: Optional[str] = Query(None, description="국가 코드")
):
"""Google Custom Search API를 사용한 검색"""
if not settings.google_api_key or not settings.google_search_engine_id:
raise HTTPException(
status_code=503,
detail="Google Custom Search API credentials not configured"
)
result = await search_service.search_with_custom_api(q, num, lang, country)
if "error" in result and result["error"]:
raise HTTPException(status_code=500, detail=result["error"])
return result
@app.get("/api/search/serpapi")
async def search_serpapi(
q: str = Query(..., description="검색 키워드"),
num: int = Query(10, description="결과 개수", ge=1, le=50),
lang: Optional[str] = Query(None, description="언어 코드"),
country: Optional[str] = Query(None, description="국가 코드")
):
"""SerpAPI를 사용한 검색"""
if not settings.serpapi_key:
raise HTTPException(
status_code=503,
detail="SerpAPI key not configured"
)
result = await search_service.search_with_serpapi(q, num, lang, country)
if "error" in result and result["error"]:
raise HTTPException(status_code=500, detail=result["error"])
return result
@app.get("/api/search/scraping")
async def search_scraping(
q: str = Query(..., description="검색 키워드"),
num: int = Query(10, description="결과 개수", ge=1, le=20),
lang: Optional[str] = Query(None, description="언어 코드")
):
"""웹 스크래핑을 사용한 검색 (제한적)"""
result = await search_service.search_with_scraping(q, num, lang)
if "error" in result and result["error"]:
raise HTTPException(status_code=500, detail=result["error"])
result["warning"] = "웹 스크래핑은 제한적이고 불안정할 수 있습니다"
return result
@app.get("/api/search/full")
async def search_with_full_content(
q: str = Query(..., description="검색 키워드"),
num: int = Query(5, description="결과 개수", ge=1, le=10),
lang: Optional[str] = Query(None, description="언어 코드 (ko, en 등)"),
country: Optional[str] = Query(None, description="국가 코드 (kr, us 등)")
):
"""
Google 검색 후 각 결과 페이지의 전체 내용을 가져오기
주의: 시간이 오래 걸릴 수 있음
"""
result = await search_service.search_with_full_content(q, num, lang, country)
if "error" in result and result["error"]:
raise HTTPException(status_code=500, detail=result["error"])
return result
@app.get("/api/trending")
async def get_trending(
country: Optional[str] = Query(None, description="국가 코드 (kr, us 등)")
):
"""실시간 트렌딩 검색어 조회"""
result = await search_service.get_trending_searches(country)
if "error" in result and result["error"]:
raise HTTPException(status_code=500, detail=result["error"])
return result
@app.post("/api/clear-cache")
async def clear_cache():
"""캐시 초기화"""
try:
search_service.redis_client.flushdb()
return {
"status": "success",
"message": "캐시가 초기화되었습니다"
}
except Exception as e:
raise HTTPException(status_code=500, detail=str(e))

View File

@ -0,0 +1,540 @@
import httpx
import json
import redis
from typing import List, Dict, Optional
from datetime import datetime
import hashlib
from bs4 import BeautifulSoup
from .config import settings
class GoogleSearchService:
def __init__(self):
# Redis 연결
self.redis_client = redis.Redis(
host=settings.redis_host,
port=settings.redis_port,
db=settings.redis_db,
decode_responses=True
)
def _get_cache_key(self, query: str, **kwargs) -> str:
"""캐시 키 생성"""
cache_data = f"{query}_{kwargs}"
return f"google_search:{hashlib.md5(cache_data.encode()).hexdigest()}"
async def search_with_custom_api(
self,
query: str,
num_results: int = 10,
language: str = None,
country: str = None,
date_restrict: str = None,
sort_by_date: bool = False
) -> Dict:
"""Google Custom Search API 사용"""
if not settings.google_api_key or not settings.google_search_engine_id:
return {
"error": "Google API credentials not configured",
"results": []
}
# 캐시 확인
cache_key = self._get_cache_key(query, num=num_results, lang=language, country=country)
cached = self.redis_client.get(cache_key)
if cached:
return json.loads(cached)
url = "https://www.googleapis.com/customsearch/v1"
all_results = []
total_results_info = None
# Google API는 한 번에 최대 10개만 반환, 20개를 원하면 2번 요청
num_requests = min((num_results + 9) // 10, 2) # 최대 2번 요청 (20개까지)
async with httpx.AsyncClient() as client:
for page in range(num_requests):
start_index = page * 10 + 1
current_num = min(10, num_results - page * 10)
params = {
"key": settings.google_api_key,
"cx": settings.google_search_engine_id,
"q": query,
"num": current_num,
"start": start_index, # 시작 인덱스
"hl": language or settings.default_language,
"gl": country or settings.default_country
}
# 날짜 제한 추가 (d7 = 일주일, m1 = 한달, y1 = 1년)
if date_restrict:
params["dateRestrict"] = date_restrict
# 날짜순 정렬 (Google Custom Search API에서는 sort=date 옵션)
if sort_by_date:
params["sort"] = "date"
try:
response = await client.get(url, params=params)
response.raise_for_status()
data = response.json()
# 첫 번째 요청에서만 전체 정보 저장
if page == 0:
total_results_info = {
"total_results": data.get("searchInformation", {}).get("totalResults"),
"search_time": data.get("searchInformation", {}).get("searchTime"),
"query": data.get("queries", {}).get("request", [{}])[0].get("searchTerms")
}
# 결과 추가
for item in data.get("items", []):
all_results.append({
"title": item.get("title"),
"link": item.get("link"),
"snippet": item.get("snippet"),
"display_link": item.get("displayLink"),
"thumbnail": item.get("pagemap", {}).get("cse_thumbnail", [{}])[0].get("src") if "pagemap" in item else None
})
except Exception as e:
# 첫 번째 요청이 실패하면 에러 반환
if page == 0:
return {
"error": str(e),
"results": []
}
# 두 번째 요청이 실패하면 첫 번째 결과만 반환
break
results = {
"query": total_results_info.get("query") if total_results_info else query,
"total_results": total_results_info.get("total_results") if total_results_info else "0",
"search_time": total_results_info.get("search_time") if total_results_info else 0,
"results": all_results[:num_results], # 요청한 개수만큼만 반환
"timestamp": datetime.utcnow().isoformat()
}
# 캐시 저장
self.redis_client.setex(
cache_key,
settings.cache_ttl,
json.dumps(results)
)
return results
async def search_with_serpapi(
self,
query: str,
num_results: int = 10,
language: str = None,
country: str = None
) -> Dict:
"""SerpAPI 사용 (유료 서비스)"""
if not settings.serpapi_key:
return {
"error": "SerpAPI key not configured",
"results": []
}
# 캐시 확인
cache_key = self._get_cache_key(query, num=num_results, lang=language, country=country)
cached = self.redis_client.get(cache_key)
if cached:
return json.loads(cached)
from serpapi import GoogleSearch
params = {
"q": query,
"api_key": settings.serpapi_key,
"num": num_results,
"hl": language or settings.default_language,
"gl": country or settings.default_country
}
try:
search = GoogleSearch(params)
results = search.get_dict()
formatted_results = self._format_serpapi_results(results)
# 캐시 저장
self.redis_client.setex(
cache_key,
settings.cache_ttl,
json.dumps(formatted_results)
)
return formatted_results
except Exception as e:
return {
"error": str(e),
"results": []
}
async def search_with_scraping(
self,
query: str,
num_results: int = 10,
language: str = None
) -> Dict:
"""웹 스크래핑으로 검색 (비추천, 제한적)"""
# 캐시 확인
cache_key = self._get_cache_key(query, num=num_results, lang=language)
cached = self.redis_client.get(cache_key)
if cached:
return json.loads(cached)
headers = {
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36"
}
params = {
"q": query,
"num": num_results,
"hl": language or settings.default_language
}
async with httpx.AsyncClient() as client:
try:
response = await client.get(
"https://www.google.com/search",
params=params,
headers=headers,
follow_redirects=True
)
soup = BeautifulSoup(response.text, 'html.parser')
results = self._parse_google_html(soup)
formatted_results = {
"query": query,
"total_results": len(results),
"results": results,
"timestamp": datetime.utcnow().isoformat()
}
# 캐시 저장
self.redis_client.setex(
cache_key,
settings.cache_ttl,
json.dumps(formatted_results)
)
return formatted_results
except Exception as e:
return {
"error": str(e),
"results": []
}
def _format_google_results(self, data: Dict) -> Dict:
"""Google API 결과 포맷팅"""
results = []
for item in data.get("items", []):
results.append({
"title": item.get("title"),
"link": item.get("link"),
"snippet": item.get("snippet"),
"display_link": item.get("displayLink"),
"thumbnail": item.get("pagemap", {}).get("cse_thumbnail", [{}])[0].get("src") if "pagemap" in item else None
})
return {
"query": data.get("queries", {}).get("request", [{}])[0].get("searchTerms"),
"total_results": data.get("searchInformation", {}).get("totalResults"),
"search_time": data.get("searchInformation", {}).get("searchTime"),
"results": results,
"timestamp": datetime.utcnow().isoformat()
}
def _format_serpapi_results(self, data: Dict) -> Dict:
"""SerpAPI 결과 포맷팅"""
results = []
for item in data.get("organic_results", []):
results.append({
"title": item.get("title"),
"link": item.get("link"),
"snippet": item.get("snippet"),
"position": item.get("position"),
"thumbnail": item.get("thumbnail"),
"date": item.get("date")
})
# 관련 검색어
related_searches = [
item.get("query") for item in data.get("related_searches", [])
]
return {
"query": data.get("search_parameters", {}).get("q"),
"total_results": data.get("search_information", {}).get("total_results"),
"search_time": data.get("search_information", {}).get("time_taken_displayed"),
"results": results,
"related_searches": related_searches,
"timestamp": datetime.utcnow().isoformat()
}
def _parse_google_html(self, soup: BeautifulSoup) -> List[Dict]:
"""HTML 파싱으로 검색 결과 추출"""
results = []
# 검색 결과 컨테이너 찾기
for g in soup.find_all('div', class_='g'):
anchors = g.find_all('a')
if anchors:
link = anchors[0].get('href', '')
title_elem = g.find('h3')
snippet_elem = g.find('span', class_='st') or g.find('div', class_='s')
if title_elem and link:
results.append({
"title": title_elem.get_text(),
"link": link,
"snippet": snippet_elem.get_text() if snippet_elem else ""
})
return results
async def fetch_page_content(self, url: str) -> Dict:
"""웹 페이지의 전체 내용을 가져오기"""
try:
headers = {
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36"
}
async with httpx.AsyncClient(timeout=10.0) as client:
response = await client.get(url, headers=headers, follow_redirects=True)
response.raise_for_status()
soup = BeautifulSoup(response.text, 'html.parser')
# 불필요한 태그 제거
for script in soup(["script", "style", "nav", "header", "footer"]):
script.decompose()
# 본문 내용 추출 시도
main_content = None
# 1. article 태그 찾기
article = soup.find('article')
if article:
main_content = article.get_text()
# 2. main 태그 찾기
if not main_content:
main = soup.find('main')
if main:
main_content = main.get_text()
# 3. 일반적인 콘텐츠 div 찾기
if not main_content:
content_divs = soup.find_all('div', class_=lambda x: x and ('content' in x.lower() or 'article' in x.lower() or 'post' in x.lower()))
if content_divs:
main_content = ' '.join([div.get_text() for div in content_divs[:3]])
# 4. 전체 body에서 텍스트 추출
if not main_content:
body = soup.find('body')
if body:
main_content = body.get_text()
else:
main_content = soup.get_text()
# 텍스트 정리
main_content = ' '.join(main_content.split())
# 제목 추출
title = soup.find('title')
title_text = title.get_text() if title else ""
# 메타 설명 추출
meta_desc = soup.find('meta', attrs={'name': 'description'})
description = meta_desc.get('content', '') if meta_desc else ""
return {
"url": url,
"title": title_text,
"description": description,
"content": main_content[:5000], # 최대 5000자
"content_length": len(main_content),
"success": True
}
except Exception as e:
return {
"url": url,
"error": str(e),
"success": False
}
async def search_with_extended_snippet(
self,
query: str,
num_results: int = 10,
language: str = None,
country: str = None
) -> Dict:
"""검색 후 확장된 snippet 가져오기 (메타 설명 + 첫 500자)"""
# 먼저 일반 검색 수행
search_results = await self.search_with_custom_api(
query, num_results, language, country
)
if "error" in search_results:
return search_results
# 각 결과의 확장된 snippet 가져오기
import asyncio
async def fetch_extended_snippet(result):
"""개별 페이지의 확장된 snippet 가져오기"""
enhanced_result = result.copy()
if result.get("link"):
try:
headers = {
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36"
}
async with httpx.AsyncClient(timeout=5.0) as client:
response = await client.get(result["link"], headers=headers, follow_redirects=True)
response.raise_for_status()
soup = BeautifulSoup(response.text, 'html.parser')
# 메타 설명 추출
meta_desc = soup.find('meta', attrs={'name': 'description'})
if not meta_desc:
meta_desc = soup.find('meta', attrs={'property': 'og:description'})
description = meta_desc.get('content', '') if meta_desc else ""
# 본문 첫 부분 추출
for script in soup(["script", "style"]):
script.decompose()
# 본문 텍스트 찾기
text_content = ""
for tag in ['article', 'main', 'div']:
elements = soup.find_all(tag)
for elem in elements:
text = elem.get_text().strip()
if len(text) > 200: # 의미있는 텍스트만
text_content = ' '.join(text.split())[:1000]
break
if text_content:
break
# 기존 snippet과 병합
extended_snippet = result.get("snippet", "")
if description and description not in extended_snippet:
extended_snippet = description + " ... " + extended_snippet
if text_content and len(extended_snippet) < 500:
extended_snippet = extended_snippet + " ... " + text_content[:500-len(extended_snippet)]
enhanced_result["snippet"] = extended_snippet[:1000] # 최대 1000자
enhanced_result["extended"] = True
except Exception as e:
# 실패 시 원본 snippet 유지
enhanced_result["extended"] = False
enhanced_result["fetch_error"] = str(e)
return enhanced_result
# 병렬로 모든 페이지 처리
tasks = [fetch_extended_snippet(result) for result in search_results.get("results", [])]
enhanced_results = await asyncio.gather(*tasks)
return {
**search_results,
"results": enhanced_results,
"snippet_extended": True
}
async def search_with_full_content(
self,
query: str,
num_results: int = 5,
language: str = None,
country: str = None
) -> Dict:
"""검색 후 각 결과의 전체 내용 가져오기"""
# 먼저 일반 검색 수행
search_results = await self.search_with_custom_api(
query, num_results, language, country
)
if "error" in search_results:
return search_results
# 각 결과의 전체 내용 가져오기
enhanced_results = []
for result in search_results.get("results", [])[:num_results]:
# 원본 검색 결과 복사
enhanced_result = result.copy()
# 페이지 내용 가져오기
if result.get("link"):
content_data = await self.fetch_page_content(result["link"])
enhanced_result["full_content"] = content_data
enhanced_results.append(enhanced_result)
return {
**search_results,
"results": enhanced_results,
"content_fetched": True
}
async def get_trending_searches(self, country: str = None) -> Dict:
"""트렌딩 검색어 가져오기"""
# Google Trends 비공식 API 사용
url = f"https://trends.google.com/trends/api/dailytrends"
params = {
"geo": country or settings.default_country.upper()
}
async with httpx.AsyncClient() as client:
try:
response = await client.get(url, params=params)
# Google Trends API는 ")]}',\n"로 시작하는 응답을 반환
json_data = response.text[6:]
data = json.loads(json_data)
trending = []
for date_data in data.get("default", {}).get("trendingSearchesDays", []):
for search in date_data.get("trendingSearches", []):
trending.append({
"title": search.get("title", {}).get("query"),
"traffic": search.get("formattedTraffic"),
"articles": [
{
"title": article.get("title"),
"url": article.get("url"),
"source": article.get("source")
}
for article in search.get("articles", [])[:3]
]
})
return {
"country": country or settings.default_country,
"trending": trending[:10],
"timestamp": datetime.utcnow().isoformat()
}
except Exception as e:
return {
"error": str(e),
"trending": []
}

View File

@ -0,0 +1,9 @@
fastapi==0.109.0
uvicorn[standard]==0.27.0
httpx==0.26.0
pydantic==2.5.3
pydantic-settings==2.1.0
google-api-python-client==2.108.0
beautifulsoup4==4.12.2
redis==5.0.1
serpapi==0.1.5

View File

@ -0,0 +1,13 @@
FROM python:3.11-slim
WORKDIR /app
# Install dependencies
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
# Copy application
COPY . .
# Run the application
CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "8000", "--reload"]

View File

@ -0,0 +1,365 @@
"""
News Aggregator Service
RSS 피드 제목을 구글 검색으로 확장하는 통합 서비스
"""
from fastapi import FastAPI, HTTPException, Query, BackgroundTasks
from fastapi.middleware.cors import CORSMiddleware
from typing import List, Optional, Dict, Any
from datetime import datetime
import httpx
import asyncio
from pydantic import BaseModel
import logging
# Configure logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
app = FastAPI(
title="News Aggregator Service",
description="RSS 피드와 구글 검색을 통합한 뉴스 수집 서비스",
version="1.0.0"
)
# CORS 설정
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
# Configuration
RSS_SERVICE_URL = "http://rss-feed-backend:8000"
GOOGLE_SEARCH_SERVICE_URL = "http://google-search-backend:8000"
# Response Models
class NewsItem(BaseModel):
"""뉴스 항목"""
rss_title: str
rss_link: Optional[str] = None
google_results: List[Dict[str, Any]] = []
search_keyword: str
timestamp: datetime = None
class AggregatedNews(BaseModel):
"""통합 뉴스 결과"""
keyword: str
rss_feed_url: str
total_rss_entries: int
processed_entries: int
news_items: List[NewsItem]
processing_time: float
# HTTP Client
client = httpx.AsyncClient(timeout=30.0)
@app.on_event("startup")
async def startup():
"""서비스 시작"""
logger.info("News Aggregator Service starting...")
@app.on_event("shutdown")
async def shutdown():
"""서비스 종료"""
await client.aclose()
logger.info("News Aggregator Service stopped")
@app.get("/")
async def root():
return {
"service": "News Aggregator Service",
"version": "1.0.0",
"description": "RSS 피드와 구글 검색 통합 서비스",
"endpoints": {
"aggregate": "GET /api/aggregate",
"aggregate_by_location": "GET /api/aggregate/location",
"aggregate_by_topic": "GET /api/aggregate/topic",
"health": "GET /health"
}
}
@app.get("/health")
async def health_check():
"""헬스 체크"""
try:
# Check RSS service
rss_response = await client.get(f"{RSS_SERVICE_URL}/health")
rss_healthy = rss_response.status_code == 200
# Check Google Search service
google_response = await client.get(f"{GOOGLE_SEARCH_SERVICE_URL}/health")
google_healthy = google_response.status_code == 200
return {
"status": "healthy" if (rss_healthy and google_healthy) else "degraded",
"services": {
"rss_feed": "healthy" if rss_healthy else "unhealthy",
"google_search": "healthy" if google_healthy else "unhealthy"
},
"timestamp": datetime.now().isoformat()
}
except Exception as e:
return {
"status": "unhealthy",
"error": str(e),
"timestamp": datetime.now().isoformat()
}
@app.get("/api/aggregate", response_model=AggregatedNews)
async def aggregate_news(
q: str = Query(..., description="검색 키워드"),
limit: int = Query(10, description="처리할 RSS 항목 수", ge=1, le=50),
google_results_per_title: int = Query(5, description="각 제목당 구글 검색 결과 수", ge=1, le=10),
lang: str = Query("ko", description="언어 코드"),
country: str = Query("KR", description="국가 코드")
):
"""
키워드로 RSS 피드를 검색하고, 각 제목을 구글에서 재검색
1. 키워드로 Google News RSS 피드 가져오기
2. RSS 피드의 각 제목을 구글 검색
3. 통합 결과 반환
"""
start_time = datetime.now()
try:
# Step 1: Get RSS feed from keyword
logger.info(f"Fetching RSS feed for keyword: {q}")
rss_response = await client.get(
f"{RSS_SERVICE_URL}/api/google-rss/search",
params={"q": q, "lang": lang, "country": country}
)
rss_response.raise_for_status()
rss_data = rss_response.json()
if not rss_data.get("success"):
raise HTTPException(status_code=500, detail=f"RSS 피드 가져오기 실패: {rss_data.get('error')}")
# Step 2: Process each RSS entry with Google search
news_items = []
entries = rss_data.get("entries", [])
# If no entries field, fallback to sample_titles
if not entries:
titles = rss_data.get("sample_titles", [])[:limit]
entries = [{"title": title, "link": "", "published": ""} for title in titles]
else:
entries = entries[:limit]
# Create tasks for parallel processing
search_tasks = []
for entry in entries:
title = entry.get("title", "")
# Clean title for better search results
clean_title = title.split(" - ")[-1] if " - " in title else title
search_tasks.append(
search_google(clean_title, google_results_per_title, lang, country)
)
# Execute searches in parallel
logger.info(f"Searching Google for {len(search_tasks)} RSS entries")
search_results = await asyncio.gather(*search_tasks, return_exceptions=True)
# Combine results
for i, entry in enumerate(entries):
google_results = []
if not isinstance(search_results[i], Exception):
google_results = search_results[i]
title = entry.get("title", "")
news_items.append(NewsItem(
rss_title=title,
rss_link=entry.get("link", ""),
google_results=google_results,
search_keyword=title.split(" - ")[-1] if " - " in title else title,
timestamp=datetime.now()
))
# Calculate processing time
processing_time = (datetime.now() - start_time).total_seconds()
return AggregatedNews(
keyword=q,
rss_feed_url=rss_data.get("feed_url", ""),
total_rss_entries=rss_data.get("entry_count", 0),
processed_entries=len(news_items),
news_items=news_items,
processing_time=processing_time
)
except httpx.HTTPStatusError as e:
logger.error(f"HTTP error: {e}")
raise HTTPException(status_code=e.response.status_code, detail=str(e))
except Exception as e:
logger.error(f"Error in aggregate_news: {e}")
raise HTTPException(status_code=500, detail=str(e))
async def search_google(query: str, num_results: int, lang: str, country: str) -> List[Dict[str, Any]]:
"""구글 검색 서비스 호출 - 전체 콘텐츠 포함"""
try:
# Full content API 직접 호출
response = await client.get(
f"{GOOGLE_SEARCH_SERVICE_URL}/api/search/full",
params={
"q": query,
"num": num_results,
"lang": lang,
"country": country
}
)
response.raise_for_status()
data = response.json()
results = data.get("results", [])
# full_content가 이미 포함되어 있으므로 그대로 반환
logger.info(f"Google search for '{query}' returned {len(results)} results with full content")
return results
except Exception as e:
logger.error(f"Google search error for '{query}': {e}")
# Fallback to basic search without full content
try:
response = await client.get(
f"{GOOGLE_SEARCH_SERVICE_URL}/api/search",
params={
"q": query,
"num": num_results,
"lang": lang,
"country": country
}
)
response.raise_for_status()
data = response.json()
return data.get("results", [])
except:
return []
@app.get("/api/aggregate/location", response_model=AggregatedNews)
async def aggregate_news_by_location(
location: str = Query(..., description="지역명 (예: Seoul, Tokyo)"),
limit: int = Query(10, description="처리할 RSS 항목 수", ge=1, le=50),
google_results_per_title: int = Query(5, description="각 제목당 구글 검색 결과 수", ge=1, le=10),
lang: str = Query("ko", description="언어 코드"),
country: str = Query("KR", description="국가 코드")
):
"""지역 기반 RSS 피드를 가져와서 각 제목을 구글 검색"""
start_time = datetime.now()
try:
# Get location-based RSS feed
logger.info(f"Fetching RSS feed for location: {location}")
rss_response = await client.get(
f"{RSS_SERVICE_URL}/api/google-rss/location",
params={"location": location, "lang": lang, "country": country}
)
rss_response.raise_for_status()
rss_data = rss_response.json()
if not rss_data.get("success"):
raise HTTPException(status_code=500, detail=f"RSS 피드 가져오기 실패: {rss_data.get('error')}")
# Process titles
news_items = []
titles = rss_data.get("sample_titles", [])[:limit]
search_tasks = []
for title in titles:
clean_title = title.split(" - ")[-1] if " - " in title else title
search_tasks.append(
search_google(clean_title, google_results_per_title, lang, country)
)
search_results = await asyncio.gather(*search_tasks, return_exceptions=True)
for i, title in enumerate(titles):
google_results = []
if not isinstance(search_results[i], Exception):
google_results = search_results[i]
news_items.append(NewsItem(
rss_title=title,
google_results=google_results,
search_keyword=title.split(" - ")[-1] if " - " in title else title,
timestamp=datetime.now()
))
processing_time = (datetime.now() - start_time).total_seconds()
return AggregatedNews(
keyword=f"Location: {location}",
rss_feed_url=rss_data.get("feed_url", ""),
total_rss_entries=rss_data.get("entry_count", 0),
processed_entries=len(news_items),
news_items=news_items,
processing_time=processing_time
)
except Exception as e:
logger.error(f"Error in aggregate_news_by_location: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.get("/api/aggregate/topic", response_model=AggregatedNews)
async def aggregate_news_by_topic(
category: str = Query(..., description="카테고리 (TECHNOLOGY, BUSINESS, HEALTH 등)"),
limit: int = Query(10, description="처리할 RSS 항목 수", ge=1, le=50),
google_results_per_title: int = Query(5, description="각 제목당 구글 검색 결과 수", ge=1, le=10),
lang: str = Query("ko", description="언어 코드"),
country: str = Query("KR", description="국가 코드")
):
"""주제별 RSS 피드를 가져와서 각 제목을 구글 검색"""
start_time = datetime.now()
try:
# Get topic-based RSS feed
logger.info(f"Fetching RSS feed for topic: {category}")
rss_response = await client.get(
f"{RSS_SERVICE_URL}/api/google-rss/topic",
params={"category": category, "lang": lang, "country": country}
)
rss_response.raise_for_status()
rss_data = rss_response.json()
if not rss_data.get("success"):
raise HTTPException(status_code=500, detail=f"RSS 피드 가져오기 실패: {rss_data.get('error')}")
# Process titles
news_items = []
titles = rss_data.get("sample_titles", [])[:limit]
search_tasks = []
for title in titles:
clean_title = title.split(" - ")[-1] if " - " in title else title
search_tasks.append(
search_google(clean_title, google_results_per_title, lang, country)
)
search_results = await asyncio.gather(*search_tasks, return_exceptions=True)
for i, title in enumerate(titles):
google_results = []
if not isinstance(search_results[i], Exception):
google_results = search_results[i]
news_items.append(NewsItem(
rss_title=title,
google_results=google_results,
search_keyword=title.split(" - ")[-1] if " - " in title else title,
timestamp=datetime.now()
))
processing_time = (datetime.now() - start_time).total_seconds()
return AggregatedNews(
keyword=f"Topic: {category}",
rss_feed_url=rss_data.get("feed_url", ""),
total_rss_entries=rss_data.get("entry_count", 0),
processed_entries=len(news_items),
news_items=news_items,
processing_time=processing_time
)
except Exception as e:
logger.error(f"Error in aggregate_news_by_topic: {e}")
raise HTTPException(status_code=500, detail=str(e))

View File

@ -0,0 +1,5 @@
fastapi==0.104.1
uvicorn[standard]==0.24.0
httpx==0.25.2
pydantic==2.5.0
python-multipart==0.0.6

View File

@ -0,0 +1,214 @@
#!/usr/bin/env python3
"""
News Aggregator Service Test
RSS 피드 제목을 구글 full content 검색으로 확장하는 통합 테스트
"""
import asyncio
import httpx
import json
from datetime import datetime
from typing import Dict, Any
# Service URL
SERVICE_URL = "http://localhost:8018"
async def test_aggregate_with_full_content():
"""키워드로 RSS 피드를 검색하고 full content 구글 검색 테스트"""
async with httpx.AsyncClient(timeout=60.0) as client:
print("\n" + "="*60)
print("뉴스 통합 서비스 Full Content 테스트")
print("="*60)
# Test with keyword "인공지능"
print("\n1. 키워드 '인공지능'으로 RSS 피드 검색 및 구글 full content 검색")
print("-" * 40)
response = await client.get(
f"{SERVICE_URL}/api/aggregate",
params={
"q": "인공지능",
"limit": 3, # 테스트용으로 3개만
"google_results_per_title": 2, # 각 제목당 2개 구글 결과
"lang": "ko",
"country": "KR"
}
)
if response.status_code == 200:
data = response.json()
print(f"✓ RSS 피드 URL: {data['rss_feed_url']}")
print(f"✓ 전체 RSS 항목 수: {data['total_rss_entries']}")
print(f"✓ 처리된 항목 수: {data['processed_entries']}")
print(f"✓ 처리 시간: {data['processing_time']:.2f}")
# Check each news item for full content
for i, item in enumerate(data['news_items'], 1):
print(f"\n [{i}] RSS 제목: {item['rss_title'][:50]}...")
print(f" 검색 키워드: {item['search_keyword'][:50]}...")
print(f" 구글 검색 결과 수: {len(item['google_results'])}")
# Check if google results have full_content
for j, result in enumerate(item['google_results'], 1):
has_full_content = 'full_content' in result
if has_full_content:
full_content = result.get('full_content', '')
if isinstance(full_content, str):
content_length = len(full_content)
else:
content_length = len(str(full_content))
else:
content_length = 0
print(f" - 결과 {j}: {result.get('title', 'N/A')[:40]}...")
print(f" Full Content 포함: {'' if has_full_content else ''}")
if has_full_content:
print(f" Content 길이: {content_length:,} 문자")
# Show first 200 chars of content
if isinstance(result['full_content'], str):
preview = result['full_content'][:200].replace('\n', ' ')
print(f" 미리보기: {preview}...")
else:
print(f" Content 타입: {type(result['full_content'])}")
print(f" Content 데이터: {str(result['full_content'])[:200]}...")
else:
print(f"✗ 오류: {response.status_code}")
print(f" 상세: {response.text}")
async def test_aggregate_by_location():
"""지역 기반 RSS 피드 및 full content 테스트"""
async with httpx.AsyncClient(timeout=60.0) as client:
print("\n" + "="*60)
print("지역 기반 뉴스 통합 Full Content 테스트")
print("="*60)
print("\n2. 지역 'Seoul'로 RSS 피드 검색 및 구글 full content 검색")
print("-" * 40)
response = await client.get(
f"{SERVICE_URL}/api/aggregate/location",
params={
"location": "Seoul",
"limit": 2,
"google_results_per_title": 2,
"lang": "ko",
"country": "KR"
}
)
if response.status_code == 200:
data = response.json()
print(f"✓ 지역: {data['keyword']}")
print(f"✓ RSS 피드 URL: {data['rss_feed_url']}")
print(f"✓ 처리된 항목 수: {data['processed_entries']}")
# Check full content availability
full_content_count = 0
total_content_size = 0
for item in data['news_items']:
for result in item['google_results']:
if 'full_content' in result:
full_content_count += 1
content = result['full_content']
if isinstance(content, str):
total_content_size += len(content)
else:
total_content_size += len(str(content))
print(f"\n📊 Full Content 통계:")
print(f" - Full Content 포함 결과: {full_content_count}")
print(f" - 전체 Content 크기: {total_content_size:,} 문자")
print(f" - 평균 Content 크기: {total_content_size//max(full_content_count, 1):,} 문자")
else:
print(f"✗ 오류: {response.status_code}")
async def test_aggregate_by_topic():
"""주제별 RSS 피드 및 full content 테스트"""
async with httpx.AsyncClient(timeout=60.0) as client:
print("\n" + "="*60)
print("주제별 뉴스 통합 Full Content 테스트")
print("="*60)
print("\n3. 주제 'TECHNOLOGY'로 RSS 피드 검색 및 구글 full content 검색")
print("-" * 40)
response = await client.get(
f"{SERVICE_URL}/api/aggregate/topic",
params={
"category": "TECHNOLOGY",
"limit": 2,
"google_results_per_title": 3,
"lang": "ko",
"country": "KR"
}
)
if response.status_code == 200:
data = response.json()
print(f"✓ 주제: {data['keyword']}")
print(f"✓ 처리 시간: {data['processing_time']:.2f}")
# Analyze content quality for AI summarization
print("\n📝 AI 요약을 위한 Content 품질 분석:")
for i, item in enumerate(data['news_items'], 1):
print(f"\n 뉴스 항목 {i}:")
for j, result in enumerate(item['google_results'], 1):
if 'full_content' in result:
content = result['full_content']
if isinstance(content, str):
# Check content quality indicators
has_paragraphs = '\n\n' in content or '</p>' in content
has_sufficient_length = len(content) > 500
has_korean = any(ord(char) >= 0xAC00 and ord(char) <= 0xD7A3 for char in content[:min(100, len(content))])
else:
content_str = str(content)
has_paragraphs = '\n\n' in content_str or '</p>' in content_str
has_sufficient_length = len(content_str) > 500
has_korean = any(ord(char) >= 0xAC00 and ord(char) <= 0xD7A3 for char in content_str[:min(100, len(content_str))])
print(f" 결과 {j} 품질 체크:")
print(f" - 충분한 길이 (>500자): {'' if has_sufficient_length else ''}")
print(f" - 단락 구조 포함: {'' if has_paragraphs else ''}")
print(f" - 한국어 콘텐츠: {'' if has_korean else ''}")
print(f" - AI 요약 가능: {'' if (has_sufficient_length and has_paragraphs) else ''}")
else:
print(f"✗ 오류: {response.status_code}")
async def test_health_check():
"""서비스 상태 확인"""
async with httpx.AsyncClient() as client:
print("\n" + "="*60)
print("서비스 Health Check")
print("="*60)
response = await client.get(f"{SERVICE_URL}/health")
if response.status_code == 200:
data = response.json()
print(f"✓ 통합 서비스 상태: {data['status']}")
print(f" - RSS 서비스: {data['services']['rss_feed']}")
print(f" - Google 검색 서비스: {data['services']['google_search']}")
else:
print(f"✗ Health check 실패: {response.status_code}")
async def main():
"""메인 테스트 실행"""
print("\n" + "="*70)
print(" News Aggregator Full Content Integration Test ")
print(" RSS 피드 + Google Full Content 통합 테스트 ")
print("="*70)
# Run tests
await test_health_check()
await test_aggregate_with_full_content()
await test_aggregate_by_location()
await test_aggregate_by_topic()
print("\n" + "="*70)
print(" 테스트 완료 - Full Content 통합 확인 ")
print("="*70)
print("\n✅ 모든 테스트가 완료되었습니다.")
print(" RSS 피드 제목을 구글 full content로 검색하는 기능이 정상 작동합니다.")
print(" AI 요약을 위한 충분한 콘텐츠가 수집되고 있습니다.")
if __name__ == "__main__":
asyncio.run(main())

View File

@ -0,0 +1,204 @@
# RSS Feed Subscription Service
RSS/Atom 피드를 구독하고 관리하는 서비스입니다.
## 주요 기능
### 1. 피드 구독 관리
- RSS/Atom 피드 URL 구독
- 카테고리별 분류 (뉴스, 기술, 비즈니스 등)
- 자동 업데이트 스케줄링
- 피드 상태 모니터링
### 2. 엔트리 관리
- 새로운 글 자동 수집
- 읽음/안읽음 상태 관리
- 별표 표시 기능
- 전체 내용 저장
### 3. 자동 업데이트
- 설정 가능한 업데이트 주기 (기본 15분)
- 백그라운드 스케줄러
- 에러 처리 및 재시도
## API 엔드포인트
### 피드 구독
```
POST /api/feeds
{
"url": "https://example.com/rss",
"title": "Example Blog",
"category": "tech",
"update_interval": 900
}
```
### 피드 목록 조회
```
GET /api/feeds?category=tech&status=active
```
### 엔트리 조회
```
GET /api/entries?feed_id=xxx&is_read=false&limit=50
```
### 읽음 표시
```
PUT /api/entries/{entry_id}/read?is_read=true
```
### 별표 표시
```
PUT /api/entries/{entry_id}/star?is_starred=true
```
### 통계 조회
```
GET /api/stats?feed_id=xxx
```
### OPML 내보내기
```
GET /api/export/opml
```
## 사용 예제
### 1. 기술 블로그 구독
```bash
curl -X POST http://localhost:8017/api/feeds \
-H "Content-Type: application/json" \
-d '{
"url": "https://techcrunch.com/feed/",
"category": "tech"
}'
```
### 2. 한국 뉴스 RSS 구독
```bash
curl -X POST http://localhost:8017/api/feeds \
-H "Content-Type: application/json" \
-d '{
"url": "https://www.hani.co.kr/rss/",
"category": "news",
"update_interval": 600
}'
```
### 3. 안읽은 엔트리 조회
```bash
curl "http://localhost:8017/api/entries?is_read=false&limit=20"
```
### 4. 모든 엔트리 읽음 처리
```bash
curl -X POST "http://localhost:8017/api/entries/mark-all-read?feed_id=xxx"
```
## 지원 카테고리
- `news`: 뉴스
- `tech`: 기술
- `business`: 비즈니스
- `science`: 과학
- `health`: 건강
- `sports`: 스포츠
- `entertainment`: 엔터테인먼트
- `lifestyle`: 라이프스타일
- `politics`: 정치
- `other`: 기타
## 환경 설정
### 필수 설정
```env
MONGODB_URL=mongodb://mongodb:27017
DB_NAME=rss_feed_db
REDIS_URL=redis://redis:6379
REDIS_DB=3
```
### 선택 설정
```env
DEFAULT_UPDATE_INTERVAL=900 # 기본 업데이트 주기 (초)
MAX_ENTRIES_PER_FEED=100 # 피드당 최대 엔트리 수
ENABLE_SCHEDULER=true # 자동 업데이트 활성화
SCHEDULER_TIMEZONE=Asia/Seoul # 스케줄러 타임존
```
## Docker 실행
```bash
# 빌드 및 실행
docker-compose build rss-feed-backend
docker-compose up -d rss-feed-backend
# 로그 확인
docker-compose logs -f rss-feed-backend
```
## 데이터 구조
### FeedSubscription
- `title`: 피드 제목
- `url`: RSS/Atom URL
- `description`: 설명
- `category`: 카테고리
- `status`: 상태 (active/inactive/error)
- `update_interval`: 업데이트 주기
- `last_fetch`: 마지막 업데이트 시간
- `error_count`: 에러 횟수
### FeedEntry
- `feed_id`: 피드 ID
- `title`: 글 제목
- `link`: 원문 링크
- `summary`: 요약
- `content`: 전체 내용
- `author`: 작성자
- `published`: 발행일
- `categories`: 태그/카테고리
- `thumbnail`: 썸네일 이미지
- `is_read`: 읽음 상태
- `is_starred`: 별표 상태
## 추천 RSS 피드
### 한국 뉴스
- 한겨레: `https://www.hani.co.kr/rss/`
- 조선일보: `https://www.chosun.com/arc/outboundfeeds/rss/`
- 중앙일보: `https://rss.joins.com/joins_news_list.xml`
### 기술 블로그
- TechCrunch: `https://techcrunch.com/feed/`
- The Verge: `https://www.theverge.com/rss/index.xml`
- Ars Technica: `https://feeds.arstechnica.com/arstechnica/index`
### 개발자 블로그
- GitHub Blog: `https://github.blog/feed/`
- Stack Overflow Blog: `https://stackoverflow.blog/feed/`
- Dev.to: `https://dev.to/feed`
## 헬스 체크
```bash
curl http://localhost:8017/health
```
## 문제 해결
### 1. 피드 파싱 실패
- RSS/Atom 형식이 올바른지 확인
- URL이 접근 가능한지 확인
- 피드 인코딩 확인 (UTF-8 권장)
### 2. 업데이트 안됨
- 스케줄러 활성화 확인 (`ENABLE_SCHEDULER=true`)
- MongoDB 연결 상태 확인
- 피드 상태가 `active`인지 확인
### 3. 중복 엔트리
- 피드에서 고유 ID를 제공하는지 확인
- 엔트리 ID 생성 로직 확인

View File

@ -0,0 +1,21 @@
FROM python:3.11-slim
WORKDIR /app
# Install system dependencies
RUN apt-get update && apt-get install -y \
gcc \
&& rm -rf /var/lib/apt/lists/*
# Copy requirements and install Python dependencies
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
# Copy application code
COPY . .
# Expose port
EXPOSE 8000
# Run the application
CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "8000", "--reload"]

View File

@ -0,0 +1,26 @@
from pydantic_settings import BaseSettings
from typing import Optional
class Settings(BaseSettings):
# MongoDB Configuration
mongodb_url: str = "mongodb://mongodb:27017"
db_name: str = "rss_feed_db"
# Redis Configuration
redis_url: str = "redis://redis:6379"
redis_db: int = 3
# Feed Settings
default_update_interval: int = 900 # 15 minutes in seconds
max_entries_per_feed: int = 100
fetch_timeout: int = 30
# Scheduler Settings
enable_scheduler: bool = True
scheduler_timezone: str = "Asia/Seoul"
class Config:
env_file = ".env"
env_file_encoding = "utf-8"
settings = Settings()

View File

@ -0,0 +1,222 @@
import feedparser
import httpx
from typing import List, Dict, Any, Optional
from datetime import datetime
from dateutil import parser as date_parser
from bs4 import BeautifulSoup
import re
import hashlib
from .models import FeedEntry
class FeedParser:
def __init__(self):
self.client = httpx.AsyncClient(
timeout=30.0,
follow_redirects=True,
headers={
"User-Agent": "Mozilla/5.0 (compatible; RSS Feed Reader/1.0)"
}
)
async def parse_feed(self, url: str) -> Dict[str, Any]:
"""Parse RSS/Atom feed from URL"""
try:
response = await self.client.get(url)
response.raise_for_status()
# Parse the feed
feed = feedparser.parse(response.content)
if feed.bozo and feed.bozo_exception:
raise Exception(f"Feed parsing error: {feed.bozo_exception}")
return {
"success": True,
"feed": feed.feed,
"entries": feed.entries,
"error": None
}
except Exception as e:
return {
"success": False,
"feed": None,
"entries": [],
"error": str(e)
}
def extract_entry_data(self, entry: Any, feed_id: str) -> FeedEntry:
"""Extract and normalize entry data"""
# Generate unique entry ID
entry_id = self._generate_entry_id(entry)
# Extract title
title = entry.get("title", "Untitled")
# Extract link
link = entry.get("link", "")
# Extract summary/description
summary = self._extract_summary(entry)
# Extract content
content = self._extract_content(entry)
# Extract author
author = entry.get("author", "")
# Extract published date
published = self._parse_date(entry.get("published", entry.get("updated")))
# Extract updated date
updated = self._parse_date(entry.get("updated", entry.get("published")))
# Extract categories
categories = self._extract_categories(entry)
# Extract thumbnail
thumbnail = self._extract_thumbnail(entry)
# Extract enclosures (media attachments)
enclosures = self._extract_enclosures(entry)
return FeedEntry(
feed_id=feed_id,
entry_id=entry_id,
title=title,
link=link,
summary=summary,
content=content,
author=author,
published=published,
updated=updated,
categories=categories,
thumbnail=thumbnail,
enclosures=enclosures
)
def _generate_entry_id(self, entry: Any) -> str:
"""Generate unique ID for entry"""
# Try to use entry's unique ID first
if hasattr(entry, "id"):
return entry.id
# Generate from link and title
unique_str = f"{entry.get('link', '')}{entry.get('title', '')}"
return hashlib.md5(unique_str.encode()).hexdigest()
def _extract_summary(self, entry: Any) -> Optional[str]:
"""Extract and clean summary"""
summary = entry.get("summary", entry.get("description", ""))
if summary:
# Clean HTML tags
soup = BeautifulSoup(summary, "html.parser")
text = soup.get_text(separator=" ", strip=True)
# Limit length
if len(text) > 500:
text = text[:497] + "..."
return text
return None
def _extract_content(self, entry: Any) -> Optional[str]:
"""Extract full content"""
content = ""
# Try content field
if hasattr(entry, "content"):
for c in entry.content:
if c.get("type") in ["text/html", "text/plain"]:
content = c.get("value", "")
break
# Fallback to summary detail
if not content and hasattr(entry, "summary_detail"):
content = entry.summary_detail.get("value", "")
# Clean excessive whitespace
if content:
content = re.sub(r'\s+', ' ', content).strip()
return content
return None
def _parse_date(self, date_str: Optional[str]) -> Optional[datetime]:
"""Parse date string to datetime"""
if not date_str:
return None
try:
# Try parsing with dateutil
return date_parser.parse(date_str)
except:
try:
# Try feedparser's time structure
if hasattr(date_str, "tm_year"):
import time
return datetime.fromtimestamp(time.mktime(date_str))
except:
pass
return None
def _extract_categories(self, entry: Any) -> List[str]:
"""Extract categories/tags"""
categories = []
if hasattr(entry, "tags"):
for tag in entry.tags:
if hasattr(tag, "term"):
categories.append(tag.term)
elif isinstance(tag, str):
categories.append(tag)
return categories
def _extract_thumbnail(self, entry: Any) -> Optional[str]:
"""Extract thumbnail image URL"""
# Check media thumbnail
if hasattr(entry, "media_thumbnail"):
for thumb in entry.media_thumbnail:
if thumb.get("url"):
return thumb["url"]
# Check media content
if hasattr(entry, "media_content"):
for media in entry.media_content:
if media.get("type", "").startswith("image/"):
return media.get("url")
# Check enclosures
if hasattr(entry, "enclosures"):
for enc in entry.enclosures:
if enc.get("type", "").startswith("image/"):
return enc.get("href", enc.get("url"))
# Extract from content/summary
content = entry.get("summary", "") + entry.get("content", [{}])[0].get("value", "") if hasattr(entry, "content") else ""
if content:
soup = BeautifulSoup(content, "html.parser")
img = soup.find("img")
if img and img.get("src"):
return img["src"]
return None
def _extract_enclosures(self, entry: Any) -> List[Dict[str, Any]]:
"""Extract media enclosures"""
enclosures = []
if hasattr(entry, "enclosures"):
for enc in entry.enclosures:
enclosure = {
"url": enc.get("href", enc.get("url", "")),
"type": enc.get("type", ""),
"length": enc.get("length", 0)
}
if enclosure["url"]:
enclosures.append(enclosure)
return enclosures
async def close(self):
"""Close HTTP client"""
await self.client.aclose()

View File

@ -0,0 +1,115 @@
"""
Google News RSS Feed Generator
구글 뉴스 RSS 피드 URL 생성 및 구독 지원
"""
from typing import Optional, List
from urllib.parse import quote_plus
from enum import Enum
class GoogleNewsCategory(str, Enum):
"""구글 뉴스 카테고리"""
WORLD = "WORLD"
NATION = "NATION"
BUSINESS = "BUSINESS"
TECHNOLOGY = "TECHNOLOGY"
ENTERTAINMENT = "ENTERTAINMENT"
SPORTS = "SPORTS"
SCIENCE = "SCIENCE"
HEALTH = "HEALTH"
class GoogleNewsRSS:
"""Google News RSS 피드 URL 생성기"""
BASE_URL = "https://news.google.com/rss"
@staticmethod
def search_feed(query: str, lang: str = "ko", country: str = "KR") -> str:
"""
키워드 검색 RSS 피드 URL 생성
Args:
query: 검색 키워드
lang: 언어 코드 (ko, en, ja, zh-CN 등)
country: 국가 코드 (KR, US, JP, CN 등)
Returns:
RSS 피드 URL
"""
encoded_query = quote_plus(query)
return f"{GoogleNewsRSS.BASE_URL}/search?q={encoded_query}&hl={lang}&gl={country}&ceid={country}:{lang}"
@staticmethod
def topic_feed(category: GoogleNewsCategory, lang: str = "ko", country: str = "KR") -> str:
"""
카테고리별 RSS 피드 URL 생성
Args:
category: 뉴스 카테고리
lang: 언어 코드
country: 국가 코드
Returns:
RSS 피드 URL
"""
return f"{GoogleNewsRSS.BASE_URL}/headlines/section/topic/{category.value}?hl={lang}&gl={country}&ceid={country}:{lang}"
@staticmethod
def location_feed(location: str, lang: str = "ko", country: str = "KR") -> str:
"""
지역 뉴스 RSS 피드 URL 생성
Args:
location: 지역명 (예: Seoul, 서울, New York)
lang: 언어 코드
country: 국가 코드
Returns:
RSS 피드 URL
"""
encoded_location = quote_plus(location)
return f"{GoogleNewsRSS.BASE_URL}/headlines/section/geo/{encoded_location}?hl={lang}&gl={country}&ceid={country}:{lang}"
@staticmethod
def trending_feed(lang: str = "ko", country: str = "KR") -> str:
"""
트렌딩 뉴스 RSS 피드 URL 생성
Args:
lang: 언어 코드
country: 국가 코드
Returns:
RSS 피드 URL
"""
return f"{GoogleNewsRSS.BASE_URL}?hl={lang}&gl={country}&ceid={country}:{lang}"
@staticmethod
def get_common_feeds() -> List[dict]:
"""
자주 사용되는 RSS 피드 목록 반환
Returns:
피드 정보 리스트
"""
return [
{
"title": "구글 뉴스 - 한국 헤드라인",
"url": GoogleNewsRSS.trending_feed("ko", "KR"),
"description": "한국 주요 뉴스"
},
{
"title": "구글 뉴스 - 기술",
"url": GoogleNewsRSS.topic_feed(GoogleNewsCategory.TECHNOLOGY, "ko", "KR"),
"description": "기술 관련 뉴스"
},
{
"title": "구글 뉴스 - 비즈니스",
"url": GoogleNewsRSS.topic_feed(GoogleNewsCategory.BUSINESS, "ko", "KR"),
"description": "비즈니스 뉴스"
},
{
"title": "Google News - World",
"url": GoogleNewsRSS.topic_feed(GoogleNewsCategory.WORLD, "en", "US"),
"description": "World news in English"
}
]

View File

@ -0,0 +1,596 @@
from fastapi import FastAPI, HTTPException, Query, Path, BackgroundTasks
from fastapi.middleware.cors import CORSMiddleware
from typing import List, Optional
from datetime import datetime
from contextlib import asynccontextmanager
import motor.motor_asyncio
from apscheduler.schedulers.asyncio import AsyncIOScheduler
from apscheduler.triggers.interval import IntervalTrigger
import pytz
import redis.asyncio as redis
import json
from .config import settings
from .models import (
FeedSubscription, FeedEntry, CreateFeedRequest,
UpdateFeedRequest, FeedStatistics, FeedStatus, FeedCategory
)
from .feed_parser import FeedParser
from .google_rss import GoogleNewsRSS, GoogleNewsCategory
# Database connection
db_client = None
db = None
redis_client = None
scheduler = None
parser = None
@asynccontextmanager
async def lifespan(app: FastAPI):
global db_client, db, redis_client, scheduler, parser
# Connect to MongoDB
db_client = motor.motor_asyncio.AsyncIOMotorClient(settings.mongodb_url)
db = db_client[settings.db_name]
# Connect to Redis
redis_client = redis.from_url(settings.redis_url, db=settings.redis_db)
# Initialize feed parser
parser = FeedParser()
# Initialize scheduler
if settings.enable_scheduler:
scheduler = AsyncIOScheduler(timezone=pytz.timezone(settings.scheduler_timezone))
scheduler.add_job(
update_all_feeds,
trigger=IntervalTrigger(seconds=60),
id="update_feeds",
replace_existing=True
)
scheduler.start()
print("RSS Feed scheduler started")
print("RSS Feed Service starting...")
yield
# Cleanup
if scheduler:
scheduler.shutdown()
if parser:
await parser.close()
if redis_client:
await redis_client.close()
db_client.close()
print("RSS Feed Service stopping...")
app = FastAPI(
title="RSS Feed Service",
description="RSS/Atom 피드 구독 및 관리 서비스",
version="1.0.0",
lifespan=lifespan
)
# CORS 설정
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
# Helper functions
async def update_feed(feed_id: str):
"""Update a single feed"""
feed = await db.feeds.find_one({"_id": feed_id})
if not feed:
return
# Parse feed
result = await parser.parse_feed(feed["url"])
if result["success"]:
# Update feed metadata
await db.feeds.update_one(
{"_id": feed_id},
{
"$set": {
"last_fetch": datetime.now(),
"status": FeedStatus.ACTIVE,
"error_count": 0,
"last_error": None,
"updated_at": datetime.now()
}
}
)
# Process entries
for entry_data in result["entries"][:settings.max_entries_per_feed]:
entry = parser.extract_entry_data(entry_data, feed_id)
# Check if entry already exists
existing = await db.entries.find_one({
"feed_id": feed_id,
"entry_id": entry.entry_id
})
if not existing:
# Insert new entry
await db.entries.insert_one(entry.dict())
else:
# Update existing entry if newer
if entry.updated and existing.get("updated"):
if entry.updated > existing["updated"]:
await db.entries.update_one(
{"_id": existing["_id"]},
{"$set": entry.dict(exclude={"id", "created_at"})}
)
else:
# Update error status
await db.feeds.update_one(
{"_id": feed_id},
{
"$set": {
"status": FeedStatus.ERROR,
"last_error": result["error"],
"updated_at": datetime.now()
},
"$inc": {"error_count": 1}
}
)
async def update_all_feeds():
"""Update all active feeds that need updating"""
now = datetime.now()
# Find feeds that need updating
feeds = await db.feeds.find({
"status": FeedStatus.ACTIVE,
"$or": [
{"last_fetch": None},
{"last_fetch": {"$lt": now}}
]
}).to_list(100)
for feed in feeds:
# Check if it's time to update
if feed.get("last_fetch"):
time_diff = (now - feed["last_fetch"]).total_seconds()
if time_diff < feed.get("update_interval", settings.default_update_interval):
continue
# Update feed in background
await update_feed(str(feed["_id"]))
# API Endpoints
@app.get("/")
async def root():
return {
"service": "RSS Feed Service",
"version": "1.0.0",
"timestamp": datetime.now().isoformat(),
"endpoints": {
"subscribe": "POST /api/feeds",
"list_feeds": "GET /api/feeds",
"get_entries": "GET /api/entries",
"mark_read": "PUT /api/entries/{entry_id}/read",
"mark_starred": "PUT /api/entries/{entry_id}/star",
"statistics": "GET /api/stats"
}
}
@app.get("/health")
async def health_check():
return {
"status": "healthy",
"service": "rss-feed",
"timestamp": datetime.now().isoformat()
}
@app.post("/api/feeds", response_model=FeedSubscription)
async def subscribe_to_feed(request: CreateFeedRequest, background_tasks: BackgroundTasks):
"""RSS/Atom 피드 구독"""
# Check if already subscribed
existing = await db.feeds.find_one({"url": str(request.url)})
if existing:
raise HTTPException(status_code=400, detail="이미 구독 중인 피드입니다")
# Parse feed to get metadata
result = await parser.parse_feed(str(request.url))
if not result["success"]:
raise HTTPException(status_code=400, detail=f"피드 파싱 실패: {result['error']}")
# Create subscription
feed = FeedSubscription(
title=request.title or result["feed"].get("title", "Untitled Feed"),
url=request.url,
description=result["feed"].get("description", ""),
category=request.category,
update_interval=request.update_interval or settings.default_update_interval
)
# Save to database - convert URL to string
feed_dict = feed.dict()
feed_dict["url"] = str(feed_dict["url"])
result = await db.feeds.insert_one(feed_dict)
feed.id = str(result.inserted_id)
# Fetch entries in background
background_tasks.add_task(update_feed, feed.id)
return feed
@app.get("/api/feeds", response_model=List[FeedSubscription])
async def list_feeds(
category: Optional[str] = Query(None, description="카테고리 필터"),
status: Optional[FeedStatus] = Query(None, description="상태 필터")
):
"""구독 중인 피드 목록 조회"""
query = {}
if category:
query["category"] = category
if status:
query["status"] = status
feeds = await db.feeds.find(query).to_list(100)
for feed in feeds:
feed["_id"] = str(feed["_id"])
return feeds
@app.get("/api/feeds/{feed_id}", response_model=FeedSubscription)
async def get_feed(feed_id: str = Path(..., description="피드 ID")):
"""특정 피드 정보 조회"""
feed = await db.feeds.find_one({"_id": feed_id})
if not feed:
raise HTTPException(status_code=404, detail="피드를 찾을 수 없습니다")
feed["_id"] = str(feed["_id"])
return feed
@app.put("/api/feeds/{feed_id}", response_model=FeedSubscription)
async def update_feed_subscription(
feed_id: str = Path(..., description="피드 ID"),
request: UpdateFeedRequest = ...
):
"""피드 구독 정보 수정"""
update_data = request.dict(exclude_unset=True)
if update_data:
update_data["updated_at"] = datetime.now()
result = await db.feeds.update_one(
{"_id": feed_id},
{"$set": update_data}
)
if result.matched_count == 0:
raise HTTPException(status_code=404, detail="피드를 찾을 수 없습니다")
feed = await db.feeds.find_one({"_id": feed_id})
feed["_id"] = str(feed["_id"])
return feed
@app.delete("/api/feeds/{feed_id}")
async def unsubscribe_from_feed(feed_id: str = Path(..., description="피드 ID")):
"""피드 구독 취소"""
# Delete feed
result = await db.feeds.delete_one({"_id": feed_id})
if result.deleted_count == 0:
raise HTTPException(status_code=404, detail="피드를 찾을 수 없습니다")
# Delete associated entries
await db.entries.delete_many({"feed_id": feed_id})
return {"message": "구독이 취소되었습니다"}
@app.post("/api/feeds/{feed_id}/refresh")
async def refresh_feed(
feed_id: str = Path(..., description="피드 ID"),
background_tasks: BackgroundTasks = ...
):
"""피드 수동 새로고침"""
feed = await db.feeds.find_one({"_id": feed_id})
if not feed:
raise HTTPException(status_code=404, detail="피드를 찾을 수 없습니다")
background_tasks.add_task(update_feed, feed_id)
return {"message": "피드 새로고침이 시작되었습니다"}
@app.get("/api/entries", response_model=List[FeedEntry])
async def get_entries(
feed_id: Optional[str] = Query(None, description="피드 ID"),
is_read: Optional[bool] = Query(None, description="읽음 상태 필터"),
is_starred: Optional[bool] = Query(None, description="별표 상태 필터"),
limit: int = Query(50, ge=1, le=100, description="결과 개수"),
offset: int = Query(0, ge=0, description="오프셋")
):
"""피드 엔트리 목록 조회"""
query = {}
if feed_id:
query["feed_id"] = feed_id
if is_read is not None:
query["is_read"] = is_read
if is_starred is not None:
query["is_starred"] = is_starred
entries = await db.entries.find(query) \
.sort("published", -1) \
.skip(offset) \
.limit(limit) \
.to_list(limit)
for entry in entries:
entry["_id"] = str(entry["_id"])
return entries
@app.get("/api/entries/{entry_id}", response_model=FeedEntry)
async def get_entry(entry_id: str = Path(..., description="엔트리 ID")):
"""특정 엔트리 조회"""
entry = await db.entries.find_one({"_id": entry_id})
if not entry:
raise HTTPException(status_code=404, detail="엔트리를 찾을 수 없습니다")
entry["_id"] = str(entry["_id"])
return entry
@app.put("/api/entries/{entry_id}/read")
async def mark_entry_as_read(
entry_id: str = Path(..., description="엔트리 ID"),
is_read: bool = Query(True, description="읽음 상태")
):
"""엔트리 읽음 상태 변경"""
result = await db.entries.update_one(
{"_id": entry_id},
{"$set": {"is_read": is_read}}
)
if result.matched_count == 0:
raise HTTPException(status_code=404, detail="엔트리를 찾을 수 없습니다")
return {"message": f"읽음 상태가 {is_read}로 변경되었습니다"}
@app.put("/api/entries/{entry_id}/star")
async def mark_entry_as_starred(
entry_id: str = Path(..., description="엔트리 ID"),
is_starred: bool = Query(True, description="별표 상태")
):
"""엔트리 별표 상태 변경"""
result = await db.entries.update_one(
{"_id": entry_id},
{"$set": {"is_starred": is_starred}}
)
if result.matched_count == 0:
raise HTTPException(status_code=404, detail="엔트리를 찾을 수 없습니다")
return {"message": f"별표 상태가 {is_starred}로 변경되었습니다"}
@app.post("/api/entries/mark-all-read")
async def mark_all_as_read(feed_id: Optional[str] = Query(None, description="피드 ID")):
"""모든 엔트리를 읽음으로 표시"""
query = {}
if feed_id:
query["feed_id"] = feed_id
result = await db.entries.update_many(
query,
{"$set": {"is_read": True}}
)
return {"message": f"{result.modified_count}개 엔트리가 읽음으로 표시되었습니다"}
@app.get("/api/stats", response_model=List[FeedStatistics])
async def get_statistics(feed_id: Optional[str] = Query(None, description="피드 ID")):
"""피드 통계 조회"""
if feed_id:
feeds = [await db.feeds.find_one({"_id": feed_id})]
if not feeds[0]:
raise HTTPException(status_code=404, detail="피드를 찾을 수 없습니다")
else:
feeds = await db.feeds.find().to_list(100)
stats = []
for feed in feeds:
feed_id = str(feed["_id"])
# Count entries
total = await db.entries.count_documents({"feed_id": feed_id})
unread = await db.entries.count_documents({"feed_id": feed_id, "is_read": False})
starred = await db.entries.count_documents({"feed_id": feed_id, "is_starred": True})
# Calculate error rate
error_rate = 0
if feed.get("error_count", 0) > 0:
total_fetches = feed.get("error_count", 0) + (1 if feed.get("last_fetch") else 0)
error_rate = feed.get("error_count", 0) / total_fetches
stats.append(FeedStatistics(
feed_id=feed_id,
total_entries=total,
unread_entries=unread,
starred_entries=starred,
last_update=feed.get("last_fetch"),
error_rate=error_rate
))
return stats
@app.get("/api/export/opml")
async def export_opml():
"""피드 목록을 OPML 형식으로 내보내기"""
feeds = await db.feeds.find().to_list(100)
opml = """<?xml version="1.0" encoding="UTF-8"?>
<opml version="2.0">
<head>
<title>RSS Feed Subscriptions</title>
<dateCreated>{}</dateCreated>
</head>
<body>""".format(datetime.now().isoformat())
for feed in feeds:
opml += f'\n <outline text="{feed["title"]}" xmlUrl="{feed["url"]}" type="rss" category="{feed.get("category", "")}" />'
opml += "\n</body>\n</opml>"
return {
"opml": opml,
"feed_count": len(feeds)
}
# Google News RSS Endpoints
@app.get("/api/google-rss/search")
async def get_google_search_rss(
q: str = Query(..., description="검색 키워드"),
lang: str = Query("ko", description="언어 코드 (ko, en, ja, zh-CN 등)"),
country: str = Query("KR", description="국가 코드 (KR, US, JP, CN 등)")
):
"""Google News 검색 RSS 피드 URL 생성"""
feed_url = GoogleNewsRSS.search_feed(q, lang, country)
# 피드 파싱 테스트
result = await parser.parse_feed(feed_url)
return {
"keyword": q,
"feed_url": feed_url,
"success": result["success"],
"feed_title": result["feed"].get("title", "Google News") if result["success"] else None,
"entry_count": len(result["entries"]) if result["success"] else 0,
"sample_titles": [entry.get("title", "") for entry in result["entries"][:5]] if result["success"] else [],
"entries": [
{
"title": entry.get("title", ""),
"link": entry.get("link", ""),
"published": entry.get("published", ""),
"summary": entry.get("summary", "")[:200] if entry.get("summary") else ""
} for entry in result["entries"][:20]
] if result["success"] else [],
"error": result.get("error")
}
@app.get("/api/google-rss/topic")
async def get_google_topic_rss(
category: GoogleNewsCategory = Query(..., description="뉴스 카테고리"),
lang: str = Query("ko", description="언어 코드"),
country: str = Query("KR", description="국가 코드")
):
"""Google News 카테고리별 RSS 피드 URL 생성"""
feed_url = GoogleNewsRSS.topic_feed(category, lang, country)
# 피드 파싱 테스트
result = await parser.parse_feed(feed_url)
return {
"category": category,
"feed_url": feed_url,
"success": result["success"],
"feed_title": result["feed"].get("title", "Google News") if result["success"] else None,
"entry_count": len(result["entries"]) if result["success"] else 0,
"sample_titles": [entry.get("title", "") for entry in result["entries"][:5]] if result["success"] else [],
"error": result.get("error")
}
@app.get("/api/google-rss/location")
async def get_google_location_rss(
location: str = Query(..., description="지역명 (예: Seoul, 서울, New York)"),
lang: str = Query("ko", description="언어 코드"),
country: str = Query("KR", description="국가 코드")
):
"""Google News 지역 뉴스 RSS 피드 URL 생성"""
feed_url = GoogleNewsRSS.location_feed(location, lang, country)
# 피드 파싱 테스트
result = await parser.parse_feed(feed_url)
return {
"location": location,
"feed_url": feed_url,
"success": result["success"],
"feed_title": result["feed"].get("title", "Google News") if result["success"] else None,
"entry_count": len(result["entries"]) if result["success"] else 0,
"sample_titles": [entry.get("title", "") for entry in result["entries"][:5]] if result["success"] else [],
"error": result.get("error")
}
@app.get("/api/google-rss/trending")
async def get_google_trending_rss(
lang: str = Query("ko", description="언어 코드"),
country: str = Query("KR", description="국가 코드")
):
"""Google News 트렌딩 RSS 피드 URL 생성"""
feed_url = GoogleNewsRSS.trending_feed(lang, country)
# 피드 파싱 테스트
result = await parser.parse_feed(feed_url)
return {
"feed_url": feed_url,
"success": result["success"],
"feed_title": result["feed"].get("title", "Google News") if result["success"] else None,
"entry_count": len(result["entries"]) if result["success"] else 0,
"sample_titles": [entry.get("title", "") for entry in result["entries"][:5]] if result["success"] else [],
"error": result.get("error")
}
@app.post("/api/google-rss/subscribe")
async def subscribe_google_rss(
q: Optional[str] = Query(None, description="검색 키워드"),
category: Optional[GoogleNewsCategory] = Query(None, description="카테고리"),
location: Optional[str] = Query(None, description="지역명"),
trending: bool = Query(False, description="트렌딩 뉴스"),
lang: str = Query("ko", description="언어 코드"),
country: str = Query("KR", description="국가 코드"),
background_tasks: BackgroundTasks = ...
):
"""Google News RSS 피드 구독"""
# URL 생성
if q:
feed_url = GoogleNewsRSS.search_feed(q, lang, country)
feed_title = f"Google News - {q}"
elif category:
feed_url = GoogleNewsRSS.topic_feed(category, lang, country)
feed_title = f"Google News - {category.value}"
elif location:
feed_url = GoogleNewsRSS.location_feed(location, lang, country)
feed_title = f"Google News - {location}"
elif trending:
feed_url = GoogleNewsRSS.trending_feed(lang, country)
feed_title = f"Google News - Trending ({country})"
else:
raise HTTPException(status_code=400, detail="검색어, 카테고리, 지역 중 하나를 지정해주세요")
# 중복 확인
existing = await db.feeds.find_one({"url": feed_url})
if existing:
raise HTTPException(status_code=400, detail="이미 구독 중인 피드입니다")
# 피드 파싱
result = await parser.parse_feed(feed_url)
if not result["success"]:
raise HTTPException(status_code=400, detail=f"피드 파싱 실패: {result['error']}")
# 구독 생성
feed = FeedSubscription(
title=feed_title,
url=feed_url,
description=result["feed"].get("description", "Google News Feed"),
category=FeedCategory.NEWS,
update_interval=900 # 15분
)
# DB 저장
feed_dict = feed.dict()
feed_dict["url"] = str(feed_dict["url"])
result = await db.feeds.insert_one(feed_dict)
feed.id = str(result.inserted_id)
# 백그라운드 업데이트
background_tasks.add_task(update_feed, feed.id)
return feed

View File

@ -0,0 +1,74 @@
from pydantic import BaseModel, Field, HttpUrl
from typing import Optional, List, Dict, Any
from datetime import datetime
from enum import Enum
class FeedStatus(str, Enum):
ACTIVE = "active"
INACTIVE = "inactive"
ERROR = "error"
class FeedCategory(str, Enum):
NEWS = "news"
TECH = "tech"
BUSINESS = "business"
SCIENCE = "science"
HEALTH = "health"
SPORTS = "sports"
ENTERTAINMENT = "entertainment"
LIFESTYLE = "lifestyle"
POLITICS = "politics"
OTHER = "other"
class FeedSubscription(BaseModel):
id: Optional[str] = Field(None, alias="_id")
title: str
url: HttpUrl
description: Optional[str] = None
category: FeedCategory = FeedCategory.OTHER
status: FeedStatus = FeedStatus.ACTIVE
update_interval: int = 900 # seconds
last_fetch: Optional[datetime] = None
last_error: Optional[str] = None
error_count: int = 0
created_at: datetime = Field(default_factory=datetime.now)
updated_at: datetime = Field(default_factory=datetime.now)
metadata: Dict[str, Any] = {}
class FeedEntry(BaseModel):
id: Optional[str] = Field(None, alias="_id")
feed_id: str
entry_id: str # RSS entry unique ID
title: str
link: str
summary: Optional[str] = None
content: Optional[str] = None
author: Optional[str] = None
published: Optional[datetime] = None
updated: Optional[datetime] = None
categories: List[str] = []
thumbnail: Optional[str] = None
enclosures: List[Dict[str, Any]] = []
is_read: bool = False
is_starred: bool = False
created_at: datetime = Field(default_factory=datetime.now)
class CreateFeedRequest(BaseModel):
url: HttpUrl
title: Optional[str] = None
category: FeedCategory = FeedCategory.OTHER
update_interval: Optional[int] = 900
class UpdateFeedRequest(BaseModel):
title: Optional[str] = None
category: Optional[FeedCategory] = None
update_interval: Optional[int] = None
status: Optional[FeedStatus] = None
class FeedStatistics(BaseModel):
feed_id: str
total_entries: int
unread_entries: int
starred_entries: int
last_update: Optional[datetime]
error_rate: float

View File

@ -0,0 +1,14 @@
fastapi==0.109.0
uvicorn[standard]==0.27.0
pydantic==2.5.3
pydantic-settings==2.1.0
feedparser==6.0.11
httpx==0.26.0
pymongo==4.6.1
motor==3.3.2
redis==5.0.1
python-dateutil==2.8.2
beautifulsoup4==4.12.3
lxml==5.1.0
apscheduler==3.10.4
pytz==2024.1

38
check_mongodb.py Normal file
View File

@ -0,0 +1,38 @@
import pymongo
from datetime import datetime
import json
# MongoDB 연결
client = pymongo.MongoClient("mongodb://localhost:27017/")
db = client["ai_writer_db"]
collection = db["articles"]
# 최근 생성된 기사 조회
articles = collection.find().sort("created_at", -1).limit(2)
for article in articles:
print("=" * 80)
print(f"기사 ID: {article['article_id']}")
print(f"키워드: {article['keyword']}")
print(f"제목: {article['title']}")
print(f"요약: {article['summary']}")
print(f"처리 시간: {article['processing_time']:.2f}")
print(f"생성 시각: {article['created_at']}")
print(f"파이프라인 단계: {', '.join(article['pipeline_stages'])}")
print(f"카테고리: {', '.join(article['categories'])}")
print(f"태그: {', '.join(article['tags'])}")
print(f"\n내용 (첫 500자):\n{article['content'][:500]}...")
print("=" * 80)
print()
# 저장된 기사 생성
with open('generated_article.json', 'w', encoding='utf-8') as f:
# 최신 기사 하나를 다시 조회
latest = collection.find_one(sort=[("created_at", -1)])
if latest:
# ObjectId를 문자열로 변환
latest['_id'] = str(latest['_id'])
# datetime 객체를 문자열로 변환
latest['created_at'] = latest['created_at'].isoformat()
json.dump(latest, f, ensure_ascii=False, indent=2)
print(f"✅ 최신 기사가 generated_article.json 파일로 저장되었습니다.")

View File

@ -0,0 +1,18 @@
# API Keys Backup - Created on 2025-01-13
# 이 파일은 중요한 API 키를 백업한 것입니다. 안전하게 보관하세요.
# Claude API Key
CLAUDE_API_KEY=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
# Google APIs
GOOGLE_API_KEY=AIzaSyBakoCsDP_oF5V4oq_eEKs4eQb-ekqxnRM
GOOGLE_SEARCH_ENGINE_ID=35bfbdb7b6f244569
# Translation (DeepL)
DEEPL_API_KEY=3abbc796-2515-44a8-972d-22dcf27ab54a
# Image Generation (Replicate)
REPLICATE_API_TOKEN=r8_AR4puLJQYD4eeuPljw2yJvKCWKT72k119pEyp
# Additional APIs (필요시 추가)
# SERPAPI_KEY=

View File

@ -0,0 +1,21 @@
FROM python:3.11-slim
WORKDIR /app
# Install system dependencies
RUN apt-get update && apt-get install -y \
gcc \
&& rm -rf /var/lib/apt/lists/*
# Copy requirements first for better caching
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
# Copy application code
COPY . .
# Expose port
EXPOSE 8000
# Run the application
CMD ["python", "main.py"]

65
console/backend/auth.py Normal file
View File

@ -0,0 +1,65 @@
from datetime import datetime, timedelta
from typing import Optional
from jose import JWTError, jwt
from passlib.context import CryptContext
from fastapi import Depends, HTTPException, status
from fastapi.security import OAuth2PasswordBearer
from pydantic import BaseModel
import os
SECRET_KEY = os.getenv("JWT_SECRET_KEY", "your-secret-key-change-in-production")
ALGORITHM = "HS256"
ACCESS_TOKEN_EXPIRE_MINUTES = 30
pwd_context = CryptContext(schemes=["bcrypt"], deprecated="auto")
oauth2_scheme = OAuth2PasswordBearer(tokenUrl="/api/auth/login")
class Token(BaseModel):
access_token: str
token_type: str
class TokenData(BaseModel):
username: Optional[str] = None
class UserLogin(BaseModel):
username: str
password: str
class UserInDB(BaseModel):
username: str
hashed_password: str
email: str
full_name: Optional[str] = None
is_active: bool = True
def verify_password(plain_password, hashed_password):
return pwd_context.verify(plain_password, hashed_password)
def get_password_hash(password):
return pwd_context.hash(password)
def create_access_token(data: dict, expires_delta: Optional[timedelta] = None):
to_encode = data.copy()
if expires_delta:
expire = datetime.utcnow() + expires_delta
else:
expire = datetime.utcnow() + timedelta(minutes=15)
to_encode.update({"exp": expire})
encoded_jwt = jwt.encode(to_encode, SECRET_KEY, algorithm=ALGORITHM)
return encoded_jwt
async def get_current_user(token: str = Depends(oauth2_scheme)):
credentials_exception = HTTPException(
status_code=status.HTTP_401_UNAUTHORIZED,
detail="Could not validate credentials",
headers={"WWW-Authenticate": "Bearer"},
)
try:
payload = jwt.decode(token, SECRET_KEY, algorithms=[ALGORITHM])
username: str = payload.get("sub")
if username is None:
raise credentials_exception
token_data = TokenData(username=username)
except JWTError:
raise credentials_exception
return token_data

View File

@ -0,0 +1,358 @@
"""
고급 이벤트 컨슈머 with DLQ and Retry
"""
import asyncio
import json
import logging
from typing import Dict, Any, Optional, List
from datetime import datetime, timedelta
from redis import asyncio as aioredis
from aiokafka import AIOKafkaProducer
import sys
sys.path.append('/app')
from shared.kafka import KafkaConsumer, Event, EventType
from event_handlers import EventHandlers
logger = logging.getLogger(__name__)
class RetryPolicy:
"""재시도 정책"""
def __init__(
self,
max_retries: int = 3,
initial_delay: float = 1.0,
max_delay: float = 60.0,
exponential_base: float = 2.0
):
self.max_retries = max_retries
self.initial_delay = initial_delay
self.max_delay = max_delay
self.exponential_base = exponential_base
def get_delay(self, retry_count: int) -> float:
"""재시도 지연 시간 계산 (exponential backoff)"""
delay = self.initial_delay * (self.exponential_base ** retry_count)
return min(delay, self.max_delay)
class AdvancedEventConsumer:
def __init__(
self,
topics: List[str],
group_id: str,
redis_url: str = "redis://redis:6379",
bootstrap_servers: str = "kafka:9092",
enable_dlq: bool = True,
dlq_topic: str = "dead-letter-queue"
):
self.topics = topics
self.group_id = group_id
self.bootstrap_servers = bootstrap_servers
self.enable_dlq = enable_dlq
self.dlq_topic = dlq_topic
# Kafka Consumer
self.consumer = KafkaConsumer(
topics=topics,
group_id=group_id,
bootstrap_servers=bootstrap_servers
)
# DLQ Producer
self.dlq_producer: Optional[AIOKafkaProducer] = None
# Redis for retry tracking
self.redis: Optional[aioredis.Redis] = None
self.redis_url = redis_url
# Event handlers
self.handlers: Optional[EventHandlers] = None
# Retry policies per event type
self.retry_policies = {
EventType.USER_CREATED: RetryPolicy(max_retries=3),
EventType.USER_UPDATED: RetryPolicy(max_retries=2),
EventType.USER_DELETED: RetryPolicy(max_retries=5), # 중요한 이벤트
EventType.OAUTH_APP_CREATED: RetryPolicy(max_retries=3),
EventType.OAUTH_TOKEN_ISSUED: RetryPolicy(max_retries=1),
}
# Processing statistics
self.stats = {
"processed": 0,
"failed": 0,
"retried": 0,
"dlq_sent": 0
}
async def start(self):
"""컨슈머 시작"""
try:
# Redis 연결
self.redis = await aioredis.from_url(
self.redis_url,
encoding="utf-8",
decode_responses=True
)
# Event handlers 초기화
self.handlers = EventHandlers(redis_client=self.redis)
# DLQ Producer 초기화
if self.enable_dlq:
self.dlq_producer = AIOKafkaProducer(
bootstrap_servers=self.bootstrap_servers,
value_serializer=lambda v: json.dumps(v).encode()
)
await self.dlq_producer.start()
logger.info(f"DLQ Producer started for topic: {self.dlq_topic}")
# 이벤트 핸들러 등록
self._register_event_handlers()
# Kafka Consumer 시작
await self.consumer.start()
logger.info(f"Advanced Event Consumer started: {self.topics}")
# 통계 리포팅 태스크 시작
asyncio.create_task(self._report_stats())
except Exception as e:
logger.error(f"Failed to start Advanced Event Consumer: {e}")
raise
async def stop(self):
"""컨슈머 종료"""
await self.consumer.stop()
if self.dlq_producer:
await self.dlq_producer.stop()
if self.redis:
await self.redis.close()
logger.info("Advanced Event Consumer stopped")
def _register_event_handlers(self):
"""이벤트 핸들러 등록"""
# 각 이벤트 타입에 대한 핸들러를 래퍼로 감싸서 등록
self.consumer.register_handler(
EventType.USER_CREATED,
self._create_handler_with_retry(
self.handlers.handle_user_created,
EventType.USER_CREATED
)
)
self.consumer.register_handler(
EventType.USER_UPDATED,
self._create_handler_with_retry(
self.handlers.handle_user_updated,
EventType.USER_UPDATED
)
)
self.consumer.register_handler(
EventType.USER_DELETED,
self._create_handler_with_retry(
self.handlers.handle_user_deleted,
EventType.USER_DELETED
)
)
self.consumer.register_handler(
EventType.OAUTH_APP_CREATED,
self._create_handler_with_retry(
self.handlers.handle_oauth_app_created,
EventType.OAUTH_APP_CREATED
)
)
self.consumer.register_handler(
EventType.OAUTH_TOKEN_ISSUED,
self._create_handler_with_retry(
self.handlers.handle_oauth_token_issued,
EventType.OAUTH_TOKEN_ISSUED
)
)
def _create_handler_with_retry(self, handler_func, event_type: EventType):
"""재시도 로직이 포함된 핸들러 래퍼 생성"""
async def wrapper(event: Event):
event_id = f"{event.event_id}:{event.event_type}"
retry_key = f"retry:{event_id}"
try:
# 재시도 횟수 확인
retry_count = 0
if self.redis:
retry_count_str = await self.redis.get(retry_key)
retry_count = int(retry_count_str) if retry_count_str else 0
# 핸들러 실행
await handler_func(event.dict())
# 성공 시 재시도 카운터 삭제
if self.redis and retry_count > 0:
await self.redis.delete(retry_key)
self.stats["processed"] += 1
except Exception as e:
logger.error(f"Error processing {event_type}: {e}")
self.stats["failed"] += 1
# 재시도 처리
retry_policy = self.retry_policies.get(event_type)
if retry_policy and retry_count < retry_policy.max_retries:
await self._handle_retry(event, retry_count, retry_policy, retry_key)
else:
# 최대 재시도 초과 -> DLQ로 전송
await self._send_to_dlq(event, str(e), retry_count)
return wrapper
async def _handle_retry(
self,
event: Event,
retry_count: int,
retry_policy: RetryPolicy,
retry_key: str
):
"""재시도 처리"""
retry_count += 1
delay = retry_policy.get_delay(retry_count)
logger.warning(
f"Retrying event {event.event_id} "
f"(attempt {retry_count}/{retry_policy.max_retries}) "
f"after {delay}s"
)
# 재시도 카운터 저장
if self.redis:
await self.redis.setex(
retry_key,
timedelta(hours=24), # 24시간 후 자동 삭제
retry_count
)
# 지연 후 재처리를 위해 다시 큐에 추가
# 실제 프로덕션에서는 별도의 재시도 토픽 사용 권장
self.stats["retried"] += 1
# 지연 실행
await asyncio.sleep(delay)
# 이벤트 재발행 (재시도 토픽으로)
if hasattr(self, 'retry_producer'):
await self._republish_for_retry(event, retry_count)
async def _send_to_dlq(self, event: Event, error: str, retry_count: int):
"""Dead Letter Queue로 전송"""
if not self.enable_dlq or not self.dlq_producer:
logger.error(f"Failed to process event {event.event_id} after {retry_count} retries")
return
try:
dlq_message = {
"original_event": event.dict(),
"error": error,
"retry_count": retry_count,
"failed_at": datetime.now().isoformat(),
"consumer_group": self.group_id,
"topic": self.topics[0] if self.topics else None
}
await self.dlq_producer.send(
self.dlq_topic,
value=dlq_message
)
self.stats["dlq_sent"] += 1
logger.error(
f"Event {event.event_id} sent to DLQ after {retry_count} retries. "
f"Error: {error}"
)
# Redis에 DLQ 전송 기록
if self.redis:
dlq_key = f"dlq:{event.event_id}"
await self.redis.setex(
dlq_key,
timedelta(days=7), # 7일 보관
json.dumps({
"error": error,
"retry_count": retry_count,
"sent_at": datetime.now().isoformat()
})
)
except Exception as e:
logger.critical(f"Failed to send event to DLQ: {e}")
async def _republish_for_retry(self, event: Event, retry_count: int):
"""재시도를 위한 이벤트 재발행"""
# 실제 구현에서는 별도의 재시도 토픽 사용
# 여기서는 로깅만 수행
logger.info(f"Would republish event {event.event_id} for retry #{retry_count}")
async def _report_stats(self):
"""통계 리포팅 (1분마다)"""
while True:
await asyncio.sleep(60)
logger.info(
f"Event Consumer Stats - "
f"Processed: {self.stats['processed']}, "
f"Failed: {self.stats['failed']}, "
f"Retried: {self.stats['retried']}, "
f"DLQ: {self.stats['dlq_sent']}"
)
# Redis에 통계 저장
if self.redis:
stats_key = f"consumer:stats:{self.group_id}"
await self.redis.hset(
stats_key,
mapping={
**self.stats,
"updated_at": datetime.now().isoformat()
}
)
async def get_dlq_messages(self, limit: int = 10) -> List[Dict[str, Any]]:
"""DLQ 메시지 조회 (관리 목적)"""
if not self.redis:
return []
dlq_keys = await self.redis.keys("dlq:*")
messages = []
for key in dlq_keys[:limit]:
data = await self.redis.get(key)
if data:
event_id = key.replace("dlq:", "")
message = json.loads(data)
message["event_id"] = event_id
messages.append(message)
return messages
async def retry_dlq_message(self, event_id: str) -> bool:
"""DLQ 메시지 수동 재시도"""
# 실제 구현에서는 DLQ에서 메시지를 읽어 재처리
logger.info(f"Manual retry requested for event: {event_id}")
if self.redis:
# 재시도 카운터 리셋
retry_key = f"retry:{event_id}:*"
keys = await self.redis.keys(retry_key)
if keys:
await self.redis.delete(*keys)
return True
return False

View File

@ -0,0 +1,213 @@
"""
이벤트 핸들러 모듈
각 이벤트 타입별 처리 로직 구현
"""
import logging
from typing import Dict, Any, Optional
from datetime import datetime
import json
import asyncio
from redis import asyncio as aioredis
logger = logging.getLogger(__name__)
class EventHandlers:
def __init__(self, redis_client: Optional[aioredis.Redis] = None):
self.redis = redis_client
self.retry_counts: Dict[str, int] = {}
async def handle_user_created(self, event: Dict[str, Any]):
"""사용자 생성 이벤트 처리"""
try:
user_id = event.get('data', {}).get('user_id')
username = event.get('data', {}).get('username')
email = event.get('data', {}).get('email')
logger.info(f"Processing USER_CREATED: {username} ({user_id})")
# Redis 캐시 무효화
if self.redis:
await self.redis.delete(f"user:{user_id}")
await self.redis.delete("users:list")
# 추가 처리 로직
# - 환영 이메일 발송 준비
# - 초기 설정 생성
# - 분석 데이터 기록
await self._publish_notification({
"type": "user.welcome",
"user_id": user_id,
"email": email,
"username": username,
"timestamp": datetime.now().isoformat()
})
logger.info(f"Successfully processed USER_CREATED for {username}")
except Exception as e:
logger.error(f"Error handling USER_CREATED: {e}")
raise
async def handle_user_updated(self, event: Dict[str, Any]):
"""사용자 업데이트 이벤트 처리"""
try:
user_id = event.get('data', {}).get('user_id')
updated_fields = event.get('data', {}).get('updated_fields', [])
logger.info(f"Processing USER_UPDATED: {user_id}, fields: {updated_fields}")
# Redis 캐시 무효화
if self.redis:
await self.redis.delete(f"user:{user_id}")
await self.redis.delete("users:list")
# 프로필 사진 변경 시 이미지 캐시도 무효화
if 'profile_picture' in updated_fields:
await self.redis.delete(f"user:profile_picture:{user_id}")
# 프로필 완성도 계산
if 'profile_picture' in updated_fields or 'bio' in updated_fields:
await self._calculate_profile_completeness(user_id)
logger.info(f"Successfully processed USER_UPDATED for {user_id}")
except Exception as e:
logger.error(f"Error handling USER_UPDATED: {e}")
raise
async def handle_user_deleted(self, event: Dict[str, Any]):
"""사용자 삭제 이벤트 처리"""
try:
user_id = event.get('data', {}).get('user_id')
username = event.get('data', {}).get('username')
logger.info(f"Processing USER_DELETED: {username} ({user_id})")
# Redis에서 모든 관련 데이터 삭제
if self.redis:
# 사용자 캐시 삭제
await self.redis.delete(f"user:{user_id}")
await self.redis.delete("users:list")
# 세션 삭제
session_keys = await self.redis.keys(f"session:*:{user_id}")
if session_keys:
await self.redis.delete(*session_keys)
# 프로필 이미지 캐시 삭제
await self.redis.delete(f"user:profile_picture:{user_id}")
# 관련 데이터 정리 이벤트 발행
await self._publish_cleanup_event({
"user_id": user_id,
"username": username,
"timestamp": datetime.now().isoformat()
})
logger.info(f"Successfully processed USER_DELETED for {username}")
except Exception as e:
logger.error(f"Error handling USER_DELETED: {e}")
raise
async def handle_oauth_app_created(self, event: Dict[str, Any]):
"""OAuth 앱 생성 이벤트 처리"""
try:
app_id = event.get('data', {}).get('app_id')
app_name = event.get('data', {}).get('name')
owner_id = event.get('data', {}).get('owner_id')
logger.info(f"Processing OAUTH_APP_CREATED: {app_name} ({app_id})")
# 앱 생성 알림
await self._publish_notification({
"type": "oauth.app_created",
"app_id": app_id,
"app_name": app_name,
"owner_id": owner_id,
"timestamp": datetime.now().isoformat()
})
logger.info(f"Successfully processed OAUTH_APP_CREATED for {app_name}")
except Exception as e:
logger.error(f"Error handling OAUTH_APP_CREATED: {e}")
raise
async def handle_oauth_token_issued(self, event: Dict[str, Any]):
"""OAuth 토큰 발급 이벤트 처리"""
try:
client_id = event.get('data', {}).get('client_id')
user_id = event.get('data', {}).get('user_id')
scopes = event.get('data', {}).get('scopes', [])
logger.info(f"Processing OAUTH_TOKEN_ISSUED: client={client_id}, user={user_id}")
# 보안 감사 로그
await self._log_security_event({
"type": "oauth.token_issued",
"client_id": client_id,
"user_id": user_id,
"scopes": scopes,
"timestamp": datetime.now().isoformat()
})
# 사용 통계 업데이트
if self.redis:
await self.redis.hincrby(f"oauth:stats:{client_id}", "tokens_issued", 1)
await self.redis.sadd(f"oauth:users:{client_id}", user_id)
logger.info(f"Successfully processed OAUTH_TOKEN_ISSUED")
except Exception as e:
logger.error(f"Error handling OAUTH_TOKEN_ISSUED: {e}")
raise
async def _publish_notification(self, notification: Dict[str, Any]):
"""알림 이벤트 발행"""
# 향후 Notification 서비스로 이벤트 발행
logger.debug(f"Publishing notification: {notification}")
if self.redis:
await self.redis.lpush(
"notifications:queue",
json.dumps(notification)
)
async def _publish_cleanup_event(self, cleanup_data: Dict[str, Any]):
"""정리 이벤트 발행"""
# 향후 각 서비스로 정리 이벤트 발행
logger.debug(f"Publishing cleanup event: {cleanup_data}")
if self.redis:
await self.redis.lpush(
"cleanup:queue",
json.dumps(cleanup_data)
)
async def _calculate_profile_completeness(self, user_id: str):
"""프로필 완성도 계산"""
# 향후 프로필 완성도 계산 로직
logger.debug(f"Calculating profile completeness for user: {user_id}")
if self.redis:
# 임시로 Redis에 저장
await self.redis.hset(
f"user:stats:{user_id}",
"profile_updated_at",
datetime.now().isoformat()
)
async def _log_security_event(self, event_data: Dict[str, Any]):
"""보안 이벤트 로깅"""
logger.info(f"Security event: {event_data}")
if self.redis:
await self.redis.lpush(
"security:audit_log",
json.dumps(event_data)
)
# 최근 100개만 유지
await self.redis.ltrim("security:audit_log", 0, 99)

328
console/backend/main.py Normal file
View File

@ -0,0 +1,328 @@
from fastapi import FastAPI, HTTPException, Request, Response, Depends, status
from fastapi.middleware.cors import CORSMiddleware
from fastapi.security import OAuth2PasswordRequestForm
import uvicorn
from datetime import datetime, timedelta
import httpx
import os
import asyncio
import logging
from typing import Any
from contextlib import asynccontextmanager
from auth import (
Token, UserLogin, UserInDB,
verify_password, get_password_hash,
create_access_token, get_current_user,
ACCESS_TOKEN_EXPIRE_MINUTES
)
# Import event consumer
from event_consumer import AdvancedEventConsumer
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
# Global event consumer instance
event_consumer = None
@asynccontextmanager
async def lifespan(app: FastAPI):
# Startup
global event_consumer
try:
# Initialize and start event consumer
event_consumer = AdvancedEventConsumer(
topics=["user-events", "oauth-events"],
group_id="console-consumer-group",
redis_url=os.getenv("REDIS_URL", "redis://redis:6379"),
bootstrap_servers=os.getenv("KAFKA_BOOTSTRAP_SERVERS", "kafka:9092"),
enable_dlq=True,
dlq_topic="dead-letter-queue"
)
await event_consumer.start()
logger.info("Event consumer started successfully")
except Exception as e:
logger.error(f"Failed to start event consumer: {e}")
# Continue without event consumer (degraded mode)
event_consumer = None
yield
# Shutdown
if event_consumer:
await event_consumer.stop()
logger.info("Event consumer stopped")
app = FastAPI(
title="Console API Gateway",
description="Central orchestrator for microservices",
version="0.1.0",
lifespan=lifespan
)
# Service URLs from environment
USERS_SERVICE_URL = os.getenv("USERS_SERVICE_URL", "http://users-backend:8000")
IMAGES_SERVICE_URL = os.getenv("IMAGES_SERVICE_URL", "http://images-backend:8000")
# CORS middleware
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
@app.get("/")
async def root():
return {
"message": "Console API Gateway",
"status": "running",
"timestamp": datetime.now().isoformat()
}
@app.get("/health")
async def health_check():
return {
"status": "healthy",
"service": "console",
"timestamp": datetime.now().isoformat(),
"event_consumer": "running" if event_consumer else "not running"
}
# Event Management Endpoints
@app.get("/api/events/stats")
async def get_event_stats(current_user = Depends(get_current_user)):
"""Get event consumer statistics"""
if not event_consumer:
raise HTTPException(status_code=503, detail="Event consumer not available")
return {
"stats": event_consumer.stats,
"timestamp": datetime.now().isoformat()
}
@app.get("/api/events/dlq")
async def get_dlq_messages(
limit: int = 10,
current_user = Depends(get_current_user)
):
"""Get messages from Dead Letter Queue"""
if not event_consumer:
raise HTTPException(status_code=503, detail="Event consumer not available")
messages = await event_consumer.get_dlq_messages(limit=limit)
return {
"messages": messages,
"count": len(messages),
"timestamp": datetime.now().isoformat()
}
@app.post("/api/events/dlq/{event_id}/retry")
async def retry_dlq_message(
event_id: str,
current_user = Depends(get_current_user)
):
"""Manually retry a message from DLQ"""
if not event_consumer:
raise HTTPException(status_code=503, detail="Event consumer not available")
success = await event_consumer.retry_dlq_message(event_id)
if not success:
raise HTTPException(status_code=404, detail="Event not found in DLQ")
return {
"status": "retry_initiated",
"event_id": event_id,
"timestamp": datetime.now().isoformat()
}
@app.get("/api/events/schemas")
async def get_event_schemas():
"""Get all event schemas documentation"""
from shared.kafka.schema_registry import SchemaRegistry
schemas = SchemaRegistry.get_all_schemas()
return {
"schemas": schemas,
"version": "1.0.0",
"timestamp": datetime.now().isoformat()
}
# Authentication endpoints
@app.post("/api/auth/login", response_model=Token)
async def login(form_data: OAuth2PasswordRequestForm = Depends()):
"""Login endpoint for authentication"""
# For demo purposes - in production, check against database
# This is temporary until we integrate with Users service
demo_users = {
"admin": {
"username": "admin",
"hashed_password": get_password_hash("admin123"),
"email": "admin@site11.com",
"full_name": "Administrator",
"is_active": True
},
"user": {
"username": "user",
"hashed_password": get_password_hash("user123"),
"email": "user@site11.com",
"full_name": "Test User",
"is_active": True
}
}
user = demo_users.get(form_data.username)
if not user or not verify_password(form_data.password, user["hashed_password"]):
raise HTTPException(
status_code=status.HTTP_401_UNAUTHORIZED,
detail="Incorrect username or password",
headers={"WWW-Authenticate": "Bearer"},
)
access_token_expires = timedelta(minutes=ACCESS_TOKEN_EXPIRE_MINUTES)
access_token = create_access_token(
data={"sub": user["username"]}, expires_delta=access_token_expires
)
return {"access_token": access_token, "token_type": "bearer"}
@app.get("/api/auth/me")
async def get_me(current_user = Depends(get_current_user)):
"""Get current user information"""
return {
"username": current_user.username,
"email": f"{current_user.username}@site11.com",
"is_active": True
}
@app.post("/api/auth/logout")
async def logout(current_user = Depends(get_current_user)):
"""Logout endpoint"""
# In a real application, you might want to blacklist the token
return {"message": "Successfully logged out"}
@app.get("/api/status")
async def system_status():
services_status = {}
# Check Users service
try:
async with httpx.AsyncClient() as client:
response = await client.get(f"{USERS_SERVICE_URL}/health", timeout=2.0)
services_status["users"] = "online" if response.status_code == 200 else "error"
except:
services_status["users"] = "offline"
# Check Images service
try:
async with httpx.AsyncClient() as client:
response = await client.get(f"{IMAGES_SERVICE_URL}/health", timeout=2.0)
services_status["images"] = "online" if response.status_code == 200 else "error"
except:
services_status["images"] = "offline"
# Other services (not yet implemented)
services_status["oauth"] = "pending"
services_status["applications"] = "pending"
services_status["data"] = "pending"
services_status["statistics"] = "pending"
return {
"console": "online",
"services": services_status,
"timestamp": datetime.now().isoformat()
}
# Protected endpoint example
@app.get("/api/protected")
async def protected_route(current_user = Depends(get_current_user)):
"""Example of a protected route"""
return {
"message": "This is a protected route",
"user": current_user.username
}
# API Gateway - Route to Images service
@app.api_route("/api/images/{path:path}", methods=["GET", "POST", "PUT", "DELETE", "PATCH"])
async def proxy_to_images(path: str, request: Request):
"""Proxy requests to Images service (public for image proxy)"""
try:
async with httpx.AsyncClient() as client:
# Build the target URL
url = f"{IMAGES_SERVICE_URL}/api/v1/{path}"
# Get request body if exists
body = None
if request.method in ["POST", "PUT", "PATCH"]:
body = await request.body()
# Forward the request
response = await client.request(
method=request.method,
url=url,
headers={
key: value for key, value in request.headers.items()
if key.lower() not in ["host", "content-length"]
},
content=body,
params=request.query_params
)
# Return the response
return Response(
content=response.content,
status_code=response.status_code,
headers=dict(response.headers)
)
except httpx.ConnectError:
raise HTTPException(status_code=503, detail="Images service unavailable")
except Exception as e:
raise HTTPException(status_code=500, detail=str(e))
# API Gateway - Route to Users service
@app.api_route("/api/users/{path:path}", methods=["GET", "POST", "PUT", "DELETE", "PATCH"])
async def proxy_to_users(path: str, request: Request, current_user = Depends(get_current_user)):
"""Proxy requests to Users service (protected)"""
try:
async with httpx.AsyncClient() as client:
# Build the target URL
url = f"{USERS_SERVICE_URL}/{path}"
# Get request body if exists
body = None
if request.method in ["POST", "PUT", "PATCH"]:
body = await request.body()
# Forward the request
response = await client.request(
method=request.method,
url=url,
headers={
key: value for key, value in request.headers.items()
if key.lower() not in ["host", "content-length"]
},
content=body,
params=request.query_params
)
# Return the response
return Response(
content=response.content,
status_code=response.status_code,
headers=dict(response.headers)
)
except httpx.ConnectError:
raise HTTPException(status_code=503, detail="Users service unavailable")
except Exception as e:
raise HTTPException(status_code=500, detail=str(e))
if __name__ == "__main__":
uvicorn.run(
"main:app",
host="0.0.0.0",
port=8000,
reload=True
)

View File

@ -0,0 +1,10 @@
fastapi==0.109.0
uvicorn[standard]==0.27.0
python-dotenv==1.0.0
pydantic==2.5.3
httpx==0.26.0
python-jose[cryptography]==3.3.0
passlib[bcrypt]==1.7.4
python-multipart==0.0.6
redis==5.0.1
aiokafka==0.10.0

View File

@ -0,0 +1,6 @@
from .producer import KafkaProducer
from .consumer import KafkaConsumer
from .events import Event, EventType
from .schema_registry import SchemaRegistry
__all__ = ['KafkaProducer', 'KafkaConsumer', 'Event', 'EventType', 'SchemaRegistry']

View File

@ -0,0 +1,125 @@
import json
import asyncio
from typing import Optional, Callable, Dict, Any, List
from aiokafka import AIOKafkaConsumer
from aiokafka.errors import KafkaError
import logging
from .events import Event, EventType
logger = logging.getLogger(__name__)
class KafkaConsumer:
def __init__(
self,
topics: List[str],
group_id: str,
bootstrap_servers: str = "kafka:9092"
):
self.topics = topics
self.group_id = group_id
self.bootstrap_servers = bootstrap_servers
self._consumer: Optional[AIOKafkaConsumer] = None
self._handlers: Dict[EventType, List[Callable]] = {}
self._running = False
def register_handler(self, event_type: EventType, handler: Callable):
"""이벤트 타입별 핸들러 등록"""
if event_type not in self._handlers:
self._handlers[event_type] = []
self._handlers[event_type].append(handler)
logger.info(f"Registered handler for {event_type}")
async def start(self):
"""Kafka Consumer 시작"""
try:
self._consumer = AIOKafkaConsumer(
*self.topics,
bootstrap_servers=self.bootstrap_servers,
group_id=self.group_id,
value_deserializer=lambda v: json.loads(v.decode()),
auto_offset_reset='earliest',
enable_auto_commit=True,
auto_commit_interval_ms=1000,
session_timeout_ms=30000,
heartbeat_interval_ms=10000
)
await self._consumer.start()
self._running = True
logger.info(f"Kafka Consumer started: {self.topics} (group: {self.group_id})")
# 메시지 처리 루프 시작
asyncio.create_task(self._consume_messages())
except Exception as e:
logger.error(f"Failed to start Kafka Consumer: {e}")
raise
async def stop(self):
"""Kafka Consumer 종료"""
self._running = False
if self._consumer:
await self._consumer.stop()
logger.info("Kafka Consumer stopped")
async def _consume_messages(self):
"""메시지 소비 루프"""
if not self._consumer:
return
while self._running:
try:
# 메시지 배치로 가져오기 (최대 100ms 대기)
msg_batch = await self._consumer.getmany(timeout_ms=100)
for tp, messages in msg_batch.items():
for msg in messages:
await self._process_message(msg.value)
except KafkaError as e:
logger.error(f"Kafka error: {e}")
await asyncio.sleep(1)
except Exception as e:
logger.error(f"Error processing messages: {e}")
await asyncio.sleep(1)
async def _process_message(self, message: Dict[str, Any]):
"""개별 메시지 처리"""
try:
# Event 객체로 변환
event = Event(**message)
# 등록된 핸들러 실행
handlers = self._handlers.get(event.event_type, [])
for handler in handlers:
try:
if asyncio.iscoroutinefunction(handler):
await handler(event)
else:
handler(event)
except Exception as e:
logger.error(f"Handler error for {event.event_type}: {e}")
if not handlers:
logger.debug(f"No handlers for event type: {event.event_type}")
except Exception as e:
logger.error(f"Failed to process message: {e}")
async def consume_one(self, timeout: float = 1.0) -> Optional[Event]:
"""단일 메시지 소비 (테스트/디버깅용)"""
if not self._consumer:
return None
try:
msg = await asyncio.wait_for(
self._consumer.getone(),
timeout=timeout
)
return Event(**msg.value)
except asyncio.TimeoutError:
return None
except Exception as e:
logger.error(f"Error consuming message: {e}")
return None

View File

@ -0,0 +1,31 @@
from enum import Enum
from pydantic import BaseModel, Field
from datetime import datetime
from typing import Any, Optional, Dict
class EventType(str, Enum):
USER_CREATED = "user.created"
USER_UPDATED = "user.updated"
USER_DELETED = "user.deleted"
USER_LOGIN = "user.login"
IMAGE_UPLOADED = "image.uploaded"
IMAGE_CACHED = "image.cached"
IMAGE_DELETED = "image.deleted"
TASK_CREATED = "task.created"
TASK_COMPLETED = "task.completed"
TASK_FAILED = "task.failed"
class Event(BaseModel):
event_type: EventType
timestamp: datetime = Field(default_factory=datetime.now)
service: str
data: Dict[str, Any]
correlation_id: Optional[str] = None
user_id: Optional[str] = None
class Config:
json_encoders = {
datetime: lambda v: v.isoformat()
}

View File

@ -0,0 +1,101 @@
import json
import asyncio
from typing import Optional, Dict, Any
from aiokafka import AIOKafkaProducer
from aiokafka.errors import KafkaError
import logging
from .events import Event
logger = logging.getLogger(__name__)
class KafkaProducer:
def __init__(self, bootstrap_servers: str = "kafka:9092"):
self.bootstrap_servers = bootstrap_servers
self._producer: Optional[AIOKafkaProducer] = None
async def start(self):
"""Kafka Producer 시작"""
try:
self._producer = AIOKafkaProducer(
bootstrap_servers=self.bootstrap_servers,
value_serializer=lambda v: json.dumps(v).encode(),
compression_type="gzip",
acks='all',
retry_backoff_ms=100
)
await self._producer.start()
logger.info(f"Kafka Producer started: {self.bootstrap_servers}")
except Exception as e:
logger.error(f"Failed to start Kafka Producer: {e}")
raise
async def stop(self):
"""Kafka Producer 종료"""
if self._producer:
await self._producer.stop()
logger.info("Kafka Producer stopped")
async def send_event(self, topic: str, event: Event) -> bool:
"""이벤트 전송"""
if not self._producer:
logger.error("Producer not started")
return False
try:
event_dict = event.dict()
event_dict['timestamp'] = event.timestamp.isoformat()
await self._producer.send_and_wait(
topic,
value=event_dict,
key=event.correlation_id.encode() if event.correlation_id else None
)
logger.info(f"Event sent to {topic}: {event.event_type}")
return True
except KafkaError as e:
logger.error(f"Failed to send event to {topic}: {e}")
return False
except Exception as e:
logger.error(f"Unexpected error sending event: {e}")
return False
async def send_batch(self, topic: str, events: list[Event]) -> int:
"""여러 이벤트를 배치로 전송"""
if not self._producer:
logger.error("Producer not started")
return 0
sent_count = 0
batch = self._producer.create_batch()
for event in events:
event_dict = event.dict()
event_dict['timestamp'] = event.timestamp.isoformat()
metadata = batch.append(
key=event.correlation_id.encode() if event.correlation_id else None,
value=json.dumps(event_dict).encode(),
timestamp=None
)
if metadata is None:
# 배치가 가득 찼으면 전송하고 새 배치 생성
await self._producer.send_batch(batch, topic)
sent_count += len(batch)
batch = self._producer.create_batch()
batch.append(
key=event.correlation_id.encode() if event.correlation_id else None,
value=json.dumps(event_dict).encode(),
timestamp=None
)
# 남은 배치 전송
if batch:
await self._producer.send_batch(batch, topic)
sent_count += len(batch)
logger.info(f"Sent {sent_count} events to {topic}")
return sent_count

View File

@ -0,0 +1,333 @@
"""
이벤트 스키마 레지스트리
이벤트 스키마 정의 및 버전 관리
"""
from typing import Dict, Any, Optional, List, Literal
from enum import Enum
from pydantic import BaseModel, Field, field_validator
from datetime import datetime
import json
class SchemaVersion(str, Enum):
V1 = "1.0.0"
V2 = "2.0.0"
class EventSchemaBase(BaseModel):
"""이벤트 스키마 베이스"""
event_id: str = Field(..., description="고유 이벤트 ID")
event_type: str = Field(..., description="이벤트 타입")
timestamp: datetime = Field(default_factory=datetime.now, description="이벤트 발생 시간")
version: str = Field(default=SchemaVersion.V1, description="스키마 버전")
service: str = Field(..., description="이벤트 발생 서비스")
class Config:
json_encoders = {
datetime: lambda v: v.isoformat()
}
# User Events Schemas
class UserCreatedSchema(EventSchemaBase):
"""사용자 생성 이벤트 스키마"""
event_type: Literal["USER_CREATED"] = "USER_CREATED"
data: Dict[str, Any] = Field(..., description="이벤트 데이터")
@field_validator('data')
@classmethod
def validate_data(cls, v):
required_fields = ['user_id', 'username', 'email']
for field in required_fields:
if field not in v:
raise ValueError(f"Missing required field: {field}")
return v
class UserUpdatedSchema(EventSchemaBase):
"""사용자 업데이트 이벤트 스키마"""
event_type: Literal["USER_UPDATED"] = "USER_UPDATED"
data: Dict[str, Any] = Field(..., description="이벤트 데이터")
@field_validator('data')
@classmethod
def validate_data(cls, v):
required_fields = ['user_id']
optional_fields = ['username', 'email', 'full_name', 'profile_picture',
'bio', 'location', 'website', 'updated_fields']
for field in required_fields:
if field not in v:
raise ValueError(f"Missing required field: {field}")
# updated_fields가 있으면 검증
if 'updated_fields' in v and not isinstance(v['updated_fields'], list):
raise ValueError("updated_fields must be a list")
return v
class UserDeletedSchema(EventSchemaBase):
"""사용자 삭제 이벤트 스키마"""
event_type: Literal["USER_DELETED"] = "USER_DELETED"
data: Dict[str, Any] = Field(..., description="이벤트 데이터")
@field_validator('data')
@classmethod
def validate_data(cls, v):
required_fields = ['user_id', 'username']
for field in required_fields:
if field not in v:
raise ValueError(f"Missing required field: {field}")
return v
# OAuth Events Schemas
class OAuthAppCreatedSchema(EventSchemaBase):
"""OAuth 앱 생성 이벤트 스키마"""
event_type: Literal["OAUTH_APP_CREATED"] = "OAUTH_APP_CREATED"
data: Dict[str, Any] = Field(..., description="이벤트 데이터")
@field_validator('data')
@classmethod
def validate_data(cls, v):
required_fields = ['app_id', 'name', 'owner_id', 'client_id']
for field in required_fields:
if field not in v:
raise ValueError(f"Missing required field: {field}")
return v
class OAuthTokenIssuedSchema(EventSchemaBase):
"""OAuth 토큰 발급 이벤트 스키마"""
event_type: Literal["OAUTH_TOKEN_ISSUED"] = "OAUTH_TOKEN_ISSUED"
data: Dict[str, Any] = Field(..., description="이벤트 데이터")
@field_validator('data')
@classmethod
def validate_data(cls, v):
required_fields = ['client_id', 'grant_type']
optional_fields = ['user_id', 'scopes', 'expires_in']
for field in required_fields:
if field not in v:
raise ValueError(f"Missing required field: {field}")
# scopes가 있으면 리스트여야 함
if 'scopes' in v and not isinstance(v['scopes'], list):
raise ValueError("scopes must be a list")
return v
class OAuthTokenRevokedSchema(EventSchemaBase):
"""OAuth 토큰 폐기 이벤트 스키마"""
event_type: Literal["OAUTH_TOKEN_REVOKED"] = "OAUTH_TOKEN_REVOKED"
data: Dict[str, Any] = Field(..., description="이벤트 데이터")
@field_validator('data')
@classmethod
def validate_data(cls, v):
required_fields = ['token_id', 'client_id']
optional_fields = ['user_id', 'revoked_by']
for field in required_fields:
if field not in v:
raise ValueError(f"Missing required field: {field}")
return v
# Image Events Schemas
class ImageUploadedSchema(EventSchemaBase):
"""이미지 업로드 이벤트 스키마"""
event_type: Literal["IMAGE_UPLOADED"] = "IMAGE_UPLOADED"
data: Dict[str, Any] = Field(..., description="이벤트 데이터")
@field_validator('data')
@classmethod
def validate_data(cls, v):
required_fields = ['image_id', 'user_id', 'url']
optional_fields = ['size', 'mime_type', 'width', 'height', 'thumbnail_url']
for field in required_fields:
if field not in v:
raise ValueError(f"Missing required field: {field}")
return v
class ImageProcessedSchema(EventSchemaBase):
"""이미지 처리 완료 이벤트 스키마"""
event_type: Literal["IMAGE_PROCESSED"] = "IMAGE_PROCESSED"
data: Dict[str, Any] = Field(..., description="이벤트 데이터")
@field_validator('data')
@classmethod
def validate_data(cls, v):
required_fields = ['image_id', 'process_type']
optional_fields = ['original_url', 'processed_url', 'processing_time_ms']
for field in required_fields:
if field not in v:
raise ValueError(f"Missing required field: {field}")
return v
class SchemaRegistry:
"""스키마 레지스트리"""
# 스키마 매핑
SCHEMAS = {
"USER_CREATED": UserCreatedSchema,
"USER_UPDATED": UserUpdatedSchema,
"USER_DELETED": UserDeletedSchema,
"OAUTH_APP_CREATED": OAuthAppCreatedSchema,
"OAUTH_TOKEN_ISSUED": OAuthTokenIssuedSchema,
"OAUTH_TOKEN_REVOKED": OAuthTokenRevokedSchema,
"IMAGE_UPLOADED": ImageUploadedSchema,
"IMAGE_PROCESSED": ImageProcessedSchema,
}
# 스키마 버전 호환성 매트릭스
COMPATIBILITY_MATRIX = {
SchemaVersion.V1: [SchemaVersion.V1],
SchemaVersion.V2: [SchemaVersion.V1, SchemaVersion.V2], # V2는 V1과 호환
}
@classmethod
def get_schema(cls, event_type: str) -> Optional[type]:
"""이벤트 타입에 대한 스키마 반환"""
return cls.SCHEMAS.get(event_type)
@classmethod
def validate_event(cls, event_data: Dict[str, Any]) -> tuple[bool, Optional[str]]:
"""이벤트 데이터 검증"""
try:
event_type = event_data.get('event_type')
if not event_type:
return False, "Missing event_type"
schema_class = cls.get_schema(event_type)
if not schema_class:
return False, f"Unknown event type: {event_type}"
# 스키마 검증
schema_class(**event_data)
return True, None
except Exception as e:
return False, str(e)
@classmethod
def is_compatible(cls, from_version: str, to_version: str) -> bool:
"""버전 호환성 확인"""
from_v = SchemaVersion(from_version)
to_v = SchemaVersion(to_version)
compatible_versions = cls.COMPATIBILITY_MATRIX.get(to_v, [])
return from_v in compatible_versions
@classmethod
def migrate_event(
cls,
event_data: Dict[str, Any],
from_version: str,
to_version: str
) -> Dict[str, Any]:
"""이벤트 데이터 마이그레이션"""
if from_version == to_version:
return event_data
if not cls.is_compatible(from_version, to_version):
raise ValueError(f"Cannot migrate from {from_version} to {to_version}")
# 버전별 마이그레이션 로직
if from_version == SchemaVersion.V1 and to_version == SchemaVersion.V2:
# V1 -> V2 마이그레이션 예시
event_data['version'] = SchemaVersion.V2
# 새로운 필드 추가 (기본값)
if 'metadata' not in event_data:
event_data['metadata'] = {}
return event_data
@classmethod
def get_all_schemas(cls) -> Dict[str, Dict[str, Any]]:
"""모든 스키마 정보 반환 (문서화용)"""
schemas_info = {}
for event_type, schema_class in cls.SCHEMAS.items():
schemas_info[event_type] = {
"description": schema_class.__doc__,
"fields": schema_class.schema(),
"version": SchemaVersion.V1,
"example": cls._generate_example(schema_class)
}
return schemas_info
@classmethod
def _generate_example(cls, schema_class: type) -> Dict[str, Any]:
"""스키마 예시 생성"""
examples = {
"USER_CREATED": {
"event_id": "evt_123456",
"event_type": "USER_CREATED",
"timestamp": datetime.now().isoformat(),
"version": "1.0.0",
"service": "users",
"data": {
"user_id": "usr_abc123",
"username": "johndoe",
"email": "john@example.com"
}
},
"USER_UPDATED": {
"event_id": "evt_123457",
"event_type": "USER_UPDATED",
"timestamp": datetime.now().isoformat(),
"version": "1.0.0",
"service": "users",
"data": {
"user_id": "usr_abc123",
"updated_fields": ["profile_picture", "bio"],
"profile_picture": "https://example.com/pic.jpg",
"bio": "Updated bio"
}
},
"OAUTH_TOKEN_ISSUED": {
"event_id": "evt_123458",
"event_type": "OAUTH_TOKEN_ISSUED",
"timestamp": datetime.now().isoformat(),
"version": "1.0.0",
"service": "oauth",
"data": {
"client_id": "app_xyz789",
"user_id": "usr_abc123",
"grant_type": "authorization_code",
"scopes": ["profile", "email"],
"expires_in": 3600
}
}
}
return examples.get(schema_class.__fields__['event_type'].default, {})
@classmethod
def export_schemas(cls, format: str = "json") -> str:
"""스키마 내보내기"""
schemas = cls.get_all_schemas()
if format == "json":
return json.dumps(schemas, indent=2, default=str)
elif format == "markdown":
return cls._export_as_markdown(schemas)
else:
raise ValueError(f"Unsupported format: {format}")
@classmethod
def _export_as_markdown(cls, schemas: Dict[str, Dict[str, Any]]) -> str:
"""마크다운 형식으로 내보내기"""
md = "# Event Schema Registry\n\n"
for event_type, info in schemas.items():
md += f"## {event_type}\n\n"
md += f"{info['description']}\n\n"
md += f"**Version:** {info['version']}\n\n"
md += "**Example:**\n```json\n"
md += json.dumps(info['example'], indent=2, default=str)
md += "\n```\n\n"
return md

View File

@ -0,0 +1,20 @@
# Build stage
FROM node:18-alpine as builder
WORKDIR /app
COPY package.json ./
RUN npm install
COPY . .
RUN npm run build
# Production stage
FROM nginx:alpine
COPY --from=builder /app/dist /usr/share/nginx/html
COPY nginx.conf /etc/nginx/conf.d/default.conf
EXPOSE 80
CMD ["nginx", "-g", "daemon off;"]

View File

@ -0,0 +1,13 @@
<!doctype html>
<html lang="en">
<head>
<meta charset="UTF-8" />
<link rel="icon" type="image/svg+xml" href="/vite.svg" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<title>Console - Microservices Dashboard</title>
</head>
<body>
<div id="root"></div>
<script type="module" src="/src/main.tsx"></script>
</body>
</html>

View File

@ -0,0 +1,22 @@
server {
listen 80;
server_name localhost;
root /usr/share/nginx/html;
index index.html;
location / {
try_files $uri $uri/ /index.html;
}
location /api {
proxy_pass http://console-backend:8000;
proxy_http_version 1.1;
proxy_set_header Upgrade $http_upgrade;
proxy_set_header Connection 'upgrade';
proxy_set_header Host $host;
proxy_cache_bypass $http_upgrade;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Forwarded-Proto $scheme;
}
}

View File

@ -0,0 +1,33 @@
{
"name": "console-frontend",
"private": true,
"version": "0.1.0",
"type": "module",
"scripts": {
"dev": "vite",
"build": "tsc && vite build",
"preview": "vite preview"
},
"dependencies": {
"@emotion/react": "^11.11.3",
"@emotion/styled": "^11.11.0",
"@mui/material": "^5.15.2",
"@mui/icons-material": "^5.15.2",
"axios": "^1.6.3",
"react": "^18.2.0",
"react-dom": "^18.2.0",
"react-router-dom": "^6.21.1"
},
"devDependencies": {
"@types/react": "^18.2.43",
"@types/react-dom": "^18.2.17",
"@typescript-eslint/eslint-plugin": "^6.14.0",
"@typescript-eslint/parser": "^6.14.0",
"@vitejs/plugin-react": "^4.2.1",
"eslint": "^8.55.0",
"eslint-plugin-react-hooks": "^4.6.0",
"eslint-plugin-react-refresh": "^0.4.5",
"typescript": "^5.2.2",
"vite": "^5.0.8"
}
}

View File

@ -0,0 +1,19 @@
import { Routes, Route } from 'react-router-dom'
import Layout from './components/Layout'
import Dashboard from './pages/Dashboard'
import Services from './pages/Services'
import Users from './pages/Users'
function App() {
return (
<Routes>
<Route path="/" element={<Layout />}>
<Route index element={<Dashboard />} />
<Route path="services" element={<Services />} />
<Route path="users" element={<Users />} />
</Route>
</Routes>
)
}
export default App

View File

@ -0,0 +1,102 @@
import { useState } from 'react'
import { Outlet, Link as RouterLink } from 'react-router-dom'
import {
AppBar,
Box,
Drawer,
IconButton,
List,
ListItem,
ListItemButton,
ListItemIcon,
ListItemText,
Toolbar,
Typography,
} from '@mui/material'
import {
Menu as MenuIcon,
Dashboard as DashboardIcon,
Cloud as CloudIcon,
People as PeopleIcon,
} from '@mui/icons-material'
const drawerWidth = 240
const menuItems = [
{ text: 'Dashboard', icon: <DashboardIcon />, path: '/' },
{ text: 'Services', icon: <CloudIcon />, path: '/services' },
{ text: 'Users', icon: <PeopleIcon />, path: '/users' },
]
function Layout() {
const [open, setOpen] = useState(true)
const handleDrawerToggle = () => {
setOpen(!open)
}
return (
<Box sx={{ display: 'flex' }}>
<AppBar
position="fixed"
sx={{ zIndex: (theme) => theme.zIndex.drawer + 1 }}
>
<Toolbar>
<IconButton
color="inherit"
aria-label="open drawer"
edge="start"
onClick={handleDrawerToggle}
sx={{ mr: 2 }}
>
<MenuIcon />
</IconButton>
<Typography variant="h6" noWrap component="div">
Microservices Console
</Typography>
</Toolbar>
</AppBar>
<Drawer
variant="persistent"
anchor="left"
open={open}
sx={{
width: drawerWidth,
flexShrink: 0,
'& .MuiDrawer-paper': {
width: drawerWidth,
boxSizing: 'border-box',
},
}}
>
<Toolbar />
<Box sx={{ overflow: 'auto' }}>
<List>
{menuItems.map((item) => (
<ListItem key={item.text} disablePadding>
<ListItemButton component={RouterLink} to={item.path}>
<ListItemIcon>{item.icon}</ListItemIcon>
<ListItemText primary={item.text} />
</ListItemButton>
</ListItem>
))}
</List>
</Box>
</Drawer>
<Box
component="main"
sx={{
flexGrow: 1,
p: 3,
marginLeft: open ? `${drawerWidth}px` : 0,
transition: 'margin 0.3s',
}}
>
<Toolbar />
<Outlet />
</Box>
</Box>
)
}
export default Layout

View File

@ -0,0 +1,29 @@
import React from 'react'
import ReactDOM from 'react-dom/client'
import { BrowserRouter } from 'react-router-dom'
import { ThemeProvider, createTheme } from '@mui/material/styles'
import CssBaseline from '@mui/material/CssBaseline'
import App from './App'
const theme = createTheme({
palette: {
mode: 'light',
primary: {
main: '#1976d2',
},
secondary: {
main: '#dc004e',
},
},
})
ReactDOM.createRoot(document.getElementById('root')!).render(
<React.StrictMode>
<BrowserRouter>
<ThemeProvider theme={theme}>
<CssBaseline />
<App />
</ThemeProvider>
</BrowserRouter>
</React.StrictMode>,
)

View File

@ -0,0 +1,153 @@
import { useEffect, useState } from 'react'
import {
Grid,
Paper,
Typography,
Box,
Card,
CardContent,
Chip,
} from '@mui/material'
import {
CheckCircle as CheckCircleIcon,
Error as ErrorIcon
} from '@mui/icons-material'
import axios from 'axios'
interface ServiceStatus {
name: string
status: 'healthy' | 'unhealthy'
endpoint: string
lastChecked: string
}
function Dashboard() {
const [services, setServices] = useState<ServiceStatus[]>([])
const [stats, setStats] = useState({
totalServices: 0,
healthyServices: 0,
unhealthyServices: 0,
})
useEffect(() => {
checkServices()
const interval = setInterval(checkServices, 10000)
return () => clearInterval(interval)
}, [])
const checkServices = async () => {
const serviceChecks = [
{ name: 'Console Backend', endpoint: '/api/health' },
{ name: 'Users Service', endpoint: '/api/users/health' },
]
const results = await Promise.all(
serviceChecks.map(async (service) => {
try {
await axios.get(service.endpoint)
return {
...service,
status: 'healthy' as const,
lastChecked: new Date().toLocaleTimeString(),
}
} catch {
return {
...service,
status: 'unhealthy' as const,
lastChecked: new Date().toLocaleTimeString(),
}
}
})
)
setServices(results)
const healthy = results.filter(s => s.status === 'healthy').length
setStats({
totalServices: results.length,
healthyServices: healthy,
unhealthyServices: results.length - healthy,
})
}
return (
<Box>
<Typography variant="h4" gutterBottom>
Dashboard
</Typography>
<Grid container spacing={3} sx={{ mb: 3 }}>
<Grid item xs={12} md={4}>
<Card>
<CardContent>
<Typography color="textSecondary" gutterBottom>
Total Services
</Typography>
<Typography variant="h3">
{stats.totalServices}
</Typography>
</CardContent>
</Card>
</Grid>
<Grid item xs={12} md={4}>
<Card>
<CardContent>
<Typography color="textSecondary" gutterBottom>
Healthy Services
</Typography>
<Typography variant="h3" color="success.main">
{stats.healthyServices}
</Typography>
</CardContent>
</Card>
</Grid>
<Grid item xs={12} md={4}>
<Card>
<CardContent>
<Typography color="textSecondary" gutterBottom>
Unhealthy Services
</Typography>
<Typography variant="h3" color="error.main">
{stats.unhealthyServices}
</Typography>
</CardContent>
</Card>
</Grid>
</Grid>
<Paper sx={{ p: 2 }}>
<Typography variant="h6" gutterBottom>
Service Status
</Typography>
<Grid container spacing={2}>
{services.map((service) => (
<Grid item xs={12} md={6} key={service.name}>
<Card variant="outlined">
<CardContent>
<Box display="flex" justifyContent="space-between" alignItems="center">
<Box>
<Typography variant="h6">{service.name}</Typography>
<Typography variant="body2" color="textSecondary">
{service.endpoint}
</Typography>
<Typography variant="caption" color="textSecondary">
Last checked: {service.lastChecked}
</Typography>
</Box>
<Chip
label={service.status}
color={service.status === 'healthy' ? 'success' : 'error'}
icon={service.status === 'healthy' ? <CheckCircleIcon /> : <ErrorIcon />}
/>
</Box>
</CardContent>
</Card>
</Grid>
))}
</Grid>
</Paper>
</Box>
)
}
export default Dashboard

View File

@ -0,0 +1,98 @@
import {
Box,
Typography,
Table,
TableBody,
TableCell,
TableContainer,
TableHead,
TableRow,
Paper,
Chip,
} from '@mui/material'
const servicesData = [
{
id: 1,
name: 'Console',
type: 'API Gateway',
port: 8011,
status: 'Running',
description: 'Central orchestrator and API gateway',
},
{
id: 2,
name: 'Users',
type: 'Microservice',
port: 8001,
status: 'Running',
description: 'User management service',
},
{
id: 3,
name: 'MongoDB',
type: 'Database',
port: 27017,
status: 'Running',
description: 'Document database for persistence',
},
{
id: 4,
name: 'Redis',
type: 'Cache',
port: 6379,
status: 'Running',
description: 'In-memory cache and pub/sub',
},
]
function Services() {
return (
<Box>
<Typography variant="h4" gutterBottom>
Services
</Typography>
<TableContainer component={Paper}>
<Table>
<TableHead>
<TableRow>
<TableCell>Service Name</TableCell>
<TableCell>Type</TableCell>
<TableCell>Port</TableCell>
<TableCell>Status</TableCell>
<TableCell>Description</TableCell>
</TableRow>
</TableHead>
<TableBody>
{servicesData.map((service) => (
<TableRow key={service.id}>
<TableCell>
<Typography variant="subtitle2">{service.name}</Typography>
</TableCell>
<TableCell>
<Chip
label={service.type}
size="small"
color={service.type === 'API Gateway' ? 'primary' : 'default'}
/>
</TableCell>
<TableCell>{service.port}</TableCell>
<TableCell>
<Chip
label={service.status}
size="small"
color="success"
/>
</TableCell>
<TableCell>{service.description}</TableCell>
</TableRow>
))}
</TableBody>
</Table>
</TableContainer>
</Box>
)
}
export default Services

View File

@ -0,0 +1,208 @@
import { useState, useEffect } from 'react'
import {
Box,
Typography,
Table,
TableBody,
TableCell,
TableContainer,
TableHead,
TableRow,
Paper,
Button,
IconButton,
TextField,
Dialog,
DialogTitle,
DialogContent,
DialogActions,
Stack,
} from '@mui/material'
import {
Add as AddIcon,
Edit as EditIcon,
Delete as DeleteIcon,
} from '@mui/icons-material'
import axios from 'axios'
interface User {
_id: string
username: string
email: string
full_name?: string
created_at: string
}
function Users() {
const [users, setUsers] = useState<User[]>([])
const [openDialog, setOpenDialog] = useState(false)
const [editingUser, setEditingUser] = useState<User | null>(null)
const [formData, setFormData] = useState({
username: '',
email: '',
full_name: '',
})
useEffect(() => {
fetchUsers()
}, [])
const fetchUsers = async () => {
try {
const response = await axios.get('/api/users/')
setUsers(response.data)
} catch (error) {
console.error('Failed to fetch users:', error)
}
}
const handleOpenDialog = (user?: User) => {
if (user) {
setEditingUser(user)
setFormData({
username: user.username,
email: user.email,
full_name: user.full_name || '',
})
} else {
setEditingUser(null)
setFormData({
username: '',
email: '',
full_name: '',
})
}
setOpenDialog(true)
}
const handleCloseDialog = () => {
setOpenDialog(false)
setEditingUser(null)
setFormData({
username: '',
email: '',
full_name: '',
})
}
const handleSubmit = async () => {
try {
if (editingUser) {
await axios.put(`/api/users/${editingUser._id}`, formData)
} else {
await axios.post('/api/users/', formData)
}
fetchUsers()
handleCloseDialog()
} catch (error) {
console.error('Failed to save user:', error)
}
}
const handleDelete = async (id: string) => {
if (confirm('Are you sure you want to delete this user?')) {
try {
await axios.delete(`/api/users/${id}`)
fetchUsers()
} catch (error) {
console.error('Failed to delete user:', error)
}
}
}
return (
<Box>
<Box display="flex" justifyContent="space-between" alignItems="center" mb={2}>
<Typography variant="h4">
Users
</Typography>
<Button
variant="contained"
startIcon={<AddIcon />}
onClick={() => handleOpenDialog()}
>
Add User
</Button>
</Box>
<TableContainer component={Paper}>
<Table>
<TableHead>
<TableRow>
<TableCell>Username</TableCell>
<TableCell>Email</TableCell>
<TableCell>Full Name</TableCell>
<TableCell>Created At</TableCell>
<TableCell align="right">Actions</TableCell>
</TableRow>
</TableHead>
<TableBody>
{users.map((user) => (
<TableRow key={user._id}>
<TableCell>{user.username}</TableCell>
<TableCell>{user.email}</TableCell>
<TableCell>{user.full_name || '-'}</TableCell>
<TableCell>
{new Date(user.created_at).toLocaleDateString()}
</TableCell>
<TableCell align="right">
<IconButton
size="small"
onClick={() => handleOpenDialog(user)}
>
<EditIcon />
</IconButton>
<IconButton
size="small"
onClick={() => handleDelete(user._id)}
>
<DeleteIcon />
</IconButton>
</TableCell>
</TableRow>
))}
</TableBody>
</Table>
</TableContainer>
<Dialog open={openDialog} onClose={handleCloseDialog} maxWidth="sm" fullWidth>
<DialogTitle>
{editingUser ? 'Edit User' : 'Add New User'}
</DialogTitle>
<DialogContent>
<Stack spacing={2} sx={{ mt: 1 }}>
<TextField
label="Username"
value={formData.username}
onChange={(e) => setFormData({ ...formData, username: e.target.value })}
fullWidth
required
/>
<TextField
label="Email"
type="email"
value={formData.email}
onChange={(e) => setFormData({ ...formData, email: e.target.value })}
fullWidth
required
/>
<TextField
label="Full Name"
value={formData.full_name}
onChange={(e) => setFormData({ ...formData, full_name: e.target.value })}
fullWidth
/>
</Stack>
</DialogContent>
<DialogActions>
<Button onClick={handleCloseDialog}>Cancel</Button>
<Button onClick={handleSubmit} variant="contained">
{editingUser ? 'Update' : 'Create'}
</Button>
</DialogActions>
</Dialog>
</Box>
)
}
export default Users

View File

@ -0,0 +1,25 @@
{
"compilerOptions": {
"target": "ES2020",
"useDefineForClassFields": true,
"lib": ["ES2020", "DOM", "DOM.Iterable"],
"module": "ESNext",
"skipLibCheck": true,
/* Bundler mode */
"moduleResolution": "bundler",
"allowImportingTsExtensions": true,
"resolveJsonModule": true,
"isolatedModules": true,
"noEmit": true,
"jsx": "react-jsx",
/* Linting */
"strict": true,
"noUnusedLocals": true,
"noUnusedParameters": true,
"noFallthroughCasesInSwitch": true
},
"include": ["src"],
"references": [{ "path": "./tsconfig.node.json" }]
}

View File

@ -0,0 +1,10 @@
{
"compilerOptions": {
"composite": true,
"skipLibCheck": true,
"module": "ESNext",
"moduleResolution": "bundler",
"allowSyntheticDefaultImports": true
},
"include": ["vite.config.ts"]
}

View File

@ -0,0 +1,17 @@
import { defineConfig } from 'vite'
import react from '@vitejs/plugin-react'
// https://vitejs.dev/config/
export default defineConfig({
plugins: [react()],
server: {
host: '0.0.0.0',
port: 3000,
proxy: {
'/api': {
target: 'http://console-backend:8000',
changeOrigin: true
}
}
}
})

View File

@ -0,0 +1,85 @@
version: '3.8'
services:
# 키워드별 전용 스케줄러 (AI 키워드)
pipeline-scheduler-ai:
build:
context: ./services/pipeline
dockerfile: scheduler/Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_scheduler_ai
restart: unless-stopped
depends_on:
- redis
- mongodb
- pipeline-rss-collector
environment:
- REDIS_URL=redis://redis:6379
- MONGODB_URL=mongodb://mongodb:27017
- DB_NAME=ai_writer_db
- KEYWORD=AI
- INTERVAL_MINUTES=60
- PRIORITY=1
- MAX_ARTICLES=100
- LOG_LEVEL=INFO
command: python single_keyword_scheduler.py
volumes:
- ./services/pipeline/shared:/app/shared:ro
networks:
- site11_network
# 키워드별 전용 스케줄러 (경제 키워드)
pipeline-scheduler-economy:
build:
context: ./services/pipeline
dockerfile: scheduler/Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_scheduler_economy
restart: unless-stopped
depends_on:
- redis
- mongodb
- pipeline-rss-collector
environment:
- REDIS_URL=redis://redis:6379
- MONGODB_URL=mongodb://mongodb:27017
- DB_NAME=ai_writer_db
- KEYWORD=경제
- INTERVAL_MINUTES=120
- PRIORITY=0
- MAX_ARTICLES=100
- LOG_LEVEL=INFO
command: python single_keyword_scheduler.py
volumes:
- ./services/pipeline/shared:/app/shared:ro
networks:
- site11_network
# 키워드별 전용 스케줄러 (테크놀로지 키워드)
pipeline-scheduler-tech:
build:
context: ./services/pipeline
dockerfile: scheduler/Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_scheduler_tech
restart: unless-stopped
depends_on:
- redis
- mongodb
- pipeline-rss-collector
environment:
- REDIS_URL=redis://redis:6379
- MONGODB_URL=mongodb://mongodb:27017
- DB_NAME=ai_writer_db
- KEYWORD=테크놀로지
- INTERVAL_MINUTES=60
- PRIORITY=1
- MAX_ARTICLES=100
- LOG_LEVEL=INFO
command: python single_keyword_scheduler.py
volumes:
- ./services/pipeline/shared:/app/shared:ro
networks:
- site11_network
networks:
site11_network:
external: true
name: site11_network

685
docker-compose.yml Normal file
View File

@ -0,0 +1,685 @@
services:
console-frontend:
build:
context: ./console/frontend
dockerfile: Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_console_frontend
ports:
- "${CONSOLE_FRONTEND_PORT}:80"
networks:
- site11_network
restart: unless-stopped
depends_on:
- console-backend
console-backend:
build:
context: ./console/backend
dockerfile: Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_console_backend
ports:
- "${CONSOLE_BACKEND_PORT}:8000"
environment:
- ENV=${ENV}
- PORT=8000
- USERS_SERVICE_URL=${USERS_SERVICE_URL}
- JWT_SECRET_KEY=${JWT_SECRET_KEY}
- JWT_ALGORITHM=${JWT_ALGORITHM}
- ACCESS_TOKEN_EXPIRE_MINUTES=${ACCESS_TOKEN_EXPIRE_MINUTES}
volumes:
- ./console/backend:/app
networks:
- site11_network
restart: unless-stopped
depends_on:
- users-backend
users-backend:
build:
context: ./services/users/backend
dockerfile: Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_users_backend
ports:
- "${USERS_BACKEND_PORT}:8000"
environment:
- ENV=${ENV}
- PORT=8000
- MONGODB_URL=${MONGODB_URL}
- DB_NAME=${USERS_DB_NAME}
- KAFKA_BOOTSTRAP_SERVERS=${KAFKA_BOOTSTRAP_SERVERS}
- KAFKA_GROUP_ID=${KAFKA_GROUP_ID}
volumes:
- ./services/users/backend:/app
- ./shared:/app/shared
networks:
- site11_network
restart: unless-stopped
depends_on:
- mongodb
- kafka
images-backend:
build:
context: ./services/images/backend
dockerfile: Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_images_backend
ports:
- "${IMAGES_SERVICE_PORT}:8000"
environment:
- ENV=${ENV}
- PORT=8000
- REDIS_URL=${REDIS_URL}
- MONGODB_URL=${MONGODB_URL}
- CACHE_DIR=/app/cache
- CONVERT_TO_WEBP=true
volumes:
- ./services/images/backend:/app
- ./data/images-cache:/app/cache
networks:
- site11_network
restart: unless-stopped
depends_on:
- redis
- mongodb
oauth-backend:
build:
context: ./services/oauth/backend
dockerfile: Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_oauth_backend
ports:
- "${OAUTH_SERVICE_PORT}:8000"
environment:
- ENV=${ENV}
- PORT=8000
- MONGODB_URL=${MONGODB_URL}
- OAUTH_DB_NAME=${OAUTH_DB_NAME}
- JWT_SECRET_KEY=${JWT_SECRET_KEY}
- JWT_ALGORITHM=${JWT_ALGORITHM}
- KAFKA_BOOTSTRAP_SERVERS=${KAFKA_BOOTSTRAP_SERVERS}
- KAFKA_GROUP_ID=${KAFKA_GROUP_ID}
volumes:
- ./services/oauth/backend:/app
- ./shared:/app/shared
networks:
- site11_network
restart: unless-stopped
depends_on:
- mongodb
- kafka
mongodb:
image: mongo:7.0
container_name: ${COMPOSE_PROJECT_NAME}_mongodb
environment:
- MONGO_INITDB_DATABASE=${MONGODB_DATABASE}
ports:
- "${MONGODB_PORT}:27017"
volumes:
- ./data/mongodb:/data/db
- ./data/mongodb/configdb:/data/configdb
networks:
- site11_network
restart: unless-stopped
healthcheck:
test: echo 'db.runCommand("ping").ok' | mongosh localhost:27017/test --quiet
interval: 10s
timeout: 5s
retries: 5
redis:
image: redis:7-alpine
container_name: ${COMPOSE_PROJECT_NAME}_redis
ports:
- "${REDIS_PORT}:6379"
volumes:
- ./data/redis:/data
networks:
- site11_network
restart: unless-stopped
healthcheck:
test: ["CMD", "redis-cli", "ping"]
interval: 10s
timeout: 5s
retries: 5
zookeeper:
image: confluentinc/cp-zookeeper:7.5.0
container_name: ${COMPOSE_PROJECT_NAME}_zookeeper
environment:
ZOOKEEPER_CLIENT_PORT: 2181
ZOOKEEPER_TICK_TIME: 2000
ports:
- "${KAFKA_ZOOKEEPER_PORT}:2181"
volumes:
- ./data/zookeeper/data:/var/lib/zookeeper/data
- ./data/zookeeper/logs:/var/lib/zookeeper/log
networks:
- site11_network
restart: unless-stopped
kafka:
image: confluentinc/cp-kafka:7.5.0
container_name: ${COMPOSE_PROJECT_NAME}_kafka
depends_on:
- zookeeper
ports:
- "${KAFKA_PORT}:9092"
- "9101:9101"
environment:
KAFKA_BROKER_ID: 1
KAFKA_ZOOKEEPER_CONNECT: zookeeper:2181
KAFKA_LISTENER_SECURITY_PROTOCOL_MAP: PLAINTEXT:PLAINTEXT,PLAINTEXT_HOST:PLAINTEXT
KAFKA_ADVERTISED_LISTENERS: PLAINTEXT://kafka:29092,PLAINTEXT_HOST://localhost:9092
KAFKA_OFFSETS_TOPIC_REPLICATION_FACTOR: 1
KAFKA_TRANSACTION_STATE_LOG_MIN_ISR: 1
KAFKA_TRANSACTION_STATE_LOG_REPLICATION_FACTOR: 1
KAFKA_GROUP_INITIAL_REBALANCE_DELAY_MS: 0
KAFKA_JMX_PORT: 9101
KAFKA_JMX_HOSTNAME: localhost
KAFKA_AUTO_CREATE_TOPICS_ENABLE: 'true'
volumes:
- ./data/kafka:/var/lib/kafka/data
networks:
- site11_network
restart: unless-stopped
healthcheck:
test: ["CMD", "kafka-broker-api-versions", "--bootstrap-server", "localhost:9092"]
interval: 10s
timeout: 5s
retries: 5
# Notifications Service
notifications-backend:
build:
context: ./services/notifications/backend
dockerfile: Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_notifications_backend
ports:
- "8013:8000"
environment:
- MONGODB_URL=mongodb://mongodb:27017
- REDIS_URL=redis://redis:6379
- KAFKA_BOOTSTRAP_SERVERS=kafka:9092
- SMTP_HOST=${SMTP_HOST:-smtp.gmail.com}
- SMTP_PORT=${SMTP_PORT:-587}
- SMTP_USER=${SMTP_USER:-}
- SMTP_PASSWORD=${SMTP_PASSWORD:-}
- SMS_API_KEY=${SMS_API_KEY:-}
- SMS_API_URL=${SMS_API_URL:-}
- FCM_SERVER_KEY=${FCM_SERVER_KEY:-}
depends_on:
- mongodb
- redis
- kafka
networks:
- site11_network
restart: unless-stopped
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
interval: 30s
timeout: 10s
retries: 3
# MinIO Object Storage
minio:
image: minio/minio:latest
container_name: ${COMPOSE_PROJECT_NAME}_minio
ports:
- "9000:9000"
- "9001:9001"
environment:
- MINIO_ROOT_USER=${MINIO_ROOT_USER:-minioadmin}
- MINIO_ROOT_PASSWORD=${MINIO_ROOT_PASSWORD:-minioadmin}
volumes:
- ./data/minio:/data
command: server /data --console-address ":9001"
networks:
- site11_network
restart: unless-stopped
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
interval: 30s
timeout: 20s
retries: 3
# File Management Service
files-backend:
build:
context: ./services/files/backend
dockerfile: Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_files_backend
ports:
- "8014:8000"
environment:
- ENV=${ENV}
- PORT=8000
- MONGODB_URL=${MONGODB_URL}
- FILES_DB_NAME=${FILES_DB_NAME:-files_db}
- MINIO_ENDPOINT=minio:9000
- MINIO_ACCESS_KEY=${MINIO_ACCESS_KEY:-minioadmin}
- MINIO_SECRET_KEY=${MINIO_SECRET_KEY:-minioadmin}
- MINIO_SECURE=false
volumes:
- ./services/files/backend:/app
- ./data/files-temp:/tmp
networks:
- site11_network
restart: unless-stopped
depends_on:
- mongodb
- minio
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
interval: 30s
timeout: 10s
retries: 3
# Apache Solr Search Engine
solr:
image: solr:9.4
container_name: ${COMPOSE_PROJECT_NAME}_solr
ports:
- "8983:8983"
volumes:
- ./data/solr:/var/solr
- ./services/search/solr-config:/opt/solr/server/solr/configsets/site11_config
command:
- solr-precreate
- site11
- /opt/solr/server/solr/configsets/site11_config
networks:
- site11_network
restart: unless-stopped
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8983/solr/site11/admin/ping"]
interval: 30s
timeout: 10s
retries: 3
# Search Service
search-backend:
build:
context: ./services/search/backend
dockerfile: Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_search_backend
ports:
- "8015:8000"
environment:
- ENV=${ENV}
- PORT=8000
- SOLR_URL=http://solr:8983/solr
- MONGODB_URL=${MONGODB_URL}
- KAFKA_BOOTSTRAP_SERVERS=${KAFKA_BOOTSTRAP_SERVERS}
volumes:
- ./services/search/backend:/app
networks:
- site11_network
restart: unless-stopped
depends_on:
- solr
- mongodb
- kafka
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
interval: 30s
timeout: 10s
retries: 3
# Statistics Service
statistics-backend:
build:
context: ./services/statistics/backend
dockerfile: Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_statistics_backend
ports:
- "8012:8000"
environment:
- REDIS_URL=redis://redis:6379
- KAFKA_BOOTSTRAP_SERVERS=kafka:9092
- INFLUXDB_HOST=influxdb
- INFLUXDB_PORT=8086
- INFLUXDB_DATABASE=statistics
depends_on:
- redis
- kafka
networks:
- site11_network
restart: unless-stopped
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
interval: 30s
timeout: 10s
retries: 3
# Google Search Service
google-search-backend:
build:
context: ./services/google-search/backend
dockerfile: Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_google_search_backend
ports:
- "8016:8000"
environment:
- REDIS_HOST=redis
- REDIS_PORT=6379
- REDIS_DB=2
- GOOGLE_API_KEY=AIzaSyBakoCsDP_oF5V4oq_eEKs4eQb-ekqxnRM
- GOOGLE_SEARCH_ENGINE_ID=35bfbdb7b6f244569
- SERPAPI_KEY=${SERPAPI_KEY:-}
- DEFAULT_LANGUAGE=ko
- DEFAULT_COUNTRY=kr
- CACHE_TTL=3600
depends_on:
- redis
networks:
- site11_network
restart: unless-stopped
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
interval: 30s
timeout: 10s
retries: 3
# RSS Feed Service
rss-feed-backend:
build:
context: ./services/rss-feed/backend
dockerfile: Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_rss_feed_backend
ports:
- "8017:8000"
environment:
- MONGODB_URL=mongodb://mongodb:27017
- DB_NAME=rss_feed_db
- REDIS_URL=redis://redis:6379
- REDIS_DB=3
- DEFAULT_UPDATE_INTERVAL=900
- MAX_ENTRIES_PER_FEED=100
- ENABLE_SCHEDULER=true
- SCHEDULER_TIMEZONE=Asia/Seoul
depends_on:
- mongodb
- redis
networks:
- site11_network
restart: unless-stopped
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
interval: 30s
timeout: 10s
retries: 3
# News Aggregator Service
news-aggregator-backend:
build:
context: ./services/news-aggregator/backend
dockerfile: Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_news_aggregator_backend
ports:
- "8018:8000"
environment:
- RSS_SERVICE_URL=http://rss-feed-backend:8000
- GOOGLE_SEARCH_SERVICE_URL=http://google-search-backend:8000
depends_on:
- rss-feed-backend
- google-search-backend
networks:
- site11_network
restart: unless-stopped
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
interval: 30s
timeout: 10s
retries: 3
# AI Writer Service
ai-writer-backend:
build:
context: ./services/ai-writer/backend
dockerfile: Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_ai_writer_backend
ports:
- "8019:8000"
environment:
- NEWS_AGGREGATOR_URL=http://news-aggregator-backend:8000
- CLAUDE_API_KEY=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
- MONGODB_URL=mongodb://mongodb:27017
- DB_NAME=ai_writer_db
- REDIS_URL=redis://redis:6379
depends_on:
- mongodb
- redis
- news-aggregator-backend
networks:
- site11_network
restart: unless-stopped
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
interval: 30s
timeout: 10s
retries: 3
# AI Writer Worker Service
ai-writer-worker:
build:
context: ./services/ai-writer
dockerfile: worker/Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_ai_writer_worker
environment:
- CLAUDE_API_KEY=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
- MONGODB_URL=mongodb://mongodb:27017
- DB_NAME=ai_writer_db
- REDIS_URL=redis://redis:6379
- WORKER_COUNT=3
depends_on:
- mongodb
- redis
- ai-writer-backend
networks:
- site11_network
restart: unless-stopped
# ============ Pipeline Services ============
# Pipeline Multi-threaded Scheduler Service
pipeline-scheduler:
build:
context: ./services/pipeline
dockerfile: scheduler/Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_pipeline_scheduler
restart: unless-stopped
depends_on:
- redis
- mongodb
environment:
- REDIS_URL=redis://redis:6379
- MONGODB_URL=mongodb://mongodb:27017
- DB_NAME=ai_writer_db
- LOG_LEVEL=INFO
command: python multi_thread_scheduler.py
volumes:
- ./services/pipeline/shared:/app/shared:ro
networks:
- site11_network
# Keyword Manager API Service
keyword-manager:
build:
context: ./services/pipeline
dockerfile: scheduler/Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_keyword_manager
restart: unless-stopped
depends_on:
- mongodb
environment:
- MONGODB_URL=mongodb://mongodb:27017
- DB_NAME=ai_writer_db
- API_PORT=8100
command: python keyword_manager.py
ports:
- "8100:8100"
volumes:
- ./services/pipeline/shared:/app/shared:ro
networks:
- site11_network
# Pipeline RSS Collector Worker
pipeline-rss-collector:
build:
context: ./services/pipeline
dockerfile: rss-collector/Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_pipeline_rss_collector
restart: unless-stopped
depends_on:
- redis
environment:
- REDIS_URL=redis://redis:6379
- LOG_LEVEL=INFO
- RSS_ENQUEUE_DELAY=1.0
volumes:
- ./services/pipeline/shared:/app/shared:ro
networks:
- site11_network
# Pipeline Google Search Worker
pipeline-google-search:
build:
context: ./services/pipeline
dockerfile: google-search/Dockerfile
restart: unless-stopped
depends_on:
- redis
environment:
- REDIS_URL=redis://redis:6379
- GOOGLE_API_KEY=AIzaSyBakoCsDP_oF5V4oq_eEKs4eQb-ekqxnRM
- GOOGLE_SEARCH_ENGINE_ID=35bfbdb7b6f244569
- LOG_LEVEL=INFO
volumes:
- ./services/pipeline/shared:/app/shared:ro
networks:
- site11_network
# Pipeline AI Article Generator Worker
pipeline-ai-article-generator:
build:
context: ./services/pipeline
dockerfile: ai-article-generator/Dockerfile
restart: unless-stopped
depends_on:
- redis
- mongodb
environment:
- REDIS_URL=redis://redis:6379
- MONGODB_URL=mongodb://mongodb:27017
- DB_NAME=ai_writer_db
- CLAUDE_API_KEY=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
- LOG_LEVEL=INFO
volumes:
- ./services/pipeline/shared:/app/shared:ro
networks:
- site11_network
# Pipeline Monitor (optional dashboard)
pipeline-monitor:
build:
context: ./services/pipeline
dockerfile: monitor/Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_pipeline_monitor
restart: unless-stopped
depends_on:
- redis
- mongodb
ports:
- "8100:8000"
environment:
- REDIS_URL=redis://redis:6379
- MONGODB_URL=mongodb://mongodb:27017
- DB_NAME=ai_writer_db
- LOG_LEVEL=INFO
volumes:
- ./services/pipeline/shared:/app/shared:ro
networks:
- site11_network
# Pipeline Translator
pipeline-translator:
build:
context: ./services/pipeline
dockerfile: translator/Dockerfile
restart: unless-stopped
depends_on:
- redis
- mongodb
environment:
- REDIS_URL=redis://redis:6379
- MONGODB_URL=mongodb://mongodb:27017
- DB_NAME=ai_writer_db
- DEEPL_API_KEY=3abbc796-2515-44a8-972d-22dcf27ab54a
- LOG_LEVEL=INFO
volumes:
- ./services/pipeline/shared:/app/shared:ro
networks:
- site11_network
# Pipeline Language Sync Service
pipeline-language-sync:
build:
context: ./services/pipeline
dockerfile: translator/Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_pipeline_language_sync
restart: unless-stopped
depends_on:
- mongodb
environment:
- MONGODB_URL=mongodb://mongodb:27017
- DB_NAME=ai_writer_db
- DEEPL_API_KEY=3abbc796-2515-44a8-972d-22dcf27ab54a
- LOG_LEVEL=INFO
command: ["python", "language_sync.py"]
volumes:
- ./services/pipeline/shared:/app/shared:ro
- ./services/pipeline/config:/app/config:ro
networks:
- site11_network
# Pipeline Image Generator
pipeline-image-generator:
build:
context: ./services/pipeline
dockerfile: image-generator/Dockerfile
container_name: ${COMPOSE_PROJECT_NAME}_pipeline_image_generator
restart: unless-stopped
depends_on:
- redis
- mongodb
env_file:
- ./services/pipeline/.env
environment:
- REDIS_URL=redis://redis:6379
- MONGODB_URL=mongodb://mongodb:27017
- DB_NAME=ai_writer_db
- LOG_LEVEL=INFO
volumes:
- ./services/pipeline/shared:/app/shared:ro
networks:
- site11_network
networks:
site11_network:
driver: bridge
name: site11_network
# Named volumes are replaced with bind mounts in ./data/ directory
# volumes:
# mongodb_data:
# mongodb_config:
# redis_data:
# images_cache:
# zookeeper_data:
# zookeeper_logs:
# kafka_data:
# minio_data:
# files_temp:
# solr_data:

182
docs/ARCHITECTURE.md Normal file
View File

@ -0,0 +1,182 @@
# Site11 Microservices Architecture
## 시스템 아키텍처 개요
### 메시징 및 데이터 처리 시스템
#### 1. **Apache Kafka** - 통합 메시징 플랫폼
- **역할**: 이벤트 스트리밍 + 작업 큐 + 메시지 버스
- **사용 사례**:
- 서비스 간 이벤트 발행/구독
- 비동기 작업 큐 (Celery 대체)
- 사용자 활동 로그 스트리밍
- 실시간 데이터 파이프라인
- 이벤트 소싱 패턴 구현
- CQRS (Command Query Responsibility Segregation)
- 백그라운드 작업 처리
#### 2. **Redis** - 인메모리 데이터 스토어
- **역할**: 캐싱 및 세션 관리 전용
- **사용 사례**:
- API 응답 캐싱
- 사용자 세션 저장
- Rate limiting
- 실시간 리더보드/카운터
- 임시 데이터 저장
#### 3. **MongoDB** - Document Database
- **역할**: 주요 데이터 영속성
- **사용 사례**:
- 서비스별 도메인 데이터
- 유연한 스키마 관리
- 이벤트 저장소
## 서비스 통신 패턴
### 동기 통신 (REST API)
```
Client → Nginx → Console (API Gateway) → Microservice
```
- 즉각적인 응답이 필요한 경우
- CRUD 작업
- 실시간 데이터 조회
### 비동기 통신 (Kafka Events)
```
Service A → Kafka Topic → Service B, C, D
```
- 서비스 간 느슨한 결합
- 이벤트 기반 아키텍처
- 확장 가능한 처리
### 캐싱 전략 (Redis)
```
Request → Check Redis Cache → Hit? Return : Fetch from DB → Store in Redis → Return
```
- 응답 시간 개선
- 데이터베이스 부하 감소
- 세션 관리
## 이벤트 플로우 예시
### 사용자 등록 플로우
1. **API Request**: Client → Console → Users Service
2. **User Created Event**: Users Service → Kafka
3. **Event Consumers**:
- Statistics Service: 사용자 통계 업데이트
- Email Service: 환영 이메일 발송
- Analytics Service: 가입 분석
4. **Cache Update**: Redis에 사용자 정보 캐싱
### 이미지 업로드 플로우
1. **Upload Request**: Client → Console → Images Service
2. **Image Uploaded Event**: Images Service → Kafka
3. **Event Processing**:
- Thumbnail Service: 썸네일 생성
- ML Service: 이미지 분석
- Statistics Service: 업로드 통계
4. **Job Queue**: Redis/Celery로 백그라운드 처리
## Kafka Topics 구조 (예정)
### Event Topics (이벤트 스트리밍)
```
# User Domain
user.created
user.updated
user.deleted
user.login
# Image Domain
image.uploaded
image.processed
image.deleted
# Application Domain
app.registered
app.updated
app.deployed
# System Events
service.health
service.error
audit.log
```
### Task Queue Topics (작업 큐)
```
# Background Jobs
tasks.email.send
tasks.image.resize
tasks.report.generate
tasks.data.export
tasks.notification.push
# Scheduled Jobs
tasks.cleanup.expired
tasks.backup.database
tasks.analytics.aggregate
```
## Redis 사용 패턴
### 1. 캐싱 계층
- Key: `cache:users:{user_id}`
- TTL: 3600초
- 패턴: Cache-Aside
### 2. 세션 관리
- Key: `session:{token}`
- TTL: 1800초
- 데이터: 사용자 정보, 권한
### 3. Rate Limiting
- Key: `rate_limit:{user_id}:{endpoint}`
- Window: Sliding window
- Limit: 100 requests/minute
### 4. 작업 큐 (Celery)
- Queue: `celery:tasks`
- Priority Queue 지원
- Dead Letter Queue
## 구현 로드맵
### Phase 1 (현재)
- ✅ 기본 서비스 구조
- ✅ MongoDB 연동
- ✅ Redis 설치
- 🔄 JWT 인증
### Phase 2 (Step 6-7)
- Kafka 클러스터 설정
- 기본 Producer/Consumer 구현
- Event Schema 정의
- Redis 캐싱 전략 구현
### Phase 3 (Step 8+)
- Event Sourcing 패턴
- CQRS 구현
- Saga 패턴 (분산 트랜잭션)
- 모니터링 대시보드
## 기술 스택
### 메시징 & 스트리밍
- **Kafka**: Event streaming
- **Redis**: Caching, Queue, Pub/Sub
- **Confluent Schema Registry**: Schema 관리 (향후)
### 백엔드
- **FastAPI**: REST API
- **Celery**: 비동기 작업 처리
- **kafka-python**: Kafka 클라이언트
### 데이터베이스
- **MongoDB**: Document store
- **Redis**: In-memory cache
### 모니터링 (향후)
- **Kafka Manager**: Kafka 클러스터 관리
- **RedisInsight**: Redis 모니터링
- **Prometheus + Grafana**: 메트릭 수집/시각화

140
docs/DATA_PERSISTENCE.md Normal file
View File

@ -0,0 +1,140 @@
# Data Persistence Configuration
## Overview
All data services are configured to use bind mounts to local directories for data persistence. This ensures data survives container restarts and rebuilds.
## Directory Structure
```
data/
├── mongodb/ # MongoDB database files
├── redis/ # Redis persistence files
├── kafka/ # Kafka log data
├── zookeeper/ # Zookeeper data and logs
│ ├── data/
│ └── logs/
├── minio/ # MinIO object storage
├── solr/ # Solr search index
├── files-temp/ # Temporary file storage
└── images-cache/ # Image processing cache
```
## Volume Mappings
### MongoDB
- `./data/mongodb:/data/db` - Database files
- `./data/mongodb/configdb:/data/configdb` - Configuration database
### Redis
- `./data/redis:/data` - RDB snapshots and AOF logs
### Kafka
- `./data/kafka:/var/lib/kafka/data` - Message logs
### Zookeeper
- `./data/zookeeper/data:/var/lib/zookeeper/data` - Coordination data
- `./data/zookeeper/logs:/var/lib/zookeeper/log` - Transaction logs
### MinIO
- `./data/minio:/data` - Object storage buckets
### Solr
- `./data/solr:/var/solr` - Search index and configuration
### Application Caches
- `./data/files-temp:/tmp` - Temporary file processing
- `./data/images-cache:/app/cache` - Processed image cache
## Backup and Restore
### Backup All Data
```bash
# Stop services
docker-compose down
# Create backup
tar -czf backup-$(date +%Y%m%d).tar.gz data/
# Restart services
docker-compose up -d
```
### Restore Data
```bash
# Stop services
docker-compose down
# Extract backup
tar -xzf backup-YYYYMMDD.tar.gz
# Restart services
docker-compose up -d
```
### Individual Service Backups
#### MongoDB Backup
```bash
docker exec site11_mongodb mongodump --out /data/db/backup
tar -czf mongodb-backup.tar.gz data/mongodb/backup/
```
#### Redis Backup
```bash
docker exec site11_redis redis-cli BGSAVE
# Wait for completion
cp data/redis/dump.rdb redis-backup-$(date +%Y%m%d).rdb
```
## Permissions
Ensure proper permissions for data directories:
```bash
# Set appropriate permissions
chmod -R 755 data/
```
## Disk Space Monitoring
Monitor disk usage regularly:
```bash
# Check data directory size
du -sh data/*
# Check individual services
du -sh data/mongodb
du -sh data/minio
du -sh data/kafka
```
## Clean Up Old Data
### Clear Kafka Logs (older than 7 days)
```bash
docker exec site11_kafka kafka-log-dirs.sh --describe --bootstrap-server localhost:9092
```
### Clear Image Cache
```bash
rm -rf data/images-cache/*
```
### Clear Temporary Files
```bash
rm -rf data/files-temp/*
```
## Migration from Docker Volumes
If migrating from named Docker volumes to bind mounts:
1. Export data from Docker volumes:
```bash
docker run --rm -v site11_mongodb_data:/source -v $(pwd)/data/mongodb:/dest alpine cp -av /source/. /dest/
```
2. Update docker-compose.yml (already done)
3. Restart services with new configuration
## Notes
- The `data/` directory is excluded from git via .gitignore
- Ensure sufficient disk space for data growth
- Consider setting up automated backups for production
- Monitor disk I/O performance for database services

View File

@ -0,0 +1,238 @@
# Pipeline Scheduler Guide
## 개요
Pipeline Scheduler는 등록된 키워드를 주기적으로 실행하여 자동으로 뉴스를 수집하고 AI 기사를 생성하는 시스템입니다.
## 아키텍처
### 1. 구성 요소
#### 1.1 Multi-Thread Scheduler (pipeline-scheduler)
- **역할**: 키워드별 스레드 관리 및 주기적 실행
- **특징**:
- 단일 Docker 컨테이너에서 여러 스레드 동시 실행
- 각 키워드당 하나의 독립 스레드
- 30초마다 새 키워드 체크 및 스레드 관리
- **위치**: `services/pipeline/scheduler/multi_thread_scheduler.py`
#### 1.2 Keyword Manager API (keyword-manager)
- **역할**: 키워드 CRUD 및 스레드 모니터링
- **포트**: 8100
- **주요 엔드포인트**:
- `GET /threads/status` - 모든 스레드 상태 조회
- `GET /keywords` - 모든 키워드 목록
- `POST /keywords` - 새 키워드 추가
- `PUT /keywords/{keyword}` - 키워드 수정
- `DELETE /keywords/{keyword}` - 키워드 삭제
- `POST /keywords/{keyword}/activate` - 키워드 활성화
- `POST /keywords/{keyword}/deactivate` - 키워드 비활성화
- `POST /keywords/{keyword}/trigger` - 즉시 실행
- **위치**: `services/pipeline/scheduler/keyword_manager.py`
### 2. 데이터 모델
```python
class Keyword:
keyword_id: str # UUID
keyword: str # 검색 키워드
interval_minutes: int # 실행 주기 (분)
is_active: bool # 활성 상태
priority: int # 우선순위 (높을수록 우선)
rss_feeds: List[str] # RSS 피드 URL 목록
max_articles_per_run: int # 실행당 최대 기사 수
last_run: datetime # 마지막 실행 시간
next_run: datetime # 다음 실행 예정 시간
```
## 사용 방법
### 1. 서비스 시작
```bash
# 스케줄러와 매니저 시작
docker-compose up -d pipeline-scheduler keyword-manager
# 로그 확인
docker-compose logs -f pipeline-scheduler
```
### 2. 키워드 관리
#### 2.1 키워드 추가
```bash
curl -X POST http://localhost:8100/keywords \
-H "Content-Type: application/json" \
-d '{
"keyword": "딥러닝",
"interval_minutes": 60,
"priority": 1,
"rss_feeds": [],
"max_articles_per_run": 100,
"is_active": true
}'
```
#### 2.2 키워드 수정
```bash
curl -X PUT http://localhost:8100/keywords/딥러닝 \
-H "Content-Type: application/json" \
-d '{
"interval_minutes": 30,
"priority": 2
}'
```
#### 2.3 키워드 활성화/비활성화
```bash
# 활성화
curl -X POST http://localhost:8100/keywords/딥러닝/activate
# 비활성화
curl -X POST http://localhost:8100/keywords/딥러닝/deactivate
```
#### 2.4 즉시 실행
```bash
curl -X POST http://localhost:8100/keywords/딥러닝/trigger
```
#### 2.5 키워드 삭제
```bash
curl -X DELETE http://localhost:8100/keywords/딥러닝
```
### 3. 모니터링
#### 3.1 스레드 상태 확인
```bash
curl http://localhost:8100/threads/status | python3 -m json.tool
```
응답 예시:
```json
{
"total_threads": 4,
"active_threads": 4,
"threads": [
{
"keyword": "블록체인",
"keyword_id": "5c7ac9a9-c56f-4878-94ec-adb13f105c8a",
"is_active": true,
"interval_minutes": 30,
"priority": 2,
"last_run": "2025-09-15T08:05:58.807000",
"next_run": "2025-09-15T08:35:58.807000",
"thread_status": "active",
"minutes_until_next_run": 25.3
}
]
}
```
#### 3.2 키워드 목록 조회
```bash
curl http://localhost:8100/keywords | python3 -m json.tool
```
## 작동 방식
### 1. 실행 흐름
1. **키워드 스레드 시작**
- 스케줄러 시작 시 활성 키워드 로드
- 각 키워드별 독립 스레드 생성
2. **주기적 실행**
- 각 스레드는 설정된 주기마다 실행
- 실행 시 PipelineJob 생성 후 Redis 큐에 추가
- RSS 수집 → Google 검색 → AI 기사 생성 → 번역 파이프라인 자동 진행
3. **동적 스레드 관리**
- 30초마다 새 키워드 확인
- 새 키워드 추가 시 자동으로 스레드 생성
- 비활성화/삭제 시 스레드 자동 중지
### 2. 우선순위 처리
- 높은 우선순위(priority) 키워드가 먼저 처리
- Redis 큐에서 우선순위별 정렬
### 3. 오류 처리
- 각 스레드는 독립적으로 오류 처리
- 오류 발생 시 1분 대기 후 재시도
- 스레드별 error_count, last_error 추적
## 현재 설정된 키워드
| 키워드 | 실행 주기 | 우선순위 | 상태 |
|--------|-----------|----------|------|
| 블록체인 | 30분 | 2 | 활성 |
| AI | 60분 | 1 | 활성 |
| 테크놀로지 | 60분 | 1 | 활성 |
| 경제 | 60분 | 0 | 활성 |
## 주의사항
1. **스레드 관리**
- 키워드 추가/삭제는 30초 이내 자동 반영
- 스레드 상태는 keyword-manager API로 실시간 확인 가능
2. **실행 주기**
- 최소 실행 주기: 제한 없음 (권장: 30분 이상)
- interval_minutes 변경 시 다음 실행 시간 자동 재계산
3. **중복 방지**
- 동일 키워드 중복 등록 불가
- RSS 수집 시 중복 URL 자동 필터링
## 트러블슈팅
### 스레드가 시작되지 않을 때
```bash
# 스케줄러 재시작
docker-compose restart pipeline-scheduler
# 로그 확인
docker-compose logs --tail=50 pipeline-scheduler
```
### 키워드가 실행되지 않을 때
```bash
# 키워드 상태 확인
curl http://localhost:8100/keywords/키워드명 | python3 -m json.tool
# 즉시 실행 트리거
curl -X POST http://localhost:8100/keywords/키워드명/trigger
```
### MongoDB 연결 오류
```bash
# MongoDB 상태 확인
docker-compose ps mongodb
# MongoDB 재시작
docker-compose restart mongodb
```
## 파일 구조
```
services/pipeline/scheduler/
├── multi_thread_scheduler.py # 멀티스레드 스케줄러
├── keyword_manager.py # 키워드 관리 API
├── single_keyword_scheduler.py # (deprecated) 단일 키워드 스케줄러
├── requirements.txt # Python 의존성
└── Dockerfile # Docker 이미지 정의
services/pipeline/shared/
└── models.py # Keyword, PipelineJob 모델 정의
```
## 향후 개선사항
1. **웹 대시보드**: 실시간 모니터링 UI
2. **알림 시스템**: 오류 발생 시 이메일/Slack 알림
3. **통계 기능**: 키워드별 수집 통계 및 분석
4. **스케줄 템플릿**: 자주 사용하는 설정 저장/불러오기
5. **백업/복구**: 키워드 설정 백업 및 복구 기능

401
docs/PLAN.md Normal file
View File

@ -0,0 +1,401 @@
# Microservices Architecture Implementation Plan
## Project Overview
Build a microservices-based platform with a central Console service acting as an orchestrator and API Gateway, managing multiple domain-specific services.
## Architecture Decision
**Selected Pattern**: API Gateway Pattern with Console as the central orchestrator
- Console handles authentication, routing, and monitoring
- Each microservice focuses on domain-specific logic
- Services communicate via REST APIs and Redis pub/sub
## Progressive Implementation Strategy
### Step-by-Step Approach
큰 그림을 먼저 구성하고, 핵심 기능부터 점진적으로 확장하는 전략
## Implementation Phases
### Step 1: Minimal Foundation (Day 1-2)
**목표**: 가장 기본적인 구조 확립
```
site11/
├── docker-compose.yml # 최소 구성 (Console만)
├── console/
│ └── backend/
│ └── main.py # Hello World API
└── README.md
```
**Tasks**:
- [ ] 간단한 docker-compose.yml 생성
- [ ] Console FastAPI "Hello World"
- [ ] 기본 health check endpoint
- [ ] Docker로 실행 확인
### Step 2: Add First Service (Day 3-4)
**목표**: Console과 하나의 서비스 연결
```
site11/
├── docker-compose.yml
├── console/
│ └── backend/
│ └── main.py # Gateway 역할 추가
└── services/
└── users/
└── backend/
└── main.py # Users 서비스
```
**Tasks**:
- [ ] Users 서비스 생성
- [ ] Console에서 Users로 라우팅
- [ ] 서비스 간 통신 테스트
- [ ] 간단한 CRUD API
### Step 3: Database Integration (Day 5-6)
**목표**: MongoDB 연결 및 기본 데이터 저장
**Tasks**:
- [ ] MongoDB 컨테이너 추가
- [ ] Console과 Users 서비스 DB 연결
- [ ] 기본 데이터 모델 생성
- [ ] 실제 데이터 CRUD 테스트
### Step 4: Frontend Skeleton (Week 2)
**목표**: 최소한의 UI 구성
**Tasks**:
- [ ] Console Frontend 생성 (React + Vite)
- [ ] 기본 레이아웃
- [ ] 서비스 상태 표시
- [ ] Nginx 설정
### Step 5: Authentication Basic (Week 2)
**목표**: 간단한 인증 시스템
**Tasks**:
- [ ] JWT 토큰 생성
- [ ] Login endpoint
- [ ] Token 검증 미들웨어
- [ ] Protected routes
### Step 6: Second Service (Week 3)
**목표**: 두 번째 서비스 추가로 패턴 확립
**Tasks**:
- [ ] OAuth 또는 Images 서비스 추가
- [ ] Console 라우팅 확장
- [ ] 서비스 간 통신 패턴 확립
- [ ] Service registry 기초
### Step 7: Service Communication (Week 3)
**목표**: 서비스 간 통신 패턴 구현
**Tasks**:
- [ ] Redis pub/sub 설정
- [ ] Event 기반 통신 예제
- [ ] Service discovery 구현
- [ ] Health check 자동화
### Step 8: Gradual Service Addition (Week 4-5)
**목표**: 나머지 서비스 점진적 추가
**각 서비스별로**:
- [ ] 기본 구조 생성
- [ ] Console 연결
- [ ] 핵심 API 구현
- [ ] Frontend 컴포넌트 추가
## 현재 시작점 (NOW)
### 즉시 시작할 수 있는 첫 걸음
#### 1. 최소 Docker 환경 구성
```bash
# 실행 명령
docker-compose up -d console
curl http://localhost:8000/health
```
#### 2. Console 서비스만으로 시작
- Health endpoint
- 간단한 API Gateway 구조
- 서비스 등록 준비
#### 3. 하나씩 추가하며 테스트
- Users 서비스 하나만 추가
- 통신 확인
- 패턴 확립 후 확장
### 핵심 원칙
1. **작동하는 코드 우선** - 완벽한 설계보다 동작하는 MVP
2. **점진적 복잡도** - 간단한 것부터 시작해서 기능 추가
3. **빠른 피드백** - 각 단계마다 실행하고 확인
4. **패턴 확립** - 첫 서비스로 패턴을 만들고 복제
---
## 상세 구현 계획 (참고용)
### Phase 1: Foundation Setup (Week 1)
#### Goals
- Set up project structure
- Configure Docker environment
- Establish basic infrastructure
#### Tasks
- [ ] Initialize Git repository
- [ ] Create Docker Compose configuration
- [ ] Set up Nginx reverse proxy
- [ ] Configure MongoDB and Redis containers
- [ ] Create base directory structure for all services
#### Deliverables
- Working Docker environment
- Basic networking between containers
- Database and cache ready
### Phase 2: Console Service - Core (Week 2)
#### Goals
- Implement Console as API Gateway
- Set up authentication system
- Create service registry
#### Tasks
- [ ] Console Backend
- [ ] FastAPI application setup
- [ ] JWT authentication implementation
- [ ] Service registry and discovery
- [ ] API routing mechanism
- [ ] Health check endpoints
- [ ] Console Frontend
- [ ] React + Vite setup
- [ ] Login/Register pages
- [ ] Admin dashboard layout
- [ ] Service status dashboard
#### Deliverables
- Working authentication system
- Basic API Gateway functionality
- Service health monitoring dashboard
### Phase 3: OAuth Service (Week 3)
#### Goals
- Centralized authentication service
- OAuth2 implementation
- User session management
#### Tasks
- [ ] OAuth Backend
- [ ] OAuth2 server implementation
- [ ] Token generation and validation
- [ ] User authentication endpoints
- [ ] Integration with Console
- [ ] OAuth Frontend
- [ ] OAuth consent screens
- [ ] Token management UI
- [ ] Application registration
#### Deliverables
- OAuth2 server
- Token-based authentication
- Integration with Console
### Phase 4: Users Service (Week 4)
#### Goals
- User management microservice
- Profile management
- User data CRUD operations
#### Tasks
- [ ] Users Backend
- [ ] User model and database schema
- [ ] CRUD APIs for user management
- [ ] Profile management endpoints
- [ ] Integration with OAuth service
- [ ] Users Frontend
- [ ] User list and search
- [ ] Profile editing interface
- [ ] User details view
#### Deliverables
- Complete user management system
- Profile management features
- Admin user interface
### Phase 5: Core Microservices (Weeks 5-6)
#### Goals
- Implement remaining core services
- Establish inter-service communication
#### Services to Implement
1. **Images Service**
- Image upload/download
- Image processing
- Storage management
2. **Applications Service**
- Application registration
- Configuration management
- Version control
3. **Data Service**
- Data import/export
- Data transformation
- API for data access
4. **Statistics Service**
- Metrics collection
- Analytics dashboard
- Report generation
#### Tasks per Service
- [ ] Backend implementation
- [ ] Domain models
- [ ] Business logic
- [ ] REST APIs
- [ ] Event publishing
- [ ] Frontend implementation
- [ ] Service-specific UI
- [ ] Integration with Console
- [ ] Dashboard widgets
### Phase 6: Integration & Testing (Week 7)
#### Goals
- End-to-end integration
- Performance optimization
- Security hardening
#### Tasks
- [ ] Integration Testing
- [ ] Service communication tests
- [ ] Load testing
- [ ] Security testing
- [ ] Optimization
- [ ] Redis caching implementation
- [ ] Database indexing
- [ ] API response optimization
- [ ] Documentation
- [ ] API documentation (OpenAPI)
- [ ] Deployment guide
- [ ] Developer documentation
#### Deliverables
- Fully integrated system
- Performance benchmarks
- Complete documentation
### Phase 7: Monitoring & DevOps (Week 8)
#### Goals
- Production readiness
- Monitoring and alerting
- CI/CD pipeline
#### Tasks
- [ ] Monitoring Setup
- [ ] Prometheus metrics
- [ ] Grafana dashboards
- [ ] Log aggregation (ELK stack)
- [ ] DevOps
- [ ] GitHub Actions CI/CD
- [ ] Automated testing
- [ ] Docker image optimization
- [ ] Production Configuration
- [ ] Environment variables
- [ ] Secrets management
- [ ] Backup strategies
#### Deliverables
- Production-ready deployment
- Monitoring dashboards
- Automated deployment pipeline
## Technical Implementation Details
### Service Communication Flow
```
Client Request → Nginx → Console (API Gateway) → Microservice
Authentication Check
Request Routing
Response Aggregation
```
### Database Strategy
```
MongoDB Instance
├── console_db # Console service data
├── users_db # Users service data
├── oauth_db # OAuth tokens and sessions
├── images_db # Image metadata
├── applications_db # Application data
├── data_db # Generic data storage
└── statistics_db # Analytics data
```
### API Versioning Strategy
- All APIs follow `/api/v1/` pattern
- Version in URL path for major versions
- Header-based versioning for minor updates
### Security Implementation
1. **Authentication Flow**
- User login → OAuth service
- OAuth service issues JWT
- Console validates JWT on each request
- Console forwards validated requests to services
2. **Service-to-Service Auth**
- Internal service tokens
- mTLS for production
- Network isolation via Docker networks
### Development Workflow
1. **Local Development**
```bash
docker-compose up -d [service-name]
docker-compose logs -f [service-name]
```
2. **Testing**
```bash
docker-compose exec [service-name] pytest
```
3. **Deployment**
```bash
docker-compose build
docker-compose up -d
```
## Success Criteria
- [ ] All services independently deployable
- [ ] Console successfully routes to all services
- [ ] Authentication works across all services
- [ ] Health monitoring shows all services green
- [ ] Load testing shows <100ms p95 latency
- [ ] Zero downtime deployments possible
## Risk Mitigation
1. **Service Failure**: Circuit breakers in Console
2. **Data Consistency**: Event sourcing for critical operations
3. **Performance**: Redis caching layer
4. **Security**: Regular security audits, dependency updates
## Timeline Summary
- **Week 1**: Foundation and infrastructure
- **Week 2**: Console core implementation
- **Week 3**: OAuth service
- **Week 4**: Users service
- **Weeks 5-6**: Remaining microservices
- **Week 7**: Integration and testing
- **Week 8**: Monitoring and production setup
## Next Steps
1. Review and approve plan
2. Set up Git repository
3. Begin Phase 1 implementation
4. Schedule weekly progress reviews

127
docs/PROGRESS.md Normal file
View File

@ -0,0 +1,127 @@
# Progress Tracking & Context Management
## Purpose
이 파일은 Claude의 컨텍스트가 리셋되어도 빠르게 현재 진행 상황을 파악하고 이어서 작업할 수 있도록 돕는 체크포인트 문서입니다.
## Current Status
- **Date Started**: 2025-09-09
- **Current Phase**: Step 3 Complete ✅
- **Next Action**: Step 4 - Frontend Skeleton
## Completed Checkpoints
✅ Project structure planning (CLAUDE.md)
✅ Implementation plan created (docs/PLAN.md)
✅ Progressive approach defined
✅ Step 1: Minimal Foundation - Docker + Console Hello World
- docker-compose.yml created
- console/backend with FastAPI
- Running on port 8011
✅ Step 2: Add First Service (Users)
- Users service with CRUD operations
- Console API Gateway routing to Users
- Service communication verified
- Test: curl http://localhost:8011/api/users/users
✅ Step 3: Database Integration
- MongoDB and Redis containers added
- Users service using MongoDB with Beanie ODM
- Data persistence verified
- MongoDB IDs: 68c126c0bbbe52be68495933
## Active Working Files
```
현재 작업 중인 주요 파일:
- /docs/PLAN.md (구현 계획)
- /CLAUDE.md (아키텍처 가이드)
- /docs/PROGRESS.md (이 파일)
```
## Next Immediate Steps
```bash
# 다음 작업 시작 명령
# Step 1: Create docker-compose.yml
# Step 2: Create console/backend/main.py
# Step 3: Test with docker-compose up
```
## Code Snippets Ready to Use
### 1. Minimal docker-compose.yml
```yaml
version: '3.8'
services:
console:
build: ./console/backend
ports:
- "8000:8000"
environment:
- ENV=development
```
### 2. Console main.py starter
```python
from fastapi import FastAPI
app = FastAPI(title="Console API Gateway")
@app.get("/health")
async def health():
return {"status": "healthy", "service": "console"}
```
## Important Decisions Made
1. **Architecture**: API Gateway Pattern with Console as orchestrator
2. **Tech Stack**: FastAPI + React + MongoDB + Redis + Docker
3. **Approach**: Progressive implementation (simple to complex)
4. **First Service**: Users service after Console
## Questions to Ask When Resuming
새로운 세션에서 이어서 작업할 때 확인할 사항:
1. "PROGRESS.md 파일을 확인했나요?"
2. "마지막으로 완료한 Step은 무엇인가요?"
3. "현재 에러나 블로킹 이슈가 있나요?"
## Git Commits Pattern
각 Step 완료 시 커밋 메시지:
```
Step X: [간단한 설명]
- 구현 내용 1
- 구현 내용 2
```
## Directory Structure Snapshot
```
site11/
├── CLAUDE.md ✅ Created
├── docs/
│ ├── PLAN.md ✅ Created
│ └── PROGRESS.md ✅ Created (this file)
├── console/ 🔄 Next
│ └── backend/
│ └── main.py
└── docker-compose.yml 🔄 Next
```
## Context Recovery Commands
새 세션에서 빠르게 상황 파악하기:
```bash
# 1. 현재 구조 확인
ls -la
# 2. 진행 상황 확인
cat docs/PROGRESS.md
# 3. 다음 단계 확인
grep "Step" docs/PLAN.md | head -5
# 4. 실행 중인 컨테이너 확인
docker ps
```
## Error Log
문제 발생 시 여기에 기록:
- (아직 없음)
## Notes for Next Session
- Step 1부터 시작
- docker-compose.yml 생성 필요
- console/backend/main.py 생성 필요
- 모든 문서 파일은 대문자.md 형식으로 생성 (예: README.md, SETUP.md)

170
docs/TEST_AUTH.md Normal file
View File

@ -0,0 +1,170 @@
# 인증 시스템 테스트 가이드
## 테스트 계정
- **관리자**: admin / admin123
- **일반 사용자**: user / user123
## 1. Terminal에서 테스트
### 로그인 테스트
```bash
# 관리자로 로그인
curl -X POST http://localhost:8011/api/auth/login \
-H "Content-Type: application/x-www-form-urlencoded" \
-d "username=admin&password=admin123"
# 응답 예시:
# {"access_token":"eyJhbGci...","token_type":"bearer"}
```
### 토큰 저장 및 사용
```bash
# 토큰을 변수에 저장
export TOKEN="eyJhbGci..." # 위에서 받은 토큰
# 인증된 요청 - 사용자 정보 조회
curl -X GET http://localhost:8011/api/auth/me \
-H "Authorization: Bearer $TOKEN"
# 인증된 요청 - 보호된 엔드포인트
curl -X GET http://localhost:8011/api/protected \
-H "Authorization: Bearer $TOKEN"
# 인증된 요청 - Users 서비스 접근
curl -X GET http://localhost:8011/api/users/ \
-H "Authorization: Bearer $TOKEN"
```
### 로그아웃
```bash
curl -X POST http://localhost:8011/api/auth/logout \
-H "Authorization: Bearer $TOKEN"
```
## 2. Postman/Insomnia에서 테스트
### Postman 설정
1. **로그인 요청**
- Method: POST
- URL: `http://localhost:8011/api/auth/login`
- Body: x-www-form-urlencoded
- username: admin
- password: admin123
2. **토큰 사용**
- Authorization 탭에서 Type: Bearer Token 선택
- Token 필드에 받은 토큰 붙여넣기
## 3. Python 스크립트로 테스트
```python
import requests
# 로그인
login_response = requests.post(
"http://localhost:8011/api/auth/login",
data={"username": "admin", "password": "admin123"}
)
token = login_response.json()["access_token"]
# 인증된 요청
headers = {"Authorization": f"Bearer {token}"}
me_response = requests.get(
"http://localhost:8011/api/auth/me",
headers=headers
)
print(me_response.json())
# Users 서비스 접근
users_response = requests.get(
"http://localhost:8011/api/users/",
headers=headers
)
print(users_response.json())
```
## 4. JavaScript (브라우저 콘솔)에서 테스트
```javascript
// 로그인
const loginResponse = await fetch('http://localhost:8011/api/auth/login', {
method: 'POST',
headers: {'Content-Type': 'application/x-www-form-urlencoded'},
body: 'username=admin&password=admin123'
});
const { access_token } = await loginResponse.json();
console.log('Token:', access_token);
// 인증된 요청
const meResponse = await fetch('http://localhost:8011/api/auth/me', {
headers: {'Authorization': `Bearer ${access_token}`}
});
const userData = await meResponse.json();
console.log('User:', userData);
```
## 5. Frontend에서 테스트 (React)
브라우저에서 http://localhost:3000 접속 후 개발자 도구 콘솔에서:
```javascript
// 로그인 함수
async function testLogin() {
const response = await fetch('/api/auth/login', {
method: 'POST',
headers: {'Content-Type': 'application/x-www-form-urlencoded'},
body: 'username=admin&password=admin123'
});
const data = await response.json();
localStorage.setItem('token', data.access_token);
console.log('Logged in!', data);
return data.access_token;
}
// 인증 테스트
async function testAuth() {
const token = localStorage.getItem('token');
const response = await fetch('/api/auth/me', {
headers: {'Authorization': `Bearer ${token}`}
});
const data = await response.json();
console.log('User info:', data);
}
// 실행
await testLogin();
await testAuth();
```
## 오류 테스트
### 잘못된 비밀번호
```bash
curl -X POST http://localhost:8011/api/auth/login \
-H "Content-Type: application/x-www-form-urlencoded" \
-d "username=admin&password=wrong"
# 응답: 401 Unauthorized
```
### 토큰 없이 보호된 엔드포인트 접근
```bash
curl -X GET http://localhost:8011/api/auth/me
# 응답: 401 Unauthorized
```
### 잘못된 토큰
```bash
curl -X GET http://localhost:8011/api/auth/me \
-H "Authorization: Bearer invalid_token"
# 응답: 401 Unauthorized
```
## 토큰 정보
- **유효 기간**: 30분 (환경 변수 ACCESS_TOKEN_EXPIRE_MINUTES로 설정 가능)
- **알고리즘**: HS256
- **페이로드**: username 정보 포함
## 다음 단계
Frontend에 로그인 페이지를 추가하면 UI에서 직접 테스트 가능합니다.

36
generated_article.json Normal file
View File

@ -0,0 +1,36 @@
{
"_id": "68c5798162bde7a1947d35a7",
"article_id": "17ee889c-a01e-4791-9f51-9336074c842b",
"job_id": "616c1c65-6b43-42a8-98e0-9547208106c8",
"keyword_id": "test_starcraft_001",
"keyword": "스타크래프트",
"title": "Is StarCraft a new tipping point after 15 years?",
"content": "**Nexon secures new StarCraft development rights***.\n\nSouth Korean gaming company Nexon has announced that it has won the rights to develop a new StarCraft game, beating out Blizzard in a bidding war. This is expected to mark a new turning point for the StarCraft IP. Nexon also secured the distribution rights to Blizzard's mobile game called Overwatch 3.\n\n**StarCraft 2, still the benchmark for RTS games***.\n\nFifteen years after its release, StarCraft 2 is still considered the pinnacle of the real-time strategy (RTS) gaming genre. According to a recently published analysis, today's RTS games are still being developed with the same standards of polish and playability that StarCraft 2 set. Blizzard recently released the 5.0.14 PTR update for StarCraft 2, demonstrating its continued support for the game.\n\n**Breaking records on the esports stage\n\nAt the Esports World Cup 2025, Joel \"Serral\" Larsson of Finland won the StarCraft 2 category, setting a new viewership record. This shows that StarCraft 2 still has a strong appeal to gamers around the world.\n\n**The failure of Stormgate, a classic for sequels\n\nOn the other hand, the spiritual successor to StarCraft, Stormgate, created by ex-Blizzard developers, has been criticized as a flop. The developers controversially blamed gamers for its failure. This shows how difficult it is to replicate StarCraft's winning formula.\n\n**Current events and outlook\n\nNexon's acquisition of StarCraft development rights is significant for the Korean gaming industry. It increases the likelihood of new StarCraft titles being developed in Korea, the home of the franchise. However, it will be a challenge to create a game that surpasses StarCraft 2, which has been a standard in the industry for 15 years.",
"summary": "Nexon's acquisition of the new StarCraft development rights marks a turning point for the StarCraft IP, which has been a benchmark in RTS gaming for 15 years.",
"source_items": [],
"images": [
"https://replicate.delivery/xezq/tYOKIl3SG35WPBVBg3ipefCOLUJPm9FCXwyOOrppS1WztyUVA/out-0.png"
],
"categories": [
"游戏",
"Esports"
],
"tags": [
"StarCraft",
"Nexon",
"RTS games",
"Blizzard",
"Esports"
],
"created_at": "2025-09-13T14:02:41.857000",
"pipeline_stages": [
"rss_collection",
"search_enrichment",
"ai_article_generation",
"image_generation",
"translation"
],
"processing_time": 19.447839,
"language": "en",
"ref_news_id": "20a6bb85-8c61-41db-82fc-d52e0d88204d"
}

View File

@ -0,0 +1,27 @@
FROM python:3.11-slim
WORKDIR /app
# Install system dependencies for Pillow and file type detection
RUN apt-get update && apt-get install -y \
gcc \
libmagic1 \
libjpeg-dev \
zlib1g-dev \
&& rm -rf /var/lib/apt/lists/*
# Copy requirements first for better caching
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
# Copy application code
COPY . .
# Create directories for thumbnails cache
RUN mkdir -p /tmp/thumbnails
# Expose port
EXPOSE 8000
# Run the application
CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "8000", "--reload"]

View File

@ -0,0 +1,247 @@
"""
File Processor for handling file uploads and processing
"""
import hashlib
import mimetypes
from datetime import datetime
from typing import Dict, Any, Optional
import logging
import uuid
from fastapi import UploadFile
from models import FileType, FileStatus
logger = logging.getLogger(__name__)
class FileProcessor:
def __init__(self, minio_client, metadata_manager, thumbnail_generator):
self.minio_client = minio_client
self.metadata_manager = metadata_manager
self.thumbnail_generator = thumbnail_generator
def _determine_file_type(self, content_type: str) -> FileType:
"""Determine file type from content type"""
if content_type.startswith('image/'):
return FileType.IMAGE
elif content_type.startswith('video/'):
return FileType.VIDEO
elif content_type.startswith('audio/'):
return FileType.AUDIO
elif content_type in ['application/pdf', 'application/msword',
'application/vnd.openxmlformats-officedocument',
'text/plain', 'text/html', 'text/csv']:
return FileType.DOCUMENT
elif content_type in ['application/zip', 'application/x-rar-compressed',
'application/x-tar', 'application/gzip']:
return FileType.ARCHIVE
else:
return FileType.OTHER
def _calculate_file_hash(self, file_data: bytes) -> str:
"""Calculate SHA256 hash of file data"""
return hashlib.sha256(file_data).hexdigest()
async def process_upload(self, file: UploadFile, user_id: str,
bucket: str = "default",
public: bool = False,
generate_thumbnail: bool = True,
tags: Optional[Dict[str, Any]] = None) -> Dict[str, Any]:
"""Process file upload"""
try:
# Read file data
file_data = await file.read()
file_size = len(file_data)
# Get content type
content_type = file.content_type or mimetypes.guess_type(file.filename)[0] or 'application/octet-stream'
# Generate file ID and object name
file_id = str(uuid.uuid4())
timestamp = datetime.now().strftime('%Y%m%d')
file_extension = file.filename.split('.')[-1] if '.' in file.filename else ''
object_name = f"{timestamp}/{user_id}/{file_id}.{file_extension}" if file_extension else f"{timestamp}/{user_id}/{file_id}"
# Calculate file hash
file_hash = self._calculate_file_hash(file_data)
# Check for duplicates
duplicates = await self.metadata_manager.find_duplicate_files(file_hash)
if duplicates and not public: # Allow duplicates for public files
# Return existing file info
existing = duplicates[0]
logger.info(f"Duplicate file detected: {existing['id']}")
return {
"file_id": existing["id"],
"filename": existing["filename"],
"size": existing["size"],
"content_type": existing["content_type"],
"file_type": existing["file_type"],
"bucket": existing["bucket"],
"public": existing["public"],
"has_thumbnail": existing.get("has_thumbnail", False),
"thumbnail_url": existing.get("thumbnail_url"),
"created_at": existing["created_at"],
"duplicate": True
}
# Upload to MinIO
upload_result = await self.minio_client.upload_file(
bucket=bucket,
object_name=object_name,
file_data=file_data,
content_type=content_type,
metadata={
"user_id": user_id,
"original_name": file.filename,
"upload_date": datetime.now().isoformat()
}
)
# Determine file type
file_type = self._determine_file_type(content_type)
# Generate thumbnail if applicable
has_thumbnail = False
thumbnail_url = None
if generate_thumbnail and file_type == FileType.IMAGE:
thumbnail_data = await self.thumbnail_generator.generate_thumbnail(
file_data=file_data,
content_type=content_type
)
if thumbnail_data:
has_thumbnail = True
# Generate multiple sizes
await self.thumbnail_generator.generate_multiple_sizes(
file_data=file_data,
content_type=content_type,
file_id=file_id
)
if public:
thumbnail_url = await self.minio_client.generate_presigned_download_url(
bucket="thumbnails",
object_name=f"thumbnails/{file_id}_medium.jpg",
expires_in=86400 * 30 # 30 days
)
# Create metadata
metadata = {
"id": file_id,
"filename": file.filename,
"original_name": file.filename,
"size": file_size,
"content_type": content_type,
"file_type": file_type.value,
"bucket": bucket,
"object_name": object_name,
"user_id": user_id,
"hash": file_hash,
"public": public,
"has_thumbnail": has_thumbnail,
"thumbnail_url": thumbnail_url,
"tags": tags or {},
"metadata": {
"etag": upload_result.get("etag"),
"version_id": upload_result.get("version_id")
}
}
# Save metadata to database
await self.metadata_manager.create_file_metadata(metadata)
# Generate download URL if public
download_url = None
if public:
download_url = await self.minio_client.generate_presigned_download_url(
bucket=bucket,
object_name=object_name,
expires_in=86400 * 30 # 30 days
)
logger.info(f"File uploaded successfully: {file_id}")
return {
"file_id": file_id,
"filename": file.filename,
"size": file_size,
"content_type": content_type,
"file_type": file_type.value,
"bucket": bucket,
"public": public,
"has_thumbnail": has_thumbnail,
"thumbnail_url": thumbnail_url,
"download_url": download_url,
"created_at": datetime.now()
}
except Exception as e:
logger.error(f"File processing error: {e}")
raise
async def process_large_file(self, file: UploadFile, user_id: str,
bucket: str = "default",
chunk_size: int = 1024 * 1024 * 5) -> Dict[str, Any]:
"""Process large file upload in chunks"""
try:
file_id = str(uuid.uuid4())
timestamp = datetime.now().strftime('%Y%m%d')
file_extension = file.filename.split('.')[-1] if '.' in file.filename else ''
object_name = f"{timestamp}/{user_id}/{file_id}.{file_extension}"
# Initialize multipart upload
hasher = hashlib.sha256()
total_size = 0
# Process file in chunks
chunks = []
while True:
chunk = await file.read(chunk_size)
if not chunk:
break
chunks.append(chunk)
hasher.update(chunk)
total_size += len(chunk)
# Combine chunks and upload
file_data = b''.join(chunks)
file_hash = hasher.hexdigest()
# Upload to MinIO
content_type = file.content_type or 'application/octet-stream'
await self.minio_client.upload_file(
bucket=bucket,
object_name=object_name,
file_data=file_data,
content_type=content_type
)
# Create metadata
metadata = {
"id": file_id,
"filename": file.filename,
"original_name": file.filename,
"size": total_size,
"content_type": content_type,
"file_type": self._determine_file_type(content_type).value,
"bucket": bucket,
"object_name": object_name,
"user_id": user_id,
"hash": file_hash,
"public": False,
"has_thumbnail": False
}
await self.metadata_manager.create_file_metadata(metadata)
return {
"file_id": file_id,
"filename": file.filename,
"size": total_size,
"message": "Large file uploaded successfully"
}
except Exception as e:
logger.error(f"Large file processing error: {e}")
raise

View File

@ -0,0 +1,541 @@
"""
File Management Service - S3-compatible Object Storage with MinIO
"""
from fastapi import FastAPI, File, UploadFile, HTTPException, Depends, Query, Form
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import StreamingResponse, FileResponse
import uvicorn
from datetime import datetime, timedelta
from typing import Optional, List, Dict, Any
import asyncio
import os
import hashlib
import magic
import io
from contextlib import asynccontextmanager
import logging
from pathlib import Path
import json
# Import custom modules
from models import FileMetadata, FileUploadResponse, FileListResponse, StorageStats
from minio_client import MinIOManager
from thumbnail_generator import ThumbnailGenerator
from metadata_manager import MetadataManager
from file_processor import FileProcessor
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
# Global instances
minio_manager = None
thumbnail_generator = None
metadata_manager = None
file_processor = None
@asynccontextmanager
async def lifespan(app: FastAPI):
# Startup
global minio_manager, thumbnail_generator, metadata_manager, file_processor
try:
# Initialize MinIO client
minio_manager = MinIOManager(
endpoint=os.getenv("MINIO_ENDPOINT", "minio:9000"),
access_key=os.getenv("MINIO_ACCESS_KEY", "minioadmin"),
secret_key=os.getenv("MINIO_SECRET_KEY", "minioadmin"),
secure=os.getenv("MINIO_SECURE", "false").lower() == "true"
)
await minio_manager.initialize()
logger.info("MinIO client initialized")
# Initialize Metadata Manager (MongoDB)
metadata_manager = MetadataManager(
mongodb_url=os.getenv("MONGODB_URL", "mongodb://mongodb:27017"),
database=os.getenv("FILES_DB_NAME", "files_db")
)
await metadata_manager.connect()
logger.info("Metadata manager connected to MongoDB")
# Initialize Thumbnail Generator
thumbnail_generator = ThumbnailGenerator(
minio_client=minio_manager,
cache_dir="/tmp/thumbnails"
)
logger.info("Thumbnail generator initialized")
# Initialize File Processor
file_processor = FileProcessor(
minio_client=minio_manager,
metadata_manager=metadata_manager,
thumbnail_generator=thumbnail_generator
)
logger.info("File processor initialized")
except Exception as e:
logger.error(f"Failed to start File service: {e}")
raise
yield
# Shutdown
if metadata_manager:
await metadata_manager.close()
logger.info("File service shutdown complete")
app = FastAPI(
title="File Management Service",
description="S3-compatible object storage with MinIO",
version="1.0.0",
lifespan=lifespan
)
# CORS middleware
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
@app.get("/")
async def root():
return {
"service": "File Management Service",
"status": "running",
"timestamp": datetime.now().isoformat()
}
@app.get("/health")
async def health_check():
return {
"status": "healthy",
"service": "files",
"components": {
"minio": "connected" if minio_manager and minio_manager.is_connected else "disconnected",
"mongodb": "connected" if metadata_manager and metadata_manager.is_connected else "disconnected",
"thumbnail_generator": "ready" if thumbnail_generator else "not_initialized"
},
"timestamp": datetime.now().isoformat()
}
# File Upload Endpoints
@app.post("/api/files/upload")
async def upload_file(
file: UploadFile = File(...),
user_id: str = Form(...),
bucket: str = Form("default"),
public: bool = Form(False),
generate_thumbnail: bool = Form(True),
tags: Optional[str] = Form(None)
):
"""Upload a file to object storage"""
try:
# Validate file
if not file.filename:
raise HTTPException(status_code=400, detail="No file provided")
# Process file upload
result = await file_processor.process_upload(
file=file,
user_id=user_id,
bucket=bucket,
public=public,
generate_thumbnail=generate_thumbnail,
tags=json.loads(tags) if tags else {}
)
return FileUploadResponse(**result)
except Exception as e:
logger.error(f"File upload error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.post("/api/files/upload-multiple")
async def upload_multiple_files(
files: List[UploadFile] = File(...),
user_id: str = Form(...),
bucket: str = Form("default"),
public: bool = Form(False)
):
"""Upload multiple files"""
try:
results = []
for file in files:
result = await file_processor.process_upload(
file=file,
user_id=user_id,
bucket=bucket,
public=public,
generate_thumbnail=True
)
results.append(result)
return {
"uploaded": len(results),
"files": results
}
except Exception as e:
logger.error(f"Multiple file upload error: {e}")
raise HTTPException(status_code=500, detail=str(e))
# File Retrieval Endpoints
@app.get("/api/files/{file_id}")
async def get_file(file_id: str):
"""Get file by ID"""
try:
# Get metadata
metadata = await metadata_manager.get_file_metadata(file_id)
if not metadata:
raise HTTPException(status_code=404, detail="File not found")
# Get file from MinIO
file_stream = await minio_manager.get_file(
bucket=metadata["bucket"],
object_name=metadata["object_name"]
)
return StreamingResponse(
file_stream,
media_type=metadata.get("content_type", "application/octet-stream"),
headers={
"Content-Disposition": f'attachment; filename="{metadata["filename"]}"'
}
)
except HTTPException:
raise
except Exception as e:
logger.error(f"File retrieval error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.get("/api/files/{file_id}/metadata")
async def get_file_metadata(file_id: str):
"""Get file metadata"""
try:
metadata = await metadata_manager.get_file_metadata(file_id)
if not metadata:
raise HTTPException(status_code=404, detail="File not found")
return FileMetadata(**metadata)
except HTTPException:
raise
except Exception as e:
logger.error(f"Metadata retrieval error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.get("/api/files/{file_id}/thumbnail")
async def get_thumbnail(
file_id: str,
width: int = Query(200, ge=50, le=1000),
height: int = Query(200, ge=50, le=1000)
):
"""Get file thumbnail"""
try:
# Get metadata
metadata = await metadata_manager.get_file_metadata(file_id)
if not metadata:
raise HTTPException(status_code=404, detail="File not found")
# Check if file has thumbnail
if not metadata.get("has_thumbnail"):
raise HTTPException(status_code=404, detail="No thumbnail available")
# Get or generate thumbnail
thumbnail = await thumbnail_generator.get_thumbnail(
file_id=file_id,
bucket=metadata["bucket"],
object_name=metadata["object_name"],
width=width,
height=height
)
return StreamingResponse(
io.BytesIO(thumbnail),
media_type="image/jpeg"
)
except HTTPException:
raise
except Exception as e:
logger.error(f"Thumbnail retrieval error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.get("/api/files/{file_id}/download")
async def download_file(file_id: str):
"""Download file with proper headers"""
try:
# Get metadata
metadata = await metadata_manager.get_file_metadata(file_id)
if not metadata:
raise HTTPException(status_code=404, detail="File not found")
# Update download count
await metadata_manager.increment_download_count(file_id)
# Get file from MinIO
file_stream = await minio_manager.get_file(
bucket=metadata["bucket"],
object_name=metadata["object_name"]
)
return StreamingResponse(
file_stream,
media_type=metadata.get("content_type", "application/octet-stream"),
headers={
"Content-Disposition": f'attachment; filename="{metadata["filename"]}"',
"Content-Length": str(metadata["size"])
}
)
except HTTPException:
raise
except Exception as e:
logger.error(f"File download error: {e}")
raise HTTPException(status_code=500, detail=str(e))
# File Management Endpoints
@app.delete("/api/files/{file_id}")
async def delete_file(file_id: str, user_id: str):
"""Delete a file"""
try:
# Get metadata
metadata = await metadata_manager.get_file_metadata(file_id)
if not metadata:
raise HTTPException(status_code=404, detail="File not found")
# Check ownership
if metadata["user_id"] != user_id:
raise HTTPException(status_code=403, detail="Permission denied")
# Delete from MinIO
await minio_manager.delete_file(
bucket=metadata["bucket"],
object_name=metadata["object_name"]
)
# Delete thumbnail if exists
if metadata.get("has_thumbnail"):
await thumbnail_generator.delete_thumbnail(file_id)
# Delete metadata
await metadata_manager.delete_file_metadata(file_id)
return {"status": "deleted", "file_id": file_id}
except HTTPException:
raise
except Exception as e:
logger.error(f"File deletion error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.patch("/api/files/{file_id}")
async def update_file_metadata(
file_id: str,
user_id: str,
updates: Dict[str, Any]
):
"""Update file metadata"""
try:
# Get existing metadata
metadata = await metadata_manager.get_file_metadata(file_id)
if not metadata:
raise HTTPException(status_code=404, detail="File not found")
# Check ownership
if metadata["user_id"] != user_id:
raise HTTPException(status_code=403, detail="Permission denied")
# Update metadata
updated = await metadata_manager.update_file_metadata(file_id, updates)
return {"status": "updated", "file_id": file_id, "metadata": updated}
except HTTPException:
raise
except Exception as e:
logger.error(f"Metadata update error: {e}")
raise HTTPException(status_code=500, detail=str(e))
# File Listing Endpoints
@app.get("/api/files")
async def list_files(
user_id: Optional[str] = None,
bucket: str = Query("default"),
limit: int = Query(20, le=100),
offset: int = Query(0),
search: Optional[str] = None,
file_type: Optional[str] = None,
sort_by: str = Query("created_at", pattern="^(created_at|filename|size)$"),
order: str = Query("desc", pattern="^(asc|desc)$")
):
"""List files with filtering and pagination"""
try:
files = await metadata_manager.list_files(
user_id=user_id,
bucket=bucket,
limit=limit,
offset=offset,
search=search,
file_type=file_type,
sort_by=sort_by,
order=order
)
return FileListResponse(**files)
except Exception as e:
logger.error(f"File listing error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.get("/api/files/user/{user_id}")
async def get_user_files(
user_id: str,
limit: int = Query(20, le=100),
offset: int = Query(0)
):
"""Get all files for a specific user"""
try:
files = await metadata_manager.list_files(
user_id=user_id,
limit=limit,
offset=offset
)
return FileListResponse(**files)
except Exception as e:
logger.error(f"User files listing error: {e}")
raise HTTPException(status_code=500, detail=str(e))
# Storage Management Endpoints
@app.get("/api/storage/stats")
async def get_storage_stats():
"""Get storage statistics"""
try:
stats = await minio_manager.get_storage_stats()
db_stats = await metadata_manager.get_storage_stats()
return StorageStats(
total_files=db_stats["total_files"],
total_size=db_stats["total_size"],
buckets=stats["buckets"],
users_count=db_stats["users_count"],
file_types=db_stats["file_types"]
)
except Exception as e:
logger.error(f"Storage stats error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.post("/api/storage/buckets")
async def create_bucket(bucket_name: str, public: bool = False):
"""Create a new storage bucket"""
try:
await minio_manager.create_bucket(bucket_name, public=public)
return {"status": "created", "bucket": bucket_name}
except Exception as e:
logger.error(f"Bucket creation error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.get("/api/storage/buckets")
async def list_buckets():
"""List all storage buckets"""
try:
buckets = await minio_manager.list_buckets()
return {"buckets": buckets}
except Exception as e:
logger.error(f"Bucket listing error: {e}")
raise HTTPException(status_code=500, detail=str(e))
# Presigned URL Endpoints
@app.post("/api/files/presigned-upload")
async def generate_presigned_upload_url(
filename: str,
content_type: str,
bucket: str = "default",
expires_in: int = Query(3600, ge=60, le=86400)
):
"""Generate presigned URL for direct upload to MinIO"""
try:
url = await minio_manager.generate_presigned_upload_url(
bucket=bucket,
object_name=f"{datetime.now().strftime('%Y%m%d')}/{filename}",
expires_in=expires_in
)
return {
"upload_url": url,
"expires_in": expires_in,
"method": "PUT"
}
except Exception as e:
logger.error(f"Presigned URL generation error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.get("/api/files/{file_id}/share")
async def generate_share_link(
file_id: str,
expires_in: int = Query(86400, ge=60, le=604800) # 1 day default, max 7 days
):
"""Generate a shareable link for a file"""
try:
# Get metadata
metadata = await metadata_manager.get_file_metadata(file_id)
if not metadata:
raise HTTPException(status_code=404, detail="File not found")
# Generate presigned URL
url = await minio_manager.generate_presigned_download_url(
bucket=metadata["bucket"],
object_name=metadata["object_name"],
expires_in=expires_in
)
return {
"share_url": url,
"expires_in": expires_in,
"file_id": file_id,
"filename": metadata["filename"]
}
except HTTPException:
raise
except Exception as e:
logger.error(f"Share link generation error: {e}")
raise HTTPException(status_code=500, detail=str(e))
# Batch Operations
@app.post("/api/files/batch-delete")
async def batch_delete_files(file_ids: List[str], user_id: str):
"""Delete multiple files at once"""
try:
deleted = []
errors = []
for file_id in file_ids:
try:
# Get metadata
metadata = await metadata_manager.get_file_metadata(file_id)
if metadata and metadata["user_id"] == user_id:
# Delete from MinIO
await minio_manager.delete_file(
bucket=metadata["bucket"],
object_name=metadata["object_name"]
)
# Delete metadata
await metadata_manager.delete_file_metadata(file_id)
deleted.append(file_id)
else:
errors.append({"file_id": file_id, "error": "Not found or permission denied"})
except Exception as e:
errors.append({"file_id": file_id, "error": str(e)})
return {
"deleted": deleted,
"errors": errors,
"total_deleted": len(deleted)
}
except Exception as e:
logger.error(f"Batch delete error: {e}")
raise HTTPException(status_code=500, detail=str(e))
if __name__ == "__main__":
uvicorn.run(
"main:app",
host="0.0.0.0",
port=8000,
reload=True
)

View File

@ -0,0 +1,331 @@
"""
Metadata Manager for file information storage in MongoDB
"""
from motor.motor_asyncio import AsyncIOMotorClient
from datetime import datetime
from typing import Optional, Dict, Any, List
import logging
import uuid
from models import FileType, FileStatus
logger = logging.getLogger(__name__)
class MetadataManager:
def __init__(self, mongodb_url: str, database: str = "files_db"):
self.mongodb_url = mongodb_url
self.database_name = database
self.client = None
self.db = None
self.collection = None
self.is_connected = False
async def connect(self):
"""Connect to MongoDB"""
try:
self.client = AsyncIOMotorClient(self.mongodb_url)
self.db = self.client[self.database_name]
self.collection = self.db.files
# Create indexes
await self._create_indexes()
# Test connection
await self.client.admin.command('ping')
self.is_connected = True
logger.info(f"Connected to MongoDB at {self.mongodb_url}")
except Exception as e:
logger.error(f"Failed to connect to MongoDB: {e}")
self.is_connected = False
raise
async def _create_indexes(self):
"""Create database indexes for better performance"""
try:
# Create indexes
await self.collection.create_index("user_id")
await self.collection.create_index("bucket")
await self.collection.create_index("created_at")
await self.collection.create_index("file_type")
await self.collection.create_index([("filename", "text")])
await self.collection.create_index([("user_id", 1), ("created_at", -1)])
logger.info("Database indexes created")
except Exception as e:
logger.error(f"Failed to create indexes: {e}")
async def create_file_metadata(self, metadata: Dict[str, Any]) -> str:
"""Create new file metadata"""
try:
# Add timestamps
metadata["created_at"] = datetime.now()
metadata["updated_at"] = datetime.now()
metadata["download_count"] = 0
metadata["status"] = FileStatus.READY.value
# Generate unique ID if not provided
if "id" not in metadata:
metadata["id"] = str(uuid.uuid4())
# Insert document
result = await self.collection.insert_one(metadata)
logger.info(f"Created metadata for file: {metadata['id']}")
return metadata["id"]
except Exception as e:
logger.error(f"Failed to create file metadata: {e}")
raise
async def get_file_metadata(self, file_id: str) -> Optional[Dict[str, Any]]:
"""Get file metadata by ID"""
try:
metadata = await self.collection.find_one({"id": file_id})
if metadata:
# Remove MongoDB's _id field
metadata.pop("_id", None)
return metadata
except Exception as e:
logger.error(f"Failed to get file metadata: {e}")
raise
async def update_file_metadata(self, file_id: str, updates: Dict[str, Any]) -> Dict[str, Any]:
"""Update file metadata"""
try:
# Add update timestamp
updates["updated_at"] = datetime.now()
# Update document
result = await self.collection.update_one(
{"id": file_id},
{"$set": updates}
)
if result.modified_count == 0:
raise Exception(f"File {file_id} not found")
# Return updated metadata
return await self.get_file_metadata(file_id)
except Exception as e:
logger.error(f"Failed to update file metadata: {e}")
raise
async def delete_file_metadata(self, file_id: str) -> bool:
"""Delete file metadata (soft delete)"""
try:
# Soft delete by marking as deleted
updates = {
"status": FileStatus.DELETED.value,
"deleted_at": datetime.now(),
"updated_at": datetime.now()
}
result = await self.collection.update_one(
{"id": file_id},
{"$set": updates}
)
return result.modified_count > 0
except Exception as e:
logger.error(f"Failed to delete file metadata: {e}")
raise
async def list_files(self, user_id: Optional[str] = None,
bucket: Optional[str] = None,
limit: int = 20,
offset: int = 0,
search: Optional[str] = None,
file_type: Optional[str] = None,
sort_by: str = "created_at",
order: str = "desc") -> Dict[str, Any]:
"""List files with filtering and pagination"""
try:
# Build query
query = {"status": {"$ne": FileStatus.DELETED.value}}
if user_id:
query["user_id"] = user_id
if bucket:
query["bucket"] = bucket
if file_type:
query["file_type"] = file_type
if search:
query["$text"] = {"$search": search}
# Count total documents
total = await self.collection.count_documents(query)
# Sort order
sort_order = -1 if order == "desc" else 1
# Execute query with pagination
cursor = self.collection.find(query)\
.sort(sort_by, sort_order)\
.skip(offset)\
.limit(limit)
files = []
async for doc in cursor:
doc.pop("_id", None)
files.append(doc)
return {
"files": files,
"total": total,
"limit": limit,
"offset": offset,
"has_more": (offset + limit) < total
}
except Exception as e:
logger.error(f"Failed to list files: {e}")
raise
async def increment_download_count(self, file_id: str):
"""Increment download counter for a file"""
try:
await self.collection.update_one(
{"id": file_id},
{
"$inc": {"download_count": 1},
"$set": {"last_accessed": datetime.now()}
}
)
except Exception as e:
logger.error(f"Failed to increment download count: {e}")
async def get_storage_stats(self) -> Dict[str, Any]:
"""Get storage statistics"""
try:
# Aggregation pipeline for statistics
pipeline = [
{"$match": {"status": {"$ne": FileStatus.DELETED.value}}},
{
"$group": {
"_id": None,
"total_files": {"$sum": 1},
"total_size": {"$sum": "$size"},
"users": {"$addToSet": "$user_id"}
}
}
]
cursor = self.collection.aggregate(pipeline)
result = await cursor.to_list(length=1)
if result:
stats = result[0]
users_count = len(stats.get("users", []))
else:
stats = {"total_files": 0, "total_size": 0}
users_count = 0
# Get file type distribution
type_pipeline = [
{"$match": {"status": {"$ne": FileStatus.DELETED.value}}},
{
"$group": {
"_id": "$file_type",
"count": {"$sum": 1}
}
}
]
type_cursor = self.collection.aggregate(type_pipeline)
type_results = await type_cursor.to_list(length=None)
file_types = {
item["_id"]: item["count"]
for item in type_results if item["_id"]
}
return {
"total_files": stats.get("total_files", 0),
"total_size": stats.get("total_size", 0),
"users_count": users_count,
"file_types": file_types
}
except Exception as e:
logger.error(f"Failed to get storage stats: {e}")
raise
async def find_duplicate_files(self, file_hash: str) -> List[Dict[str, Any]]:
"""Find duplicate files by hash"""
try:
cursor = self.collection.find({
"hash": file_hash,
"status": {"$ne": FileStatus.DELETED.value}
})
duplicates = []
async for doc in cursor:
doc.pop("_id", None)
duplicates.append(doc)
return duplicates
except Exception as e:
logger.error(f"Failed to find duplicate files: {e}")
raise
async def get_user_storage_usage(self, user_id: str) -> Dict[str, Any]:
"""Get storage usage for a specific user"""
try:
pipeline = [
{
"$match": {
"user_id": user_id,
"status": {"$ne": FileStatus.DELETED.value}
}
},
{
"$group": {
"_id": "$file_type",
"count": {"$sum": 1},
"size": {"$sum": "$size"}
}
}
]
cursor = self.collection.aggregate(pipeline)
results = await cursor.to_list(length=None)
total_size = sum(item["size"] for item in results)
total_files = sum(item["count"] for item in results)
breakdown = {
item["_id"]: {
"count": item["count"],
"size": item["size"]
}
for item in results if item["_id"]
}
return {
"user_id": user_id,
"total_files": total_files,
"total_size": total_size,
"breakdown": breakdown
}
except Exception as e:
logger.error(f"Failed to get user storage usage: {e}")
raise
async def close(self):
"""Close MongoDB connection"""
if self.client:
self.client.close()
self.is_connected = False
logger.info("MongoDB connection closed")

View File

@ -0,0 +1,333 @@
"""
MinIO Client for S3-compatible object storage
"""
from minio import Minio
from minio.error import S3Error
import asyncio
import io
from typing import Optional, Dict, Any, List
import logging
from datetime import timedelta
logger = logging.getLogger(__name__)
class MinIOManager:
def __init__(self, endpoint: str, access_key: str, secret_key: str, secure: bool = False):
self.endpoint = endpoint
self.access_key = access_key
self.secret_key = secret_key
self.secure = secure
self.client = None
self.is_connected = False
async def initialize(self):
"""Initialize MinIO client and create default buckets"""
try:
self.client = Minio(
self.endpoint,
access_key=self.access_key,
secret_key=self.secret_key,
secure=self.secure
)
# Create default buckets
default_buckets = ["default", "public", "thumbnails", "temp"]
for bucket in default_buckets:
await self.create_bucket(bucket, public=(bucket == "public"))
self.is_connected = True
logger.info(f"Connected to MinIO at {self.endpoint}")
except Exception as e:
logger.error(f"Failed to initialize MinIO: {e}")
self.is_connected = False
raise
async def create_bucket(self, bucket_name: str, public: bool = False):
"""Create a new bucket"""
try:
# Run in executor to avoid blocking
loop = asyncio.get_event_loop()
# Check if bucket exists
exists = await loop.run_in_executor(
None,
self.client.bucket_exists,
bucket_name
)
if not exists:
await loop.run_in_executor(
None,
self.client.make_bucket,
bucket_name
)
logger.info(f"Created bucket: {bucket_name}")
# Set bucket policy if public
if public:
policy = {
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Principal": {"AWS": ["*"]},
"Action": ["s3:GetObject"],
"Resource": [f"arn:aws:s3:::{bucket_name}/*"]
}
]
}
import json
await loop.run_in_executor(
None,
self.client.set_bucket_policy,
bucket_name,
json.dumps(policy)
)
logger.info(f"Set public policy for bucket: {bucket_name}")
except Exception as e:
logger.error(f"Failed to create bucket {bucket_name}: {e}")
raise
async def upload_file(self, bucket: str, object_name: str, file_data: bytes,
content_type: str = "application/octet-stream",
metadata: Optional[Dict[str, str]] = None):
"""Upload a file to MinIO"""
try:
loop = asyncio.get_event_loop()
# Convert bytes to BytesIO
file_stream = io.BytesIO(file_data)
length = len(file_data)
# Upload file
result = await loop.run_in_executor(
None,
self.client.put_object,
bucket,
object_name,
file_stream,
length,
content_type,
metadata
)
logger.info(f"Uploaded {object_name} to {bucket}")
return {
"bucket": bucket,
"object_name": object_name,
"etag": result.etag,
"version_id": result.version_id
}
except Exception as e:
logger.error(f"Failed to upload file: {e}")
raise
async def get_file(self, bucket: str, object_name: str) -> io.BytesIO:
"""Get a file from MinIO"""
try:
loop = asyncio.get_event_loop()
# Get object
response = await loop.run_in_executor(
None,
self.client.get_object,
bucket,
object_name
)
# Read data
data = response.read()
response.close()
response.release_conn()
return io.BytesIO(data)
except Exception as e:
logger.error(f"Failed to get file: {e}")
raise
async def delete_file(self, bucket: str, object_name: str):
"""Delete a file from MinIO"""
try:
loop = asyncio.get_event_loop()
await loop.run_in_executor(
None,
self.client.remove_object,
bucket,
object_name
)
logger.info(f"Deleted {object_name} from {bucket}")
except Exception as e:
logger.error(f"Failed to delete file: {e}")
raise
async def list_files(self, bucket: str, prefix: Optional[str] = None,
recursive: bool = True) -> List[Dict[str, Any]]:
"""List files in a bucket"""
try:
loop = asyncio.get_event_loop()
objects = await loop.run_in_executor(
None,
lambda: list(self.client.list_objects(
bucket,
prefix=prefix,
recursive=recursive
))
)
files = []
for obj in objects:
files.append({
"name": obj.object_name,
"size": obj.size,
"last_modified": obj.last_modified,
"etag": obj.etag,
"content_type": obj.content_type
})
return files
except Exception as e:
logger.error(f"Failed to list files: {e}")
raise
async def get_file_info(self, bucket: str, object_name: str) -> Dict[str, Any]:
"""Get file information"""
try:
loop = asyncio.get_event_loop()
stat = await loop.run_in_executor(
None,
self.client.stat_object,
bucket,
object_name
)
return {
"size": stat.size,
"etag": stat.etag,
"content_type": stat.content_type,
"last_modified": stat.last_modified,
"metadata": stat.metadata
}
except Exception as e:
logger.error(f"Failed to get file info: {e}")
raise
async def generate_presigned_download_url(self, bucket: str, object_name: str,
expires_in: int = 3600) -> str:
"""Generate a presigned URL for downloading"""
try:
loop = asyncio.get_event_loop()
url = await loop.run_in_executor(
None,
self.client.presigned_get_object,
bucket,
object_name,
timedelta(seconds=expires_in)
)
return url
except Exception as e:
logger.error(f"Failed to generate presigned URL: {e}")
raise
async def generate_presigned_upload_url(self, bucket: str, object_name: str,
expires_in: int = 3600) -> str:
"""Generate a presigned URL for uploading"""
try:
loop = asyncio.get_event_loop()
url = await loop.run_in_executor(
None,
self.client.presigned_put_object,
bucket,
object_name,
timedelta(seconds=expires_in)
)
return url
except Exception as e:
logger.error(f"Failed to generate presigned upload URL: {e}")
raise
async def copy_file(self, source_bucket: str, source_object: str,
dest_bucket: str, dest_object: str):
"""Copy a file within MinIO"""
try:
loop = asyncio.get_event_loop()
await loop.run_in_executor(
None,
self.client.copy_object,
dest_bucket,
dest_object,
f"/{source_bucket}/{source_object}"
)
logger.info(f"Copied {source_object} to {dest_object}")
except Exception as e:
logger.error(f"Failed to copy file: {e}")
raise
async def list_buckets(self) -> List[str]:
"""List all buckets"""
try:
loop = asyncio.get_event_loop()
buckets = await loop.run_in_executor(
None,
self.client.list_buckets
)
return [bucket.name for bucket in buckets]
except Exception as e:
logger.error(f"Failed to list buckets: {e}")
raise
async def get_storage_stats(self) -> Dict[str, Any]:
"""Get storage statistics"""
try:
buckets = await self.list_buckets()
stats = {
"buckets": buckets,
"bucket_count": len(buckets),
"bucket_stats": {}
}
# Get stats for each bucket
for bucket in buckets:
files = await self.list_files(bucket)
total_size = sum(f["size"] for f in files)
stats["bucket_stats"][bucket] = {
"file_count": len(files),
"total_size": total_size
}
return stats
except Exception as e:
logger.error(f"Failed to get storage stats: {e}")
raise
async def check_file_exists(self, bucket: str, object_name: str) -> bool:
"""Check if a file exists"""
try:
await self.get_file_info(bucket, object_name)
return True
except:
return False

View File

@ -0,0 +1,112 @@
"""
Data models for File Management Service
"""
from pydantic import BaseModel, Field
from datetime import datetime
from typing import Optional, List, Dict, Any
from enum import Enum
class FileType(str, Enum):
IMAGE = "image"
VIDEO = "video"
AUDIO = "audio"
DOCUMENT = "document"
ARCHIVE = "archive"
OTHER = "other"
class FileStatus(str, Enum):
PENDING = "pending"
PROCESSING = "processing"
READY = "ready"
ERROR = "error"
DELETED = "deleted"
class FileMetadata(BaseModel):
id: str
filename: str
original_name: str
size: int
content_type: str
file_type: FileType
bucket: str
object_name: str
user_id: str
hash: str
status: FileStatus = FileStatus.READY
public: bool = False
has_thumbnail: bool = False
thumbnail_url: Optional[str] = None
tags: Dict[str, Any] = {}
metadata: Dict[str, Any] = {}
download_count: int = 0
created_at: datetime
updated_at: datetime
deleted_at: Optional[datetime] = None
class FileUploadResponse(BaseModel):
file_id: str
filename: str
size: int
content_type: str
file_type: FileType
bucket: str
public: bool
has_thumbnail: bool
thumbnail_url: Optional[str] = None
download_url: Optional[str] = None
created_at: datetime
message: str = "File uploaded successfully"
class FileListResponse(BaseModel):
files: List[FileMetadata]
total: int
limit: int
offset: int
has_more: bool
class StorageStats(BaseModel):
total_files: int
total_size: int
buckets: List[str]
users_count: int
file_types: Dict[str, int]
storage_used_percentage: Optional[float] = None
class ThumbnailRequest(BaseModel):
file_id: str
width: int = Field(200, ge=50, le=1000)
height: int = Field(200, ge=50, le=1000)
quality: int = Field(85, ge=50, le=100)
format: str = Field("jpeg", pattern="^(jpeg|png|webp)$")
class PresignedUrlResponse(BaseModel):
url: str
expires_in: int
method: str
headers: Optional[Dict[str, str]] = None
class BatchOperationResult(BaseModel):
successful: List[str]
failed: List[Dict[str, str]]
total_processed: int
total_successful: int
total_failed: int
class FileShareLink(BaseModel):
share_url: str
expires_in: int
file_id: str
filename: str
created_at: datetime
expires_at: datetime
class FileProcessingJob(BaseModel):
job_id: str
file_id: str
job_type: str # thumbnail, compress, convert, etc.
status: str # pending, processing, completed, failed
progress: Optional[float] = None
result: Optional[Dict[str, Any]] = None
error: Optional[str] = None
created_at: datetime
completed_at: Optional[datetime] = None

View File

@ -0,0 +1,11 @@
fastapi==0.109.0
uvicorn[standard]==0.27.0
pydantic==2.5.3
python-dotenv==1.0.0
motor==3.5.1
pymongo==4.6.1
minio==7.2.3
pillow==10.2.0
python-magic==0.4.27
aiofiles==23.2.1
python-multipart==0.0.6

View File

@ -0,0 +1,281 @@
#!/usr/bin/env python3
"""
Test script for File Management Service
"""
import asyncio
import httpx
import os
import json
from datetime import datetime
import base64
BASE_URL = "http://localhost:8014"
# Sample image for testing (1x1 pixel PNG)
TEST_IMAGE_BASE64 = "iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mNkYPhfDwAChwGA60e6kgAAAABJRU5ErkJggg=="
TEST_IMAGE_DATA = base64.b64decode(TEST_IMAGE_BASE64)
async def test_file_api():
"""Test file management API endpoints"""
async with httpx.AsyncClient() as client:
print("\n📁 Testing File Management Service API...")
# Test health check
print("\n1. Testing health check...")
response = await client.get(f"{BASE_URL}/health")
print(f"Health check: {response.json()}")
# Test file upload
print("\n2. Testing file upload...")
files = {
'file': ('test_image.png', TEST_IMAGE_DATA, 'image/png')
}
data = {
'user_id': 'test_user_123',
'bucket': 'default',
'public': 'false',
'generate_thumbnail': 'true',
'tags': json.dumps({"test": "true", "category": "sample"})
}
response = await client.post(
f"{BASE_URL}/api/files/upload",
files=files,
data=data
)
if response.status_code == 200:
upload_result = response.json()
print(f"File uploaded: {upload_result}")
file_id = upload_result.get("file_id")
else:
print(f"Upload failed: {response.status_code} - {response.text}")
file_id = None
# Test multiple file upload
print("\n3. Testing multiple file upload...")
files = [
('files', ('test1.png', TEST_IMAGE_DATA, 'image/png')),
('files', ('test2.png', TEST_IMAGE_DATA, 'image/png')),
('files', ('test3.png', TEST_IMAGE_DATA, 'image/png'))
]
data = {
'user_id': 'test_user_123',
'bucket': 'default',
'public': 'false'
}
response = await client.post(
f"{BASE_URL}/api/files/upload-multiple",
files=files,
data=data
)
if response.status_code == 200:
print(f"Multiple files uploaded: {response.json()}")
else:
print(f"Multiple upload failed: {response.status_code}")
# Test file metadata retrieval
if file_id:
print("\n4. Testing file metadata retrieval...")
response = await client.get(f"{BASE_URL}/api/files/{file_id}/metadata")
if response.status_code == 200:
print(f"File metadata: {response.json()}")
else:
print(f"Metadata retrieval failed: {response.status_code}")
# Test thumbnail generation
print("\n5. Testing thumbnail retrieval...")
response = await client.get(
f"{BASE_URL}/api/files/{file_id}/thumbnail",
params={"width": 100, "height": 100}
)
if response.status_code == 200:
print(f"Thumbnail retrieved: {len(response.content)} bytes")
else:
print(f"Thumbnail retrieval failed: {response.status_code}")
# Test file download
print("\n6. Testing file download...")
response = await client.get(f"{BASE_URL}/api/files/{file_id}/download")
if response.status_code == 200:
print(f"File downloaded: {len(response.content)} bytes")
else:
print(f"Download failed: {response.status_code}")
# Test share link generation
print("\n7. Testing share link generation...")
response = await client.get(
f"{BASE_URL}/api/files/{file_id}/share",
params={"expires_in": 3600}
)
if response.status_code == 200:
share_result = response.json()
print(f"Share link generated: {share_result.get('share_url', 'N/A')[:50]}...")
else:
print(f"Share link generation failed: {response.status_code}")
# Test file listing
print("\n8. Testing file listing...")
response = await client.get(
f"{BASE_URL}/api/files",
params={
"user_id": "test_user_123",
"limit": 10,
"offset": 0
}
)
if response.status_code == 200:
files_list = response.json()
print(f"Files found: {files_list.get('total', 0)} files")
else:
print(f"File listing failed: {response.status_code}")
# Test storage statistics
print("\n9. Testing storage statistics...")
response = await client.get(f"{BASE_URL}/api/storage/stats")
if response.status_code == 200:
stats = response.json()
print(f"Storage stats: {stats}")
else:
print(f"Storage stats failed: {response.status_code}")
# Test bucket operations
print("\n10. Testing bucket operations...")
response = await client.post(
f"{BASE_URL}/api/storage/buckets",
params={"bucket_name": "test-bucket", "public": False}
)
if response.status_code == 200:
print(f"Bucket created: {response.json()}")
else:
print(f"Bucket creation: {response.status_code}")
response = await client.get(f"{BASE_URL}/api/storage/buckets")
if response.status_code == 200:
print(f"Available buckets: {response.json()}")
else:
print(f"Bucket listing failed: {response.status_code}")
# Test presigned URL generation
print("\n11. Testing presigned URL generation...")
response = await client.post(
f"{BASE_URL}/api/files/presigned-upload",
params={
"filename": "test_upload.txt",
"content_type": "text/plain",
"bucket": "default",
"expires_in": 3600
}
)
if response.status_code == 200:
presigned = response.json()
print(f"Presigned upload URL generated: {presigned.get('upload_url', 'N/A')[:50]}...")
else:
print(f"Presigned URL generation failed: {response.status_code}")
# Test file deletion
if file_id:
print("\n12. Testing file deletion...")
response = await client.delete(
f"{BASE_URL}/api/files/{file_id}",
params={"user_id": "test_user_123"}
)
if response.status_code == 200:
print(f"File deleted: {response.json()}")
else:
print(f"File deletion failed: {response.status_code}")
async def test_large_file_upload():
"""Test large file upload"""
print("\n\n📦 Testing Large File Upload...")
# Create a larger test file (1MB)
large_data = b"x" * (1024 * 1024) # 1MB of data
async with httpx.AsyncClient(timeout=30.0) as client:
files = {
'file': ('large_test.bin', large_data, 'application/octet-stream')
}
data = {
'user_id': 'test_user_123',
'bucket': 'default',
'public': 'false'
}
print("Uploading 1MB file...")
response = await client.post(
f"{BASE_URL}/api/files/upload",
files=files,
data=data
)
if response.status_code == 200:
result = response.json()
print(f"Large file uploaded successfully: {result.get('file_id')}")
print(f"File size: {result.get('size')} bytes")
else:
print(f"Large file upload failed: {response.status_code}")
async def test_duplicate_detection():
"""Test duplicate file detection"""
print("\n\n🔍 Testing Duplicate Detection...")
async with httpx.AsyncClient() as client:
# Upload the same file twice
files = {
'file': ('duplicate_test.png', TEST_IMAGE_DATA, 'image/png')
}
data = {
'user_id': 'test_user_123',
'bucket': 'default',
'public': 'false'
}
print("Uploading file first time...")
response1 = await client.post(
f"{BASE_URL}/api/files/upload",
files=files,
data=data
)
if response1.status_code == 200:
result1 = response1.json()
print(f"First upload: {result1.get('file_id')}")
print("Uploading same file again...")
response2 = await client.post(
f"{BASE_URL}/api/files/upload",
files=files,
data=data
)
if response2.status_code == 200:
result2 = response2.json()
print(f"Second upload: {result2.get('file_id')}")
if result2.get('duplicate'):
print("✅ Duplicate detected successfully!")
else:
print("❌ Duplicate not detected")
async def main():
"""Run all tests"""
print("=" * 60)
print("FILE MANAGEMENT SERVICE TEST SUITE")
print("=" * 60)
print(f"Started at: {datetime.now().isoformat()}")
# Run tests
await test_file_api()
await test_large_file_upload()
await test_duplicate_detection()
print("\n" + "=" * 60)
print("✅ All tests completed!")
print(f"Finished at: {datetime.now().isoformat()}")
print("=" * 60)
if __name__ == "__main__":
asyncio.run(main())

View File

@ -0,0 +1,236 @@
"""
Thumbnail Generator for image and video files
"""
from PIL import Image, ImageOps
import io
import os
import hashlib
import logging
from typing import Optional, Tuple
import asyncio
from pathlib import Path
logger = logging.getLogger(__name__)
class ThumbnailGenerator:
def __init__(self, minio_client, cache_dir: str = "/tmp/thumbnails"):
self.minio_client = minio_client
self.cache_dir = Path(cache_dir)
self.cache_dir.mkdir(parents=True, exist_ok=True)
# Supported image formats for thumbnail generation
self.supported_formats = {
'image/jpeg', 'image/jpg', 'image/png', 'image/gif',
'image/webp', 'image/bmp', 'image/tiff'
}
def _get_cache_path(self, file_id: str, width: int, height: int) -> Path:
"""Generate cache file path for thumbnail"""
cache_key = f"{file_id}_{width}x{height}"
cache_hash = hashlib.md5(cache_key.encode()).hexdigest()
return self.cache_dir / f"{cache_hash[:2]}" / f"{cache_hash}.jpg"
async def generate_thumbnail(self, file_data: bytes, content_type: str,
width: int = 200, height: int = 200) -> Optional[bytes]:
"""Generate a thumbnail from file data"""
try:
if content_type not in self.supported_formats:
logger.warning(f"Unsupported format for thumbnail: {content_type}")
return None
loop = asyncio.get_event_loop()
# Generate thumbnail in thread pool
thumbnail_data = await loop.run_in_executor(
None,
self._create_thumbnail,
file_data,
width,
height
)
return thumbnail_data
except Exception as e:
logger.error(f"Failed to generate thumbnail: {e}")
return None
def _create_thumbnail(self, file_data: bytes, width: int, height: int) -> bytes:
"""Create thumbnail using PIL"""
try:
# Open image
image = Image.open(io.BytesIO(file_data))
# Convert RGBA to RGB if necessary
if image.mode in ('RGBA', 'LA', 'P'):
# Create a white background
background = Image.new('RGB', image.size, (255, 255, 255))
if image.mode == 'P':
image = image.convert('RGBA')
background.paste(image, mask=image.split()[-1] if image.mode == 'RGBA' else None)
image = background
elif image.mode not in ('RGB', 'L'):
image = image.convert('RGB')
# Calculate thumbnail size maintaining aspect ratio
image.thumbnail((width, height), Image.Resampling.LANCZOS)
# Apply EXIF orientation if present
image = ImageOps.exif_transpose(image)
# Save thumbnail to bytes
output = io.BytesIO()
image.save(output, format='JPEG', quality=85, optimize=True)
output.seek(0)
return output.read()
except Exception as e:
logger.error(f"Thumbnail creation failed: {e}")
raise
async def get_thumbnail(self, file_id: str, bucket: str, object_name: str,
width: int = 200, height: int = 200) -> Optional[bytes]:
"""Get or generate thumbnail for a file"""
try:
# Check cache first
cache_path = self._get_cache_path(file_id, width, height)
if cache_path.exists():
logger.info(f"Thumbnail found in cache: {cache_path}")
with open(cache_path, 'rb') as f:
return f.read()
# Check if thumbnail exists in MinIO
thumbnail_object = f"thumbnails/{file_id}_{width}x{height}.jpg"
try:
thumbnail_stream = await self.minio_client.get_file(
bucket="thumbnails",
object_name=thumbnail_object
)
thumbnail_data = thumbnail_stream.read()
# Save to cache
await self._save_to_cache(cache_path, thumbnail_data)
return thumbnail_data
except:
pass # Thumbnail doesn't exist, generate it
# Get original file
file_stream = await self.minio_client.get_file(bucket, object_name)
file_data = file_stream.read()
# Get file info for content type
file_info = await self.minio_client.get_file_info(bucket, object_name)
content_type = file_info.get("content_type", "")
# Generate thumbnail
thumbnail_data = await self.generate_thumbnail(
file_data, content_type, width, height
)
if thumbnail_data:
# Save to MinIO
await self.minio_client.upload_file(
bucket="thumbnails",
object_name=thumbnail_object,
file_data=thumbnail_data,
content_type="image/jpeg"
)
# Save to cache
await self._save_to_cache(cache_path, thumbnail_data)
return thumbnail_data
except Exception as e:
logger.error(f"Failed to get thumbnail: {e}")
return None
async def _save_to_cache(self, cache_path: Path, data: bytes):
"""Save thumbnail to cache"""
try:
cache_path.parent.mkdir(parents=True, exist_ok=True)
loop = asyncio.get_event_loop()
await loop.run_in_executor(
None,
lambda: cache_path.write_bytes(data)
)
logger.info(f"Thumbnail saved to cache: {cache_path}")
except Exception as e:
logger.error(f"Failed to save to cache: {e}")
async def delete_thumbnail(self, file_id: str):
"""Delete all thumbnails for a file"""
try:
# Delete from cache
for cache_file in self.cache_dir.rglob(f"*{file_id}*"):
try:
cache_file.unlink()
logger.info(f"Deleted cache file: {cache_file}")
except:
pass
# Delete from MinIO (list and delete all sizes)
files = await self.minio_client.list_files(
bucket="thumbnails",
prefix=f"thumbnails/{file_id}_"
)
for file in files:
await self.minio_client.delete_file(
bucket="thumbnails",
object_name=file["name"]
)
logger.info(f"Deleted thumbnail: {file['name']}")
except Exception as e:
logger.error(f"Failed to delete thumbnails: {e}")
async def generate_multiple_sizes(self, file_data: bytes, content_type: str,
file_id: str) -> dict:
"""Generate thumbnails in multiple sizes"""
sizes = {
"small": (150, 150),
"medium": (300, 300),
"large": (600, 600)
}
results = {}
for size_name, (width, height) in sizes.items():
thumbnail = await self.generate_thumbnail(
file_data, content_type, width, height
)
if thumbnail:
# Save to MinIO
object_name = f"thumbnails/{file_id}_{size_name}.jpg"
await self.minio_client.upload_file(
bucket="thumbnails",
object_name=object_name,
file_data=thumbnail,
content_type="image/jpeg"
)
results[size_name] = {
"size": len(thumbnail),
"dimensions": f"{width}x{height}",
"object_name": object_name
}
return results
def clear_cache(self):
"""Clear thumbnail cache"""
try:
import shutil
shutil.rmtree(self.cache_dir)
self.cache_dir.mkdir(parents=True, exist_ok=True)
logger.info("Thumbnail cache cleared")
except Exception as e:
logger.error(f"Failed to clear cache: {e}")

View File

@ -0,0 +1,26 @@
FROM python:3.11-slim
WORKDIR /app
# 시스템 패키지 설치
RUN apt-get update && apt-get install -y \
gcc \
libheif-dev \
libde265-dev \
libjpeg-dev \
libpng-dev \
&& rm -rf /var/lib/apt/lists/*
# Python 패키지 설치
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
# 애플리케이션 코드 복사
COPY . .
# 캐시 디렉토리 생성
RUN mkdir -p /app/cache
EXPOSE 8000
CMD ["python", "main.py"]

View File

View File

@ -0,0 +1,197 @@
from fastapi import APIRouter, Query, HTTPException, Body
from fastapi.responses import Response
from typing import Optional, Dict
import mimetypes
from pathlib import Path
import hashlib
from ..core.config import settings
# MinIO 사용 여부에 따라 적절한 캐시 모듈 선택
if settings.use_minio:
from ..core.minio_cache import cache
else:
from ..core.cache import cache
router = APIRouter()
@router.get("/image")
async def get_image(
url: str = Query(..., description="원본 이미지 URL"),
size: Optional[str] = Query(None, description="이미지 크기 (thumb, card, list, detail, hero)")
):
"""
이미지 프록시 엔드포인트
- 외부 URL의 이미지를 가져와서 캐싱
- 선택적으로 리사이징 및 최적화
- WebP 포맷으로 자동 변환 (설정에 따라)
"""
try:
# 캐시 확인
cached_data = await cache.get(url, size)
if cached_data:
# 캐시된 이미지 반환
# SVG 체크
if url.lower().endswith('.svg') or cache._is_svg(cached_data):
content_type = 'image/svg+xml'
# GIF 체크 (GIF는 WebP로 변환하지 않음)
elif url.lower().endswith('.gif'):
content_type = 'image/gif'
# WebP 변환이 활성화된 경우 항상 WebP로 제공 (GIF 제외)
elif settings.convert_to_webp and size:
content_type = 'image/webp'
else:
content_type = mimetypes.guess_type(url)[0] or 'image/jpeg'
return Response(
content=cached_data,
media_type=content_type,
headers={
"Cache-Control": f"public, max-age={86400 * 7}", # 7일 브라우저 캐시
"X-Cache": "HIT",
"X-Image-Format": content_type.split('/')[-1].upper()
}
)
# 캐시 미스 - 이미지 다운로드
image_data = await cache.download_image(url)
# URL에서 MIME 타입 추측
guessed_type = mimetypes.guess_type(url)[0]
# SVG 확장자 체크 (mimetypes가 SVG를 제대로 인식하지 못할 수 있음)
if url.lower().endswith('.svg') or cache._is_svg(image_data):
content_type = 'image/svg+xml'
# GIF 체크
elif url.lower().endswith('.gif') or (guessed_type and 'gif' in guessed_type.lower()):
content_type = 'image/gif'
else:
content_type = guessed_type or 'image/jpeg'
# 리사이징 및 최적화 (SVG와 GIF는 특별 처리)
if size and content_type != 'image/svg+xml':
# GIF는 특별 처리
if content_type == 'image/gif':
image_data, content_type = cache._process_gif(image_data, settings.thumbnail_sizes[size])
else:
image_data, content_type = cache.resize_and_optimize_image(image_data, size)
# 캐시에 저장
await cache.set(url, image_data, size)
# 백그라운드에서 다른 크기들도 생성하도록 트리거
await cache.trigger_background_generation(url)
# 이미지 반환
return Response(
content=image_data,
media_type=content_type,
headers={
"Cache-Control": f"public, max-age={86400 * 7}",
"X-Cache": "MISS",
"X-Image-Format": content_type.split('/')[-1].upper()
}
)
except HTTPException:
raise
except Exception as e:
import traceback
print(f"Error processing image from {url}: {str(e)}")
traceback.print_exc()
# 403 에러를 명확히 처리
if "403" in str(e):
raise HTTPException(
status_code=403,
detail=f"이미지 접근 거부됨: {url}"
)
raise HTTPException(
status_code=500,
detail=f"이미지 처리 실패: {str(e)}"
)
@router.get("/stats")
async def get_stats():
"""캐시 통계 정보"""
cache_size = await cache.get_cache_size()
# 디렉토리 구조 통계 추가 (MinIO 또는 파일시스템)
dir_stats = await cache.get_directory_stats()
return {
"cache_size_gb": round(cache_size, 2),
"max_cache_size_gb": settings.max_cache_size_gb,
"cache_usage_percent": round((cache_size / settings.max_cache_size_gb) * 100, 2),
"directory_stats": dir_stats
}
@router.post("/cleanup")
async def cleanup_cache():
"""오래된 캐시 정리"""
await cache.cleanup_old_cache()
return {"message": "캐시 정리 완료"}
@router.post("/cache/delete")
async def delete_cache(request: Dict = Body(...)):
"""특정 URL의 캐시 삭제"""
url = request.get("url")
if not url:
raise HTTPException(status_code=400, detail="URL이 필요합니다")
try:
# URL의 모든 크기 버전 삭제
sizes = ["thumb", "card", "list", "detail", "hero", None] # None은 원본
deleted_count = 0
for size in sizes:
# 캐시 경로 계산
url_hash = hashlib.md5(url.encode()).hexdigest()
# 3단계 디렉토리 구조
level1 = url_hash[:2]
level2 = url_hash[2:4]
level3 = url_hash[4:6]
# 크기별 파일명
if size:
patterns = [
f"{url_hash}_{size}.webp",
f"{url_hash}_{size}.jpg",
f"{url_hash}_{size}.jpeg",
f"{url_hash}_{size}.png",
f"{url_hash}_{size}.gif"
]
else:
patterns = [
f"{url_hash}",
f"{url_hash}.jpg",
f"{url_hash}.jpeg",
f"{url_hash}.png",
f"{url_hash}.gif",
f"{url_hash}.webp"
]
# 각 패턴에 대해 파일 삭제 시도
for filename in patterns:
cache_path = settings.cache_dir / level1 / level2 / level3 / filename
if cache_path.exists():
cache_path.unlink()
deleted_count += 1
print(f"✅ 캐시 파일 삭제: {cache_path}")
return {
"status": "success",
"message": f"{deleted_count}개의 캐시 파일이 삭제되었습니다",
"url": url
}
except Exception as e:
print(f"❌ 캐시 삭제 오류: {e}")
raise HTTPException(
status_code=500,
detail=f"캐시 삭제 실패: {str(e)}"
)

View File

@ -0,0 +1,91 @@
import asyncio
import logging
from typing import Set, Optional
from pathlib import Path
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
class BackgroundTaskManager:
"""백그라운드 작업 관리자"""
def __init__(self):
self.processing_urls: Set[str] = set() # 현재 처리 중인 URL 목록
self.task_queue: asyncio.Queue = None
self.worker_task: Optional[asyncio.Task] = None
async def start(self):
"""백그라운드 워커 시작"""
self.task_queue = asyncio.Queue(maxsize=100)
self.worker_task = asyncio.create_task(self._worker())
logger.info("백그라운드 작업 관리자 시작됨")
async def stop(self):
"""백그라운드 워커 정지"""
if self.worker_task:
self.worker_task.cancel()
try:
await self.worker_task
except asyncio.CancelledError:
pass
logger.info("백그라운드 작업 관리자 정지됨")
async def add_task(self, url: str):
"""작업 큐에 URL 추가"""
if url not in self.processing_urls and self.task_queue:
try:
self.processing_urls.add(url)
await self.task_queue.put(url)
logger.info(f"백그라운드 작업 추가: {url}")
except asyncio.QueueFull:
self.processing_urls.discard(url)
logger.warning(f"작업 큐가 가득 참: {url}")
async def _worker(self):
"""백그라운드 워커 - 큐에서 작업을 가져와 처리"""
from .cache import cache
while True:
try:
# 큐에서 URL 가져오기
url = await self.task_queue.get()
try:
# 원본 이미지가 캐시에 있는지 확인
original_data = await cache.get(url, None)
if not original_data:
# 원본 이미지 다운로드
original_data = await cache.download_image(url)
await cache.set(url, original_data, None)
# 모든 크기의 이미지 생성
sizes = ['thumb', 'card', 'list', 'detail', 'hero']
for size in sizes:
# 이미 존재하는지 확인
existing = await cache.get(url, size)
if not existing:
try:
# 리사이징 및 최적화 - cache.resize_and_optimize_image가 WebP를 처리함
resized_data, _ = cache.resize_and_optimize_image(original_data, size)
await cache.set(url, resized_data, size)
logger.info(f"백그라운드 생성 완료: {url} ({size})")
except Exception as e:
logger.error(f"백그라운드 리사이징 실패: {url} ({size}) - {str(e)}")
import traceback
logger.error(f"Traceback: {traceback.format_exc()}")
except Exception as e:
logger.error(f"백그라운드 작업 실패: {url} - {str(e)}")
finally:
# 처리 완료된 URL 제거
self.processing_urls.discard(url)
except asyncio.CancelledError:
break
except Exception as e:
logger.error(f"백그라운드 워커 오류: {str(e)}")
await asyncio.sleep(1) # 오류 발생 시 잠시 대기
# 전역 백그라운드 작업 관리자
background_manager = BackgroundTaskManager()

View File

@ -0,0 +1,796 @@
import hashlib
import aiofiles
import os
from pathlib import Path
from datetime import datetime, timedelta
from typing import Optional
import httpx
from PIL import Image
try:
from pillow_heif import register_heif_opener, register_avif_opener
register_heif_opener() # HEIF/HEIC 지원
register_avif_opener() # AVIF 지원
print("HEIF/AVIF support enabled successfully")
except ImportError:
print("Warning: pillow_heif not installed, HEIF/AVIF support disabled")
import io
import asyncio
import ssl
from .config import settings
class ImageCache:
def __init__(self):
self.cache_dir = settings.cache_dir
self.cache_dir.mkdir(parents=True, exist_ok=True)
def _get_cache_path(self, url: str, size: Optional[str] = None) -> Path:
"""URL을 기반으로 캐시 파일 경로 생성"""
# URL을 해시하여 파일명 생성
url_hash = hashlib.md5(url.encode()).hexdigest()
# 3단계 디렉토리 구조 생성
# 예: 10f8a8f96aa1377e86fdbc6bf3c631cf -> 10/f8/a8/
level1 = url_hash[:2] # 첫 2자리
level2 = url_hash[2:4] # 다음 2자리
level3 = url_hash[4:6] # 다음 2자리
# 크기별로 다른 파일명 사용
if size:
filename = f"{url_hash}_{size}"
else:
filename = url_hash
# 확장자 추출 (WebP로 저장되는 경우 .webp 사용)
if settings.convert_to_webp and size:
filename = f"{filename}.webp"
else:
ext = self._get_extension_from_url(url)
if ext:
filename = f"{filename}.{ext}"
# 3단계 디렉토리 경로 생성
path = self.cache_dir / level1 / level2 / level3 / filename
path.parent.mkdir(parents=True, exist_ok=True)
return path
def _get_extension_from_url(self, url: str) -> Optional[str]:
"""URL에서 파일 확장자 추출"""
path = url.split('?')[0] # 쿼리 파라미터 제거
parts = path.split('.')
if len(parts) > 1:
ext = parts[-1].lower()
if ext in settings.allowed_formats:
return ext
return None
def _is_svg(self, data: bytes) -> bool:
"""SVG 파일인지 확인"""
# SVG 파일의 시작 부분 확인
if len(data) < 100:
return False
# 처음 1000바이트만 확인 (성능 최적화)
header = data[:1000].lower()
# SVG 시그니처 확인
svg_signatures = [
b'<svg',
b'<?xml',
b'<!doctype svg'
]
for sig in svg_signatures:
if sig in header:
return True
return False
def _process_gif(self, gif_data: bytes, target_size: tuple) -> tuple[bytes, str]:
"""GIF 처리 - JPEG로 변환하여 안정적으로 처리"""
try:
from PIL import Image
# GIF 열기
img = Image.open(io.BytesIO(gif_data))
# 모든 GIF를 RGB로 변환 (팔레트 모드 문제 해결)
# 팔레트 모드(P)를 RGB로 직접 변환
if img.mode != 'RGB':
img = img.convert('RGB')
# 리사이징
img = img.resize(target_size, Image.Resampling.LANCZOS)
# JPEG로 저장 (안정적)
output = io.BytesIO()
img.save(output, format='JPEG', quality=85, optimize=True)
return output.getvalue(), 'image/jpeg'
except Exception as e:
print(f"GIF 처리 중 오류: {e}")
import traceback
traceback.print_exc()
# 오류 발생 시 원본 반환
return gif_data, 'image/gif'
async def get(self, url: str, size: Optional[str] = None) -> Optional[bytes]:
"""캐시에서 이미지 가져오기"""
cache_path = self._get_cache_path(url, size)
if cache_path.exists():
# 캐시 만료 확인
stat = cache_path.stat()
age = datetime.now() - datetime.fromtimestamp(stat.st_mtime)
if age < timedelta(days=settings.cache_ttl_days):
async with aiofiles.open(cache_path, 'rb') as f:
return await f.read()
else:
# 만료된 캐시 삭제
cache_path.unlink()
return None
async def set(self, url: str, data: bytes, size: Optional[str] = None):
"""캐시에 이미지 저장"""
cache_path = self._get_cache_path(url, size)
async with aiofiles.open(cache_path, 'wb') as f:
await f.write(data)
async def download_image(self, url: str) -> bytes:
"""외부 URL에서 이미지 다운로드"""
from urllib.parse import urlparse
# URL에서 도메인 추출
parsed_url = urlparse(url)
domain = parsed_url.netloc
base_url = f"{parsed_url.scheme}://{parsed_url.netloc}/"
# 기본 헤더 설정
headers = {
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36',
'Accept': 'image/webp,image/apng,image/svg+xml,image/*,*/*;q=0.8',
'Accept-Language': 'ko-KR,ko;q=0.9,en-US;q=0.8,en;q=0.7',
'Cache-Control': 'no-cache',
'Pragma': 'no-cache',
'Sec-Fetch-Dest': 'image',
'Sec-Fetch-Mode': 'no-cors',
'Sec-Fetch-Site': 'cross-site',
'Referer': base_url # 항상 기본 Referer 설정
}
# 특정 사이트별 Referer 오버라이드
if 'yna.co.kr' in url:
headers['Referer'] = 'https://www.yna.co.kr/'
client = httpx.AsyncClient(
verify=False, # SSL 검증 비활성화
timeout=30.0,
follow_redirects=True
)
elif 'investing.com' in url:
headers['Referer'] = 'https://www.investing.com/'
client = httpx.AsyncClient()
elif 'naver.com' in url:
headers['Referer'] = 'https://news.naver.com/'
client = httpx.AsyncClient()
elif 'daum.net' in url:
headers['Referer'] = 'https://news.daum.net/'
client = httpx.AsyncClient()
elif 'chosun.com' in url:
headers['Referer'] = 'https://www.chosun.com/'
client = httpx.AsyncClient()
elif 'vietnam.vn' in url or 'vstatic.vietnam.vn' in url:
headers['Referer'] = 'https://vietnam.vn/'
client = httpx.AsyncClient()
elif 'ddaily.co.kr' in url:
# ddaily는 /photos/ 경로를 사용해야 함
headers['Referer'] = 'https://www.ddaily.co.kr/'
# URL이 잘못된 경로를 사용하는 경우 수정
if '/2025/' in url and '/photos/' not in url:
url = url.replace('/2025/', '/photos/2025/')
print(f"Fixed ddaily URL: {url}")
client = httpx.AsyncClient()
else:
# 기본적으로 도메인 기반 Referer 사용
client = httpx.AsyncClient()
async with client:
try:
response = await client.get(
url,
headers=headers,
timeout=settings.request_timeout,
follow_redirects=True
)
response.raise_for_status()
except Exception as e:
# 모든 에러에 대해 Playwright 사용 시도
error_msg = str(e)
if isinstance(e, httpx.HTTPStatusError):
error_type = f"HTTP {e.response.status_code}"
elif isinstance(e, httpx.ConnectError):
error_type = "Connection Error"
elif isinstance(e, ssl.SSLError):
error_type = "SSL Error"
elif "resolve" in error_msg.lower() or "dns" in error_msg.lower():
error_type = "DNS Resolution Error"
else:
error_type = "Network Error"
print(f"{error_type} for {url}, trying with Playwright...")
# Playwright로 이미지 가져오기 시도
try:
from playwright.async_api import async_playwright
from PIL import Image
import io
async with async_playwright() as p:
# 브라우저 실행
browser = await p.chromium.launch(
headless=True,
args=['--no-sandbox', '--disable-setuid-sandbox']
)
# Referer 설정을 위한 도메인 추출
from urllib.parse import urlparse
parsed = urlparse(url)
referer_url = f"{parsed.scheme}://{parsed.netloc}/"
context = await browser.new_context(
viewport={'width': 1920, 'height': 1080},
user_agent='Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36',
extra_http_headers={
'Referer': referer_url
}
)
page = await context.new_page()
try:
# Response를 가로채기 위한 설정
image_data = None
async def handle_response(response):
nonlocal image_data
# 이미지 URL에 대한 응답 가로채기
if url in response.url or response.url == url:
try:
image_data = await response.body()
print(f"✅ Image intercepted: {len(image_data)} bytes")
except:
pass
# Response 이벤트 리스너 등록
page.on('response', handle_response)
# 이미지 URL로 이동 (에러 무시)
try:
await page.goto(url, wait_until='networkidle', timeout=30000)
except Exception as goto_error:
print(f"⚠️ Direct navigation failed: {goto_error}")
# 직접 이동 실패 시 HTML에 img 태그 삽입
await page.set_content(f'''
<html>
<body style="margin:0;padding:0;">
<img src="{url}" style="max-width:100%;height:auto;"
crossorigin="anonymous" />
</body>
</html>
''')
await page.wait_for_timeout(3000) # 이미지 로딩 대기
# 이미지 데이터가 없으면 JavaScript로 직접 fetch
if not image_data:
# JavaScript로 이미지 fetch
image_data_base64 = await page.evaluate('''
async (url) => {
try {
const response = await fetch(url);
const blob = await response.blob();
return new Promise((resolve) => {
const reader = new FileReader();
reader.onloadend = () => resolve(reader.result.split(',')[1]);
reader.readAsDataURL(blob);
});
} catch (e) {
return null;
}
}
''', url)
if image_data_base64:
import base64
image_data = base64.b64decode(image_data_base64)
print(f"✅ Image fetched via JavaScript: {len(image_data)} bytes")
# 여전히 데이터가 없으면 스크린샷 사용
if not image_data:
# 이미지 요소 찾기
img_element = await page.query_selector('img')
if img_element:
# 이미지가 로드되었는지 확인
is_loaded = await img_element.evaluate('(img) => img.complete && img.naturalHeight > 0')
if is_loaded:
image_data = await img_element.screenshot()
print(f"✅ Screenshot from loaded image: {len(image_data)} bytes")
else:
# 이미지 로드 대기
try:
await img_element.evaluate('(img) => new Promise(r => img.onload = r)')
image_data = await img_element.screenshot()
print(f"✅ Screenshot after waiting: {len(image_data)} bytes")
except:
# 전체 페이지 스크린샷
image_data = await page.screenshot(full_page=True)
print(f"⚠️ Full page screenshot: {len(image_data)} bytes")
else:
image_data = await page.screenshot(full_page=True)
print(f"⚠️ No image element, full screenshot: {len(image_data)} bytes")
print(f"✅ Successfully fetched image with Playwright: {url}")
return image_data
finally:
await page.close()
await context.close()
await browser.close()
except Exception as pw_error:
print(f"Playwright failed: {pw_error}, returning placeholder")
# Playwright도 실패하면 세련된 placeholder 반환
from PIL import Image, ImageDraw, ImageFont
import io
import random
# 그라디언트 배경색 선택 (부드러운 색상)
gradients = [
('#667eea', '#764ba2'), # 보라 그라디언트
('#f093fb', '#f5576c'), # 핑크 그라디언트
('#4facfe', '#00f2fe'), # 하늘색 그라디언트
('#43e97b', '#38f9d7'), # 민트 그라디언트
('#fa709a', '#fee140'), # 선셋 그라디언트
('#30cfd0', '#330867'), # 딥 오션
('#a8edea', '#fed6e3'), # 파스텔
('#ffecd2', '#fcb69f'), # 피치
]
# 랜덤 그라디언트 선택
color1, color2 = random.choice(gradients)
# 이미지 생성 (16:9 비율)
width, height = 800, 450
img = Image.new('RGB', (width, height))
draw = ImageDraw.Draw(img)
# 그라디언트 배경 생성
def hex_to_rgb(hex_color):
hex_color = hex_color.lstrip('#')
return tuple(int(hex_color[i:i+2], 16) for i in (0, 2, 4))
rgb1 = hex_to_rgb(color1)
rgb2 = hex_to_rgb(color2)
# 세로 그라디언트
for y in range(height):
ratio = y / height
r = int(rgb1[0] * (1 - ratio) + rgb2[0] * ratio)
g = int(rgb1[1] * (1 - ratio) + rgb2[1] * ratio)
b = int(rgb1[2] * (1 - ratio) + rgb2[2] * ratio)
draw.rectangle([(0, y), (width, y + 1)], fill=(r, g, b))
# 반투명 오버레이 추가 (깊이감)
overlay = Image.new('RGBA', (width, height), (0, 0, 0, 0))
overlay_draw = ImageDraw.Draw(overlay)
# 중앙 원형 그라디언트 효과
center_x, center_y = width // 2, height // 2
max_radius = min(width, height) // 3
for radius in range(max_radius, 0, -2):
opacity = int(255 * (1 - radius / max_radius) * 0.3)
overlay_draw.ellipse(
[(center_x - radius, center_y - radius),
(center_x + radius, center_y + radius)],
fill=(255, 255, 255, opacity)
)
# 이미지 아이콘 그리기 (산 모양)
icon_color = (255, 255, 255, 200)
icon_size = 80
icon_x = center_x
icon_y = center_y - 20
# 산 아이콘 (사진 이미지를 나타냄)
mountain_points = [
(icon_x - icon_size, icon_y + icon_size//2),
(icon_x - icon_size//2, icon_y - icon_size//4),
(icon_x - icon_size//4, icon_y),
(icon_x + icon_size//4, icon_y - icon_size//2),
(icon_x + icon_size, icon_y + icon_size//2),
]
overlay_draw.polygon(mountain_points, fill=icon_color)
# 태양/달 원
sun_radius = icon_size // 4
overlay_draw.ellipse(
[(icon_x - icon_size//2, icon_y - icon_size//2 - sun_radius),
(icon_x - icon_size//2 + sun_radius*2, icon_y - icon_size//2 + sun_radius)],
fill=icon_color
)
# 프레임 테두리
frame_margin = 40
overlay_draw.rectangle(
[(frame_margin, frame_margin),
(width - frame_margin, height - frame_margin)],
outline=(255, 255, 255, 150),
width=3
)
# 코너 장식
corner_size = 20
corner_width = 4
corners = [
(frame_margin, frame_margin),
(width - frame_margin - corner_size, frame_margin),
(frame_margin, height - frame_margin - corner_size),
(width - frame_margin - corner_size, height - frame_margin - corner_size)
]
for x, y in corners:
# 가로선
overlay_draw.rectangle(
[(x, y), (x + corner_size, y + corner_width)],
fill=(255, 255, 255, 200)
)
# 세로선
overlay_draw.rectangle(
[(x, y), (x + corner_width, y + corner_size)],
fill=(255, 255, 255, 200)
)
# "Image Loading..." 텍스트 (작게)
try:
# 시스템 폰트 시도
font = ImageFont.truetype("/usr/share/fonts/truetype/dejavu/DejaVuSans.ttf", 16)
except:
font = ImageFont.load_default()
text = "Image Loading..."
bbox = draw.textbbox((0, 0), text, font=font)
text_width = bbox[2] - bbox[0]
text_height = bbox[3] - bbox[1]
text_x = (width - text_width) // 2
text_y = center_y + icon_size
# 텍스트 그림자
for offset in [(2, 2), (-1, -1)]:
overlay_draw.text(
(text_x + offset[0], text_y + offset[1]),
text,
font=font,
fill=(0, 0, 0, 100)
)
# 텍스트 본체
overlay_draw.text(
(text_x, text_y),
text,
font=font,
fill=(255, 255, 255, 220)
)
# 오버레이 합성
img = Image.alpha_composite(img.convert('RGBA'), overlay).convert('RGB')
# 약간의 노이즈 추가 (텍스처)
pixels = img.load()
for _ in range(1000):
x = random.randint(0, width - 1)
y = random.randint(0, height - 1)
r, g, b = pixels[x, y]
brightness = random.randint(-20, 20)
pixels[x, y] = (
max(0, min(255, r + brightness)),
max(0, min(255, g + brightness)),
max(0, min(255, b + brightness))
)
# JPEG로 변환 (높은 품질)
output = io.BytesIO()
img.save(output, format='JPEG', quality=85, optimize=True)
return output.getvalue()
raise
# 이미지 크기 확인
content_length = int(response.headers.get('content-length', 0))
max_size = settings.max_image_size_mb * 1024 * 1024
if content_length > max_size:
raise ValueError(f"Image too large: {content_length} bytes")
# 응답 데이터 확인
content = response.content
print(f"Downloaded {len(content)} bytes from {url[:50]}...")
# gzip 압축 확인 및 해제
import gzip
if len(content) > 2 and content[:2] == b'\x1f\x8b':
print("📦 Gzip compressed data detected, decompressing...")
try:
content = gzip.decompress(content)
print(f"✅ Decompressed to {len(content)} bytes")
except Exception as e:
print(f"❌ Failed to decompress gzip: {e}")
# 처음 몇 바이트로 이미지 형식 확인
if len(content) > 10:
header = content[:12]
if header[:2] == b'\xff\xd8':
print("✅ JPEG image detected")
elif header[:8] == b'\x89PNG\r\n\x1a\n':
print("✅ PNG image detected")
elif header[:6] in (b'GIF87a', b'GIF89a'):
print("✅ GIF image detected")
elif header[:4] == b'RIFF' and header[8:12] == b'WEBP':
print("✅ WebP image detected")
elif b'<svg' in header or b'<?xml' in header:
print("✅ SVG image detected")
elif header[4:12] == b'ftypavif':
print("✅ AVIF image detected")
else:
print(f"⚠️ Unknown image format. Header: {header.hex()}")
return content
def resize_and_optimize_image(self, image_data: bytes, size: str) -> tuple[bytes, str]:
"""이미지 리사이징 및 최적화"""
if size not in settings.thumbnail_sizes:
raise ValueError(f"Invalid size: {size}")
target_size = settings.thumbnail_sizes[size]
# SVG 체크 - SVG는 리사이징하지 않고 그대로 반환
if self._is_svg(image_data):
return image_data, 'image/svg+xml'
# PIL로 이미지 열기
try:
img = Image.open(io.BytesIO(image_data))
except Exception as e:
# WebP 헤더 체크 (RIFF....WEBP)
header = image_data[:12] if len(image_data) >= 12 else image_data
if header[:4] == b'RIFF' and header[8:12] == b'WEBP':
print("🎨 WebP 이미지 감지됨, 변환 시도")
# WebP 형식이지만 PIL이 열지 못하는 경우
# Pillow-SIMD 또는 추가 라이브러리가 필요할 수 있음
try:
# 재시도
from PIL import WebPImagePlugin
img = Image.open(io.BytesIO(image_data))
except:
print("❌ WebP 이미지를 열 수 없음, 원본 반환")
return image_data, 'image/webp'
else:
raise e
# GIF 애니메이션 체크 및 처리
if getattr(img, "format", None) == "GIF":
return self._process_gif(image_data, target_size)
# WebP 형식 체크
original_format = getattr(img, "format", None)
is_webp = original_format == "WEBP"
# 원본 모드와 투명도 정보 저장
original_mode = img.mode
original_has_transparency = img.mode in ('RGBA', 'LA')
original_has_palette = img.mode == 'P'
# 팔레트 모드(P) 처리 - 간단하게 PIL의 기본 변환 사용
if img.mode == 'P':
# 팔레트 모드는 RGB로 직접 변환
# PIL의 convert 메서드가 팔레트를 올바르게 처리함
img = img.convert('RGB')
# 투명도가 있는 이미지 처리
if img.mode == 'RGBA':
# RGBA는 흰색 배경과 합성
background = Image.new('RGB', img.size, (255, 255, 255))
background.paste(img, mask=img.split()[-1])
img = background
elif img.mode == 'LA':
# LA(그레이스케일+알파)는 RGBA를 거쳐 RGB로
img = img.convert('RGBA')
background = Image.new('RGB', img.size, (255, 255, 255))
background.paste(img, mask=img.split()[-1])
img = background
elif img.mode == 'L':
# 그레이스케일은 RGB로 변환
img = img.convert('RGB')
elif img.mode not in ('RGB',):
# 기타 모드는 모두 RGB로 변환
img = img.convert('RGB')
# EXIF 방향 정보 처리 (RGB 변환 후에 수행)
try:
from PIL import ImageOps
img = ImageOps.exif_transpose(img)
except:
pass
# 메타데이터 제거는 스킵 (팔레트 모드 이미지에서 문제 발생)
# RGB로 변환되었으므로 이미 메타데이터는 대부분 제거됨
# 비율 유지하며 리사이징 (크롭 없이)
img_ratio = img.width / img.height
target_width = target_size[0]
target_height = target_size[1]
# 원본 비율을 유지하면서 목표 크기에 맞추기
# 너비 또는 높이 중 하나를 기준으로 비율 계산
if img.width > target_width or img.height > target_height:
# 너비 기준 리사이징
width_ratio = target_width / img.width
# 높이 기준 리사이징
height_ratio = target_height / img.height
# 둘 중 작은 비율 사용 (목표 크기를 넘지 않도록)
ratio = min(width_ratio, height_ratio)
new_width = int(img.width * ratio)
new_height = int(img.height * ratio)
# 큰 이미지를 작게 만들 때는 2단계 리샘플링으로 품질 향상
if img.width > new_width * 2 or img.height > new_height * 2:
# 1단계: 목표 크기의 2배로 먼저 축소
intermediate_width = new_width * 2
intermediate_height = new_height * 2
img = img.resize((intermediate_width, intermediate_height), Image.Resampling.LANCZOS)
# 최종 목표 크기로 리샘플링
img = img.resize((new_width, new_height), Image.Resampling.LANCZOS)
# 샤프닝 적용 (작은 이미지에만)
if target_size[0] <= 400:
from PIL import ImageEnhance
enhancer = ImageEnhance.Sharpness(img)
img = enhancer.enhance(1.2)
# 바이트로 변환
output = io.BytesIO()
# 적응형 품질 계산 (이미지 크기에 따라 조정)
def get_adaptive_quality(base_quality: int, target_width: int) -> int:
"""이미지 크기에 따른 적응형 품질 계산"""
# 품질을 더 높게 설정하여 검정색 문제 해결
if target_width <= 150: # 썸네일
return min(base_quality + 10, 95)
elif target_width <= 360: # 카드
return min(base_quality + 5, 90)
elif target_width <= 800: # 상세
return base_quality # 85
else: # 히어로
return base_quality # 85
# WebP 변환 및 최적화 - 최고 압축률 설정
# WebP 입력은 JPEG로 변환 (WebP 리사이징 문제 회피)
if is_webp:
output_format = 'JPEG'
content_type = 'image/jpeg'
else:
output_format = 'WEBP' if settings.convert_to_webp else 'JPEG'
content_type = 'image/webp' if output_format == 'WEBP' else 'image/jpeg'
if output_format == 'WEBP':
# WebP 최적화: method=6(최고품질), lossless=False, exact=False
adaptive_quality = get_adaptive_quality(settings.webp_quality, target_size[0])
save_kwargs = {
'format': 'WEBP',
'quality': adaptive_quality,
'method': 6, # 최고 압축 알고리즘 (0-6)
'lossless': settings.webp_lossless,
'exact': False, # 약간의 품질 손실 허용하여 더 작은 크기
}
img.save(output, **save_kwargs)
elif original_has_transparency and not settings.convert_to_webp:
# PNG 최적화 (투명도가 있는 이미지)
save_kwargs = {
'format': 'PNG',
'optimize': settings.optimize_png,
'compress_level': settings.png_compress_level,
}
# 팔레트 모드로 변환 가능한지 확인 (256색 이하)
if settings.optimize_png:
try:
# 색상 수가 256개 이하이면 팔레트 모드로 변환
quantized = img.quantize(colors=256, method=Image.Quantize.MEDIANCUT)
if len(quantized.getcolors()) <= 256:
img = quantized
save_kwargs['format'] = 'PNG'
except:
pass
content_type = 'image/png'
img.save(output, **save_kwargs)
else:
# JPEG 최적화 설정 (기본값)
adaptive_quality = get_adaptive_quality(settings.jpeg_quality, target_size[0])
save_kwargs = {
'format': 'JPEG',
'quality': adaptive_quality,
'optimize': True,
'progressive': settings.progressive_jpeg,
}
img.save(output, **save_kwargs)
return output.getvalue(), content_type
async def get_cache_size(self) -> float:
"""현재 캐시 크기 (GB)"""
total_size = 0
for dirpath, dirnames, filenames in os.walk(self.cache_dir):
for filename in filenames:
filepath = os.path.join(dirpath, filename)
total_size += os.path.getsize(filepath)
return total_size / (1024 ** 3) # GB로 변환
async def cleanup_old_cache(self):
"""오래된 캐시 파일 정리"""
cutoff_time = datetime.now() - timedelta(days=settings.cache_ttl_days)
for dirpath, dirnames, filenames in os.walk(self.cache_dir):
for filename in filenames:
filepath = Path(dirpath) / filename
if filepath.stat().st_mtime < cutoff_time.timestamp():
filepath.unlink()
async def trigger_background_generation(self, url: str):
"""백그라운드에서 모든 크기의 이미지 생성 트리거"""
from .background_tasks import background_manager
# 백그라운드 작업 큐에 추가
asyncio.create_task(background_manager.add_task(url))
async def get_directory_stats(self) -> dict:
"""디렉토리 구조 통계 정보"""
total_files = 0
total_dirs = 0
files_per_dir = {}
for root, dirs, files in os.walk(self.cache_dir):
total_dirs += len(dirs)
total_files += len(files)
# 각 디렉토리의 파일 수 계산
rel_path = os.path.relpath(root, self.cache_dir)
depth = len(Path(rel_path).parts) if rel_path != '.' else 0
if files and depth == 3: # 3단계 디렉토리에서만 파일 수 계산
files_per_dir[rel_path] = len(files)
# 통계 계산
avg_files_per_dir = sum(files_per_dir.values()) / len(files_per_dir) if files_per_dir else 0
max_files_in_dir = max(files_per_dir.values()) if files_per_dir else 0
return {
"total_files": total_files,
"total_directories": total_dirs,
"average_files_per_directory": round(avg_files_per_dir, 2),
"max_files_in_single_directory": max_files_in_dir,
"directory_depth": 3
}
cache = ImageCache()

View File

@ -0,0 +1,54 @@
from pydantic_settings import BaseSettings
from pathlib import Path
class Settings(BaseSettings):
# 기본 설정
app_name: str = "Image Proxy Service"
debug: bool = True
# 캐시 설정 (MinIO 전환 시에도 로컬 임시 파일용)
cache_dir: Path = Path("/app/cache")
max_cache_size_gb: int = 10
cache_ttl_days: int = 30
# MinIO 설정
use_minio: bool = True # MinIO 사용 여부
minio_endpoint: str = "minio:9000"
minio_access_key: str = "minioadmin"
minio_secret_key: str = "minioadmin"
minio_bucket_name: str = "image-cache"
minio_secure: bool = False
# 이미지 설정
max_image_size_mb: int = 20
allowed_formats: list = ["jpg", "jpeg", "png", "gif", "webp", "svg"]
# 리사이징 설정 - 뉴스 카드 용도별 최적화
thumbnail_sizes: dict = {
"thumb": (150, 100), # 작은 썸네일 (3:2 비율)
"card": (360, 240), # 뉴스 카드용 (3:2 비율)
"list": (300, 200), # 리스트용 (3:2 비율)
"detail": (800, 533), # 상세 페이지용 (원본 비율 유지)
"hero": (1200, 800) # 히어로 이미지용 (원본 비율 유지)
}
# 이미지 최적화 설정 - 품질 보장하면서 최저 용량
jpeg_quality: int = 85 # JPEG 품질 (품질 향상)
webp_quality: int = 85 # WebP 품질 (품질 향상으로 검정색 문제 해결)
webp_lossless: bool = False # 무손실 압축 비활성화 (용량 최적화)
png_compress_level: int = 9 # PNG 최대 압축 (0-9, 9가 최고 압축)
convert_to_webp: bool = False # WebP 변환 임시 비활성화 (검정색 이미지 문제)
# 고급 최적화 설정
progressive_jpeg: bool = True # 점진적 JPEG (로딩 성능 향상)
strip_metadata: bool = True # EXIF 등 메타데이터 제거 (용량 절약)
optimize_png: bool = True # PNG 팔레트 최적화
# 외부 요청 설정
request_timeout: int = 30
user_agent: str = "ImageProxyService/1.0"
class Config:
env_file = ".env"
settings = Settings()

View File

@ -0,0 +1,414 @@
import hashlib
import os
from pathlib import Path
from datetime import datetime, timedelta
from typing import Optional, Tuple
import httpx
from PIL import Image
try:
from pillow_heif import register_heif_opener, register_avif_opener
register_heif_opener() # HEIF/HEIC 지원
register_avif_opener() # AVIF 지원
print("HEIF/AVIF support enabled successfully")
except ImportError:
print("Warning: pillow_heif not installed, HEIF/AVIF support disabled")
import io
import asyncio
import ssl
from minio import Minio
from minio.error import S3Error
import tempfile
from .config import settings
class MinIOImageCache:
def __init__(self):
# MinIO 클라이언트 초기화
self.client = Minio(
settings.minio_endpoint,
access_key=settings.minio_access_key,
secret_key=settings.minio_secret_key,
secure=settings.minio_secure
)
# 버킷 생성 (동기 호출)
self._ensure_bucket()
# 로컬 임시 디렉토리 (이미지 처리용)
self.temp_dir = Path(tempfile.gettempdir()) / "image_cache_temp"
self.temp_dir.mkdir(parents=True, exist_ok=True)
def _ensure_bucket(self):
"""버킷이 존재하는지 확인하고 없으면 생성"""
try:
if not self.client.bucket_exists(settings.minio_bucket_name):
self.client.make_bucket(settings.minio_bucket_name)
print(f"✅ Created MinIO bucket: {settings.minio_bucket_name}")
else:
print(f"✅ MinIO bucket exists: {settings.minio_bucket_name}")
except S3Error as e:
print(f"❌ Error creating bucket: {e}")
def _get_object_name(self, url: str, size: Optional[str] = None) -> str:
"""URL을 기반으로 MinIO 객체 이름 생성"""
url_hash = hashlib.md5(url.encode()).hexdigest()
# 3단계 디렉토리 구조 생성 (MinIO는 /를 디렉토리처럼 취급)
level1 = url_hash[:2]
level2 = url_hash[2:4]
level3 = url_hash[4:6]
# 크기별로 다른 파일명 사용
if size:
filename = f"{url_hash}_{size}"
else:
filename = url_hash
# 확장자 추출 (WebP로 저장되는 경우 .webp 사용)
if settings.convert_to_webp and size:
filename = f"{filename}.webp"
else:
ext = self._get_extension_from_url(url)
if ext:
filename = f"{filename}.{ext}"
# MinIO 객체 경로 생성
object_name = f"{level1}/{level2}/{level3}/{filename}"
return object_name
def _get_extension_from_url(self, url: str) -> Optional[str]:
"""URL에서 파일 확장자 추출"""
path = url.split('?')[0] # 쿼리 파라미터 제거
parts = path.split('.')
if len(parts) > 1:
ext = parts[-1].lower()
if ext in settings.allowed_formats:
return ext
return None
def _is_svg(self, data: bytes) -> bool:
"""SVG 파일인지 확인"""
if len(data) < 100:
return False
header = data[:1000].lower()
svg_signatures = [
b'<svg',
b'<?xml',
b'<!doctype svg'
]
for sig in svg_signatures:
if sig in header:
return True
return False
def _process_gif(self, gif_data: bytes, target_size: tuple) -> tuple[bytes, str]:
"""GIF 처리 - JPEG로 변환하여 안정적으로 처리"""
try:
img = Image.open(io.BytesIO(gif_data))
if img.mode != 'RGB':
if img.mode == 'P':
img = img.convert('RGBA')
if img.mode == 'RGBA':
background = Image.new('RGB', img.size, (255, 255, 255))
background.paste(img, mask=img.split()[3] if len(img.split()) == 4 else None)
img = background
elif img.mode != 'RGB':
img = img.convert('RGB')
# 리사이즈
img.thumbnail(target_size, Image.Resampling.LANCZOS)
# JPEG로 저장
output = io.BytesIO()
img.save(
output,
format='JPEG',
quality=settings.jpeg_quality,
optimize=True,
progressive=settings.progressive_jpeg
)
return output.getvalue(), 'image/jpeg'
except Exception as e:
print(f"GIF 처리 오류: {e}")
return gif_data, 'image/gif'
def resize_and_optimize_image(self, image_data: bytes, size: str) -> tuple[bytes, str]:
"""이미지 리사이징 및 최적화"""
try:
target_size = settings.thumbnail_sizes.get(size, settings.thumbnail_sizes["thumb"])
# 이미지 열기
img = Image.open(io.BytesIO(image_data))
# EXIF 회전 정보 처리
try:
from PIL import ImageOps
img = ImageOps.exif_transpose(img)
except:
pass
# 리사이즈 (원본 비율 유지)
img.thumbnail(target_size, Image.Resampling.LANCZOS)
# 출력 버퍼
output = io.BytesIO()
# WebP로 변환 설정이 활성화되어 있으면
if settings.convert_to_webp:
# RGBA를 RGB로 변환 (WebP는 투명도 지원하지만 일부 브라우저 호환성 문제)
if img.mode in ('RGBA', 'LA', 'P'):
# 투명 배경을 흰색으로
background = Image.new('RGB', img.size, (255, 255, 255))
if img.mode == 'P':
img = img.convert('RGBA')
background.paste(img, mask=img.split()[-1] if 'A' in img.mode else None)
img = background
elif img.mode != 'RGB':
img = img.convert('RGB')
# WebP로 저장
img.save(
output,
format='WEBP',
quality=settings.webp_quality,
lossless=settings.webp_lossless,
method=6 # 최고 압축
)
content_type = 'image/webp'
else:
# 원본 포맷 유지하면서 최적화
if img.format == 'PNG':
img.save(
output,
format='PNG',
compress_level=settings.png_compress_level,
optimize=settings.optimize_png
)
content_type = 'image/png'
else:
# JPEG로 변환
if img.mode != 'RGB':
img = img.convert('RGB')
img.save(
output,
format='JPEG',
quality=settings.jpeg_quality,
optimize=True,
progressive=settings.progressive_jpeg
)
content_type = 'image/jpeg'
return output.getvalue(), content_type
except Exception as e:
print(f"이미지 최적화 오류: {e}")
import traceback
traceback.print_exc()
return image_data, 'image/jpeg'
async def get(self, url: str, size: Optional[str] = None) -> Optional[bytes]:
"""MinIO에서 캐시된 이미지 가져오기"""
object_name = self._get_object_name(url, size)
try:
# MinIO에서 객체 가져오기
response = self.client.get_object(settings.minio_bucket_name, object_name)
data = response.read()
response.close()
response.release_conn()
print(f"✅ Cache HIT from MinIO: {object_name}")
return data
except S3Error as e:
if e.code == 'NoSuchKey':
print(f"📭 Cache MISS in MinIO: {object_name}")
return None
else:
print(f"❌ MinIO error: {e}")
return None
async def set(self, url: str, data: bytes, size: Optional[str] = None):
"""MinIO에 이미지 캐시 저장"""
object_name = self._get_object_name(url, size)
try:
# 바이트 데이터를 스트림으로 변환
data_stream = io.BytesIO(data)
data_length = len(data)
# content-type 결정
if url.lower().endswith('.svg') or self._is_svg(data):
content_type = 'image/svg+xml'
elif url.lower().endswith('.gif'):
content_type = 'image/gif'
elif settings.convert_to_webp and size:
content_type = 'image/webp'
else:
content_type = 'application/octet-stream'
# MinIO에 저장 (메타데이터는 ASCII만 지원하므로 URL 해시 사용)
self.client.put_object(
settings.minio_bucket_name,
object_name,
data_stream,
data_length,
content_type=content_type,
metadata={
'url_hash': hashlib.md5(url.encode()).hexdigest(),
'cached_at': datetime.utcnow().isoformat(),
'size_variant': size or 'original'
}
)
print(f"✅ Cached to MinIO: {object_name} ({data_length} bytes)")
except S3Error as e:
print(f"❌ Failed to cache to MinIO: {e}")
async def download_image(self, url: str) -> bytes:
"""외부 URL에서 이미지 다운로드"""
# SSL 검증 비활성화 (개발 환경용)
ssl_context = ssl.create_default_context()
ssl_context.check_hostname = False
ssl_context.verify_mode = ssl.CERT_NONE
async with httpx.AsyncClient(
timeout=settings.request_timeout,
verify=False,
follow_redirects=True
) as client:
headers = {
"User-Agent": settings.user_agent,
"Accept": "image/webp,image/apng,image/*,*/*;q=0.8",
"Accept-Encoding": "gzip, deflate, br",
"Cache-Control": "no-cache",
"Referer": url.split('/')[0] + '//' + url.split('/')[2] if len(url.split('/')) > 2 else url
}
response = await client.get(url, headers=headers)
if response.status_code == 403:
headers["User-Agent"] = "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36"
response = await client.get(url, headers=headers)
response.raise_for_status()
content_length = response.headers.get("content-length")
if content_length:
size_mb = int(content_length) / (1024 * 1024)
if size_mb > settings.max_image_size_mb:
raise ValueError(f"이미지 크기가 {settings.max_image_size_mb}MB를 초과합니다")
return response.content
async def get_cache_size(self) -> float:
"""MinIO 버킷 크기 조회 (GB)"""
try:
total_size = 0
objects = self.client.list_objects(settings.minio_bucket_name, recursive=True)
for obj in objects:
total_size += obj.size
return total_size / (1024 ** 3) # GB로 변환
except S3Error as e:
print(f"❌ Failed to get cache size: {e}")
return 0.0
async def get_directory_stats(self) -> dict:
"""MinIO 디렉토리 구조 통계"""
try:
total_files = 0
directories = set()
objects = self.client.list_objects(settings.minio_bucket_name, recursive=True)
for obj in objects:
total_files += 1
# 디렉토리 경로 추출
parts = obj.object_name.split('/')
if len(parts) > 1:
dir_path = '/'.join(parts[:-1])
directories.add(dir_path)
return {
"total_files": total_files,
"total_directories": len(directories),
"average_files_per_directory": total_files / max(len(directories), 1),
"bucket_name": settings.minio_bucket_name
}
except S3Error as e:
print(f"❌ Failed to get directory stats: {e}")
return {
"total_files": 0,
"total_directories": 0,
"average_files_per_directory": 0,
"bucket_name": settings.minio_bucket_name
}
async def cleanup_old_cache(self):
"""오래된 캐시 정리"""
try:
cutoff_date = datetime.utcnow() - timedelta(days=settings.cache_ttl_days)
deleted_count = 0
objects = self.client.list_objects(settings.minio_bucket_name, recursive=True)
for obj in objects:
# 객체의 마지막 수정 시간이 cutoff_date 이전이면 삭제
if obj.last_modified.replace(tzinfo=None) < cutoff_date:
self.client.remove_object(settings.minio_bucket_name, obj.object_name)
deleted_count += 1
print(f"🗑️ Deleted old cache: {obj.object_name}")
print(f"✅ Cleaned up {deleted_count} old cached files")
return deleted_count
except S3Error as e:
print(f"❌ Failed to cleanup cache: {e}")
return 0
async def trigger_background_generation(self, url: str):
"""백그라운드에서 다양한 크기 생성"""
asyncio.create_task(self._generate_all_sizes(url))
async def _generate_all_sizes(self, url: str):
"""모든 크기 버전 생성"""
try:
# 원본 이미지 다운로드
image_data = await self.download_image(url)
# SVG는 리사이징 불필요
if self._is_svg(image_data):
return
# 모든 크기 생성
for size_name in settings.thumbnail_sizes.keys():
# 이미 캐시되어 있는지 확인
existing = await self.get(url, size_name)
if not existing:
# 리사이징 및 최적화
if url.lower().endswith('.gif'):
resized_data, _ = self._process_gif(image_data, settings.thumbnail_sizes[size_name])
else:
resized_data, _ = self.resize_and_optimize_image(image_data, size_name)
# 캐시에 저장
await self.set(url, resized_data, size_name)
print(f"✅ Generated {size_name} version for {url}")
except Exception as e:
print(f"❌ Background generation failed for {url}: {e}")
# 싱글톤 인스턴스
cache = MinIOImageCache()

View File

@ -0,0 +1,65 @@
from fastapi import FastAPI
from fastapi.middleware.cors import CORSMiddleware
from contextlib import asynccontextmanager
import uvicorn
from datetime import datetime
from app.api.endpoints import router
from app.core.config import settings
@asynccontextmanager
async def lifespan(app: FastAPI):
# 시작 시
print("Images service starting...")
yield
# 종료 시
print("Images service stopping...")
app = FastAPI(
title="Images Service",
description="이미지 업로드, 프록시 및 캐싱 서비스",
version="2.0.0",
lifespan=lifespan
)
# CORS 설정
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
# 라우터 등록
app.include_router(router, prefix="/api/v1")
@app.get("/")
async def root():
return {
"service": "Images Service",
"version": "2.0.0",
"timestamp": datetime.now().isoformat(),
"endpoints": {
"proxy": "/api/v1/image?url=<image_url>&size=<optional_size>",
"upload": "/api/v1/upload",
"stats": "/api/v1/stats",
"cleanup": "/api/v1/cleanup"
}
}
@app.get("/health")
async def health_check():
return {
"status": "healthy",
"service": "images",
"timestamp": datetime.now().isoformat()
}
if __name__ == "__main__":
uvicorn.run(
"main:app",
host="0.0.0.0",
port=8000,
reload=True
)

View File

@ -0,0 +1,12 @@
fastapi==0.109.0
uvicorn[standard]==0.27.0
httpx==0.26.0
pillow==10.2.0
pillow-heif==0.20.0
aiofiles==23.2.1
python-multipart==0.0.6
pydantic==2.5.3
pydantic-settings==2.1.0
motor==3.3.2
redis==5.0.1
minio==7.2.3

View File

@ -0,0 +1,21 @@
FROM python:3.11-slim
WORKDIR /app
# Install system dependencies
RUN apt-get update && apt-get install -y \
gcc \
&& rm -rf /var/lib/apt/lists/*
# Copy requirements first for better caching
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
# Copy application code
COPY . .
# Expose port
EXPOSE 8000
# Run the application
CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "8000", "--reload"]

View File

@ -0,0 +1,335 @@
"""
Channel Handlers for different notification delivery methods
"""
import logging
import asyncio
from typing import Optional, Dict, Any
from models import Notification, NotificationStatus
import smtplib
from email.mime.text import MIMEText
from email.mime.multipart import MIMEMultipart
import httpx
import json
logger = logging.getLogger(__name__)
class BaseChannelHandler:
"""Base class for channel handlers"""
async def send(self, notification: Notification) -> bool:
"""Send notification through the channel"""
raise NotImplementedError
async def verify_delivery(self, notification: Notification) -> bool:
"""Verify if notification was delivered"""
return True
class EmailHandler(BaseChannelHandler):
"""Email notification handler"""
def __init__(self, smtp_host: str, smtp_port: int, smtp_user: str, smtp_password: str):
self.smtp_host = smtp_host
self.smtp_port = smtp_port
self.smtp_user = smtp_user
self.smtp_password = smtp_password
async def send(self, notification: Notification) -> bool:
"""Send email notification"""
try:
# In production, would use async SMTP library
# For demo, we'll simulate email sending
logger.info(f"Sending email to user {notification.user_id}")
if not self.smtp_user or not self.smtp_password:
# Simulate sending without actual SMTP config
await asyncio.sleep(0.1) # Simulate network delay
logger.info(f"Email sent (simulated) to user {notification.user_id}")
return True
# Create message
msg = MIMEMultipart()
msg['From'] = self.smtp_user
msg['To'] = f"user_{notification.user_id}@example.com" # Would fetch actual email
msg['Subject'] = notification.title
# Add body
body = notification.message
if notification.data and "html_content" in notification.data:
msg.attach(MIMEText(notification.data["html_content"], 'html'))
else:
msg.attach(MIMEText(body, 'plain'))
# Send email (would be async in production)
# server = smtplib.SMTP(self.smtp_host, self.smtp_port)
# server.starttls()
# server.login(self.smtp_user, self.smtp_password)
# server.send_message(msg)
# server.quit()
logger.info(f"Email sent successfully to user {notification.user_id}")
return True
except Exception as e:
logger.error(f"Failed to send email: {e}")
return False
class SMSHandler(BaseChannelHandler):
"""SMS notification handler"""
def __init__(self, api_key: str, api_url: str):
self.api_key = api_key
self.api_url = api_url
self.client = httpx.AsyncClient()
async def send(self, notification: Notification) -> bool:
"""Send SMS notification"""
try:
# In production, would integrate with SMS provider (Twilio, etc.)
logger.info(f"Sending SMS to user {notification.user_id}")
if not self.api_key or not self.api_url:
# Simulate sending without actual API config
await asyncio.sleep(0.1) # Simulate network delay
logger.info(f"SMS sent (simulated) to user {notification.user_id}")
return True
# Would fetch user's phone number from database
phone_number = notification.data.get("phone") if notification.data else None
if not phone_number:
phone_number = "+1234567890" # Demo number
# Send SMS via API (example structure)
payload = {
"to": phone_number,
"message": f"{notification.title}\n{notification.message}",
"api_key": self.api_key
}
# response = await self.client.post(self.api_url, json=payload)
# return response.status_code == 200
# Simulate success
await asyncio.sleep(0.1)
logger.info(f"SMS sent successfully to user {notification.user_id}")
return True
except Exception as e:
logger.error(f"Failed to send SMS: {e}")
return False
class PushHandler(BaseChannelHandler):
"""Push notification handler (FCM/APNS)"""
def __init__(self, fcm_server_key: str):
self.fcm_server_key = fcm_server_key
self.fcm_url = "https://fcm.googleapis.com/fcm/send"
self.client = httpx.AsyncClient()
async def send(self, notification: Notification) -> bool:
"""Send push notification"""
try:
logger.info(f"Sending push notification to user {notification.user_id}")
if not self.fcm_server_key:
# Simulate sending without actual FCM config
await asyncio.sleep(0.1)
logger.info(f"Push notification sent (simulated) to user {notification.user_id}")
return True
# Would fetch user's device tokens from database
device_tokens = notification.data.get("device_tokens", []) if notification.data else []
if not device_tokens:
# Simulate with dummy token
device_tokens = ["dummy_token"]
# Send to each device token
for token in device_tokens:
payload = {
"to": token,
"notification": {
"title": notification.title,
"body": notification.message,
"icon": notification.data.get("icon") if notification.data else None,
"click_action": notification.data.get("click_action") if notification.data else None
},
"data": notification.data or {}
}
headers = {
"Authorization": f"key={self.fcm_server_key}",
"Content-Type": "application/json"
}
# response = await self.client.post(
# self.fcm_url,
# json=payload,
# headers=headers
# )
# Simulate success
await asyncio.sleep(0.05)
logger.info(f"Push notification sent successfully to user {notification.user_id}")
return True
except Exception as e:
logger.error(f"Failed to send push notification: {e}")
return False
class InAppHandler(BaseChannelHandler):
"""In-app notification handler"""
def __init__(self):
self.ws_server = None
def set_ws_server(self, ws_server):
"""Set WebSocket server for real-time delivery"""
self.ws_server = ws_server
async def send(self, notification: Notification) -> bool:
"""Send in-app notification"""
try:
logger.info(f"Sending in-app notification to user {notification.user_id}")
# Store notification in database (already done in manager)
# This would be retrieved when user logs in or requests notifications
# If WebSocket connection exists, send real-time
if self.ws_server:
await self.ws_server.send_to_user(
notification.user_id,
{
"type": "notification",
"notification": {
"id": notification.id,
"title": notification.title,
"message": notification.message,
"priority": notification.priority.value,
"category": notification.category.value if hasattr(notification, 'category') else "system",
"timestamp": notification.created_at.isoformat(),
"data": notification.data
}
}
)
logger.info(f"In-app notification sent successfully to user {notification.user_id}")
return True
except Exception as e:
logger.error(f"Failed to send in-app notification: {e}")
return False
class SlackHandler(BaseChannelHandler):
"""Slack notification handler"""
def __init__(self, webhook_url: Optional[str] = None):
self.webhook_url = webhook_url
self.client = httpx.AsyncClient()
async def send(self, notification: Notification) -> bool:
"""Send Slack notification"""
try:
logger.info(f"Sending Slack notification for user {notification.user_id}")
if not self.webhook_url:
# Simulate sending
await asyncio.sleep(0.1)
logger.info(f"Slack notification sent (simulated) for user {notification.user_id}")
return True
# Format message for Slack
slack_message = {
"text": notification.title,
"blocks": [
{
"type": "header",
"text": {
"type": "plain_text",
"text": notification.title
}
},
{
"type": "section",
"text": {
"type": "mrkdwn",
"text": notification.message
}
}
]
}
# Add additional fields if present
if notification.data:
fields = []
for key, value in notification.data.items():
if key not in ["html_content", "device_tokens"]:
fields.append({
"type": "mrkdwn",
"text": f"*{key}:* {value}"
})
if fields:
slack_message["blocks"].append({
"type": "section",
"fields": fields[:10] # Slack limits to 10 fields
})
# Send to Slack
# response = await self.client.post(self.webhook_url, json=slack_message)
# return response.status_code == 200
await asyncio.sleep(0.1)
logger.info(f"Slack notification sent successfully")
return True
except Exception as e:
logger.error(f"Failed to send Slack notification: {e}")
return False
class WebhookHandler(BaseChannelHandler):
"""Generic webhook notification handler"""
def __init__(self, default_webhook_url: Optional[str] = None):
self.default_webhook_url = default_webhook_url
self.client = httpx.AsyncClient()
async def send(self, notification: Notification) -> bool:
"""Send webhook notification"""
try:
# Get webhook URL from notification data or use default
webhook_url = None
if notification.data and "webhook_url" in notification.data:
webhook_url = notification.data["webhook_url"]
else:
webhook_url = self.default_webhook_url
if not webhook_url:
logger.warning("No webhook URL configured")
return False
logger.info(f"Sending webhook notification for user {notification.user_id}")
# Prepare payload
payload = {
"notification_id": notification.id,
"user_id": notification.user_id,
"title": notification.title,
"message": notification.message,
"priority": notification.priority.value,
"timestamp": notification.created_at.isoformat(),
"data": notification.data
}
# Send webhook
# response = await self.client.post(webhook_url, json=payload)
# return response.status_code in [200, 201, 202, 204]
# Simulate success
await asyncio.sleep(0.1)
logger.info(f"Webhook notification sent successfully")
return True
except Exception as e:
logger.error(f"Failed to send webhook notification: {e}")
return False

Some files were not shown because too many files have changed in this diff Show More