[Feat] 업로드한 파일의 중복 검사

src/app.py

@@ -3,12 +3,20 @@
 from fastapi.exceptions import RequestValidationError
 from fastapi.responses import JSONResponse
 from dotenv import load_dotenv
+import logging
 
 from src.router import router
 
 
 load_dotenv()
 
+# 로깅 설정
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
+    handlers=[logging.StreamHandler()]
+)
+
 origins = [
     "https://d6jo3bhmz1u5k.cloudfront.net",
     "https://localhost:5173",

src/main/ai/data/FileDuplicateCheckQueue.py

@@ -0,0 +1,27 @@
+import json
+
+class FileDuplicateCheckQueue:
+    def __init__(self, sqs_client, queue_url):
+        self.sqs_client = sqs_client
+        self.queue_url = queue_url
+
+    def send_message(self, request_id: str, user_id: str, s3_bucket: str, s3_key: str):
+        """SQS 큐에 메시지를 전송합니다."""
+        message_body = {
+            'request_type': 'file_duplicate_check_embedding_file',
+            'request_id': request_id,
+            'user_id': user_id,
+            'payload': {
+                's3_bucket': s3_bucket,
+                's3_key': s3_key
+            }
+        }
+
+        response = self.sqs_client.send_message(
+            QueueUrl=self.queue_url,
+            MessageGroupId=str(user_id),
+            MessageDeduplicationId=str(request_id),
+            MessageBody=json.dumps(message_body),
+        )
+
+        return response 

src/main/ai/data/FileDuplicateCheckRepository.py

@@ -0,0 +1,100 @@
+from bson import ObjectId
+from datetime import datetime, timezone
+
+
+class FileDuplicateCheckRepository:
+    def __init__(self, mongo_client):
+        self.mongo_client = mongo_client
+        self.db = mongo_client.get_database()
+        self.file_checks_collection = self.db.get_collection("file_duplicate_checks")
+        self.files_collection = self.db.get_collection("files")
+        self.file_embeddings_collection = self.db.get_collection("file_embeddings")
+
+    def get_current_time(self):
+        """현재 시간을 UTC 기준으로 반환합니다."""
+        return datetime.now(timezone.utc)
+
+    def get_file_by_id(self, file_id: str):
+        """파일 ID로 파일 정보를 조회합니다."""
+        try:
+            file_obj_id = ObjectId(file_id)
+            return self.files_collection.find_one({"_id": file_obj_id})
+        except Exception as e:
+            return None
+
+    def has_file_embedding(self, file_id: str) -> bool:
+        """파일의 임베딩 존재 여부를 확인합니다."""
+        try:
+            file_obj_id = ObjectId(file_id)
+            result = self.file_embeddings_collection.find_one({"file_id": file_obj_id})
+            return result is not None
+        except Exception as e:
+            return False
+
+    def create_duplicate_check_request(self, file_id: str, user_id: str):
+        """중복 검사 요청을 생성합니다."""
+        now = self.get_current_time()
+        document = {
+            "file_id": file_id,
+            "user_id": user_id,
+            "is_completed": False,
+            "is_duplicated": None,
+            "created_at": now
+        }
+
+        result = self.file_checks_collection.insert_one(document)
+
+        return self.file_checks_collection.find_one({"_id": result.inserted_id})
+
+    def get_duplicate_check_by_file_id(self, file_id: str, user_id: str):
+        """파일 ID와 사용자 ID로 중복 검사 요청을 조회합니다."""
+        return self.file_checks_collection.find_one({
+            "file_id": file_id,
+            "user_id": user_id
+        })
+
+    def get_duplicate_check_by_id(self, request_id: str):
+        """요청 ID로 중복 검사 요청을 조회합니다."""
+        try:
+            request_obj_id = ObjectId(request_id)
+            return self.file_checks_collection.find_one({"_id": request_obj_id})
+        except Exception as e:
+            return None
+
+    def update_file_duplicate_status(self, file_id: str, is_duplicated: bool):
+        """파일의 중복 상태를 업데이트합니다."""
+        try:
+            file_obj_id = ObjectId(file_id)
+            result = self.files_collection.update_one(
+                {"_id": file_obj_id},
+                {"$set": {"is_duplicated": is_duplicated}}
+            )
+
+            if result.modified_count > 0:
+                return self.files_collection.find_one({"_id": file_obj_id})
+            return None
+        except Exception as e:
+            return None
+
+    def update_duplicate_check_result(self, request_id: str, is_duplicated: bool):
+        """중복 검사 결과를 업데이트합니다."""
+        try:
+            now = self.get_current_time()
+            request_obj_id = ObjectId(request_id)
+
+            result = self.file_checks_collection.update_one(
+                {"_id": request_obj_id},
+                {
+                    "$set": {
+                        "is_completed": True,
+                        "is_duplicated": is_duplicated,
+                        "updated_at": now
+                    }
+                }
+            )
+
+            if result.modified_count > 0:
+                return self.file_checks_collection.find_one({"_id": request_obj_id})
+            return None
+        except Exception as e:
+            return None 

src/main/ai/data/__init__.py

@@ -0,0 +1,4 @@
+from src.main.ai.data.CategoryRecommendationRepository import CategoryRecommendationRepository
+from src.main.ai.data.CategoryRecommendationQueue import CategoryRecommendationQueue
+from src.main.ai.data.FileDuplicateCheckRepository import FileDuplicateCheckRepository
+from src.main.ai.data.FileDuplicateCheckQueue import FileDuplicateCheckQueue

src/main/ai/di/dependencies.py

@@ -5,6 +5,9 @@
 from src.main.ai.data.CategoryRecommendationRepository import CategoryRecommendationRepository
 from src.main.ai.service.CategoryRecommendationService import CategoryRecommendationService
 from src.main.ai.data.CategoryRecommendationQueue import CategoryRecommendationQueue
+from src.main.ai.data.FileDuplicateCheckRepository import FileDuplicateCheckRepository
+from src.main.ai.service.FileDuplicateCheckService import FileDuplicateCheckService
+from src.main.ai.data.FileDuplicateCheckQueue import FileDuplicateCheckQueue
 from src.main.config.mongodb import get_mongo_client
 
 load_dotenv()
@@ -31,3 +34,26 @@ def get_category_recommendation_service():
     repository = get_category_recommendation_repository()
     queue = get_category_recommendation_queue()
     return CategoryRecommendationService(repository, queue)
+
+
+def get_file_duplicate_check_repository():
+    client = get_mongo_client()
+    return FileDuplicateCheckRepository(client)
+
+
+def get_file_duplicate_check_queue():
+    if os.getenv('ENV') == 'local':
+        aws_profile = os.getenv('AWS_PROFILE', 'default')
+        session = boto3.Session(profile_name=aws_profile)
+        sqs_client = session.client('sqs')
+    else:
+        sqs_client = boto3.client('sqs', region_name=os.getenv('AWS_REGION'))
+
+    queue_url = os.getenv('SQS_REQUEST_QUEUE_URL')
+    return FileDuplicateCheckQueue(sqs_client, queue_url)
+
+
+def get_file_duplicate_check_service():
+    repository = get_file_duplicate_check_repository()
+    queue = get_file_duplicate_check_queue()
+    return FileDuplicateCheckService(repository, queue)

src/main/ai/exception/BadRequestException.py

@@ -0,0 +1,13 @@
+from fastapi import HTTPException
+
+
+class BadRequestException(HTTPException):
+    """
+    요청이 잘못되었을 때 발생하는 예외
+    """
+
+    def __init__(self, detail: str = "Bad request"):
+        super().__init__(
+            status_code=400, 
+            detail=detail
+        ) 

src/main/ai/exception/FileNotFoundException.py

@@ -0,0 +1,13 @@
+from fastapi import HTTPException
+
+
+class FileNotFoundException(HTTPException):
+    """
+    파일을 찾을 수 없을 때 발생하는 예외
+    """
+
+    def __init__(self, file_id: str):
+        super().__init__(
+            status_code=404, 
+            detail=f"File with ID '{file_id}' not found"
+        ) 

src/main/ai/exception/__init__.py

@@ -0,0 +1,3 @@
+"""
+예외 패키지
+""" 

src/main/ai/models/CategoryRecommendation.py

@@ -18,4 +18,5 @@ class CategoryRecommendationStatusResponse(BaseModel):
 
 
 class CategoryRecommendationResultRequest(BaseModel):
+    request_id: Optional[str] = None
     predicted_category: str 

src/main/ai/models/FileDuplicateCheck.py

@@ -0,0 +1,34 @@
+from pydantic import BaseModel, Field
+from typing import Optional, List
+from datetime import datetime
+
+
+class FileDuplicateCheckRequest(BaseModel):
+    """파일 중복 검사 요청 모델"""
+    user_id: str
+    file_id: str
+
+
+class FileDuplicateCheckResponse(BaseModel):
+    """파일 중복 검사 응답 모델"""
+    request_id: str
+
+
+class FileDuplicateCheckStatusResponse(BaseModel):
+    """파일 중복 검사 상태 응답 모델"""
+    request_id: str
+    file_id: str
+    is_completed: bool
+    is_duplicated: Optional[bool] = None
+
+
+class FileDuplicateCheckResultRequest(BaseModel):
+    """파일 중복 검사 결과 요청 모델"""
+    request_id: str
+    is_duplicated: bool
+
+
+class FileDuplicateCheckEmbeddingsRequest(BaseModel):
+    """파일 임베딩 저장 요청 모델"""
+    file_id: str
+    embeddings: List[float] 

src/main/ai/models/__init__.py

@@ -1 +1,2 @@
-
+from src.main.ai.models.CategoryRecommendation import *
+from src.main.ai.models.FileDuplicateCheck import * 

src/main/ai/router/AIInternalAPIRouter.py

@@ -1,9 +1,11 @@
 from fastapi import APIRouter, Depends, HTTPException, status
 from fastapi.responses import JSONResponse
 
-from src.main.ai.di.dependencies import get_category_recommendation_service
+from src.main.ai.di.dependencies import get_category_recommendation_service, get_file_duplicate_check_service
 from src.main.ai.models.CategoryRecommendation import CategoryRecommendationResultRequest, CategoryRecommendationStatusResponse
 from src.main.ai.service.CategoryRecommendationService import CategoryRecommendationService
+from src.main.ai.models.FileDuplicateCheck import FileDuplicateCheckStatusResponse, FileDuplicateCheckEmbeddingsRequest, FileDuplicateCheckRequest, FileDuplicateCheckResponse, FileDuplicateCheckResultRequest
+from src.main.ai.service.FileDuplicateCheckService import FileDuplicateCheckService
 
 
 router = APIRouter(
@@ -12,13 +14,39 @@
 )
 
 
-@router.post("/category-recommendation-results/{request_id}", response_model=CategoryRecommendationStatusResponse)
+@router.post("/category-recommendation-results", response_model=dict)
 async def update_category_recommendation_result(
-    request_id: str,
     request: CategoryRecommendationResultRequest,
     service: CategoryRecommendationService = Depends(get_category_recommendation_service)
 ):
-    result = service.update_recommendation_result(request_id, request)
+    result = service.update_recommendation_result(request.request_id, request)
+
+    if not result:
+        return JSONResponse(
+            status_code=status.HTTP_404_NOT_FOUND,
+            content={
+                "detail": "요청을 찾을 수 없습니다. 존재하지 않는 ID입니다."
+            }
+        )
+
+    return {"success": True}
+
+
+@router.post("/file-duplicate-checks", response_model=FileDuplicateCheckResponse)
+async def create_file_duplicate_check(
+    request: FileDuplicateCheckRequest,
+    service: FileDuplicateCheckService = Depends(get_file_duplicate_check_service)
+):
+    result = service.create_duplicate_check_request(request)
+    return result
+
+
+@router.post("/file-duplicate-check-embeddings", response_model=dict)
+async def update_file_duplicate_check_result(
+    request: FileDuplicateCheckResultRequest,
+    service: FileDuplicateCheckService = Depends(get_file_duplicate_check_service)
+):
+    result = service.update_duplicate_check_result(request.request_id, request.is_duplicated)
 
     if not result:
         return JSONResponse(
@@ -29,5 +57,13 @@ async def update_category_recommendation_result(
                 "detail": "존재하지 않는 ID입니다."
             }
         )
-
-    return result
+
+    # 요청 정보 조회
+    check = service.repository.get_duplicate_check_by_id(request.request_id)
+
+    return {
+        "request_id": request.request_id,
+        "file_id": check["file_id"],
+        "is_completed": check["is_completed"],
+        "is_duplicated": check["is_duplicated"]
+    }

src/main/ai/router/AIPublicAPIRouter.py

@@ -2,9 +2,11 @@
 import uuid
 
 from src.main.auth.dependencies import get_current_user
-from src.main.ai.di.dependencies import get_category_recommendation_service
+from src.main.ai.di.dependencies import get_category_recommendation_service, get_file_duplicate_check_service
 from src.main.ai.models.CategoryRecommendation import CategoryRecommendationRequest, CategoryRecommendationResponse, CategoryRecommendationStatusResponse
 from src.main.ai.service.CategoryRecommendationService import CategoryRecommendationService
+from src.main.ai.models.FileDuplicateCheck import FileDuplicateCheckStatusResponse
+from src.main.ai.service.FileDuplicateCheckService import FileDuplicateCheckService
 
 
 router = APIRouter(
@@ -43,3 +45,23 @@ async def get_category_recommendation_status(
         )
 
     return result
+
+
+@router.get("/file-duplicate-checks", response_model=FileDuplicateCheckStatusResponse)
+async def get_file_duplicate_check_status(
+    file_id: str,
+    user_id: uuid.UUID = Depends(get_current_user),
+    service: FileDuplicateCheckService = Depends(get_file_duplicate_check_service)
+):
+    """
+    파일 중복 검사 결과 조회
+    """
+    result = service.get_duplicate_check_status(file_id, str(user_id))
+
+    if not result:
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND,
+            detail="요청을 찾을 수 없습니다. 존재하지 않는 ID입니다."
+        )
+
+    return result

src/main/ai/router/__init__.py

@@ -0,0 +1,2 @@
+from src.main.ai.router.AIPublicAPIRouter import router as public_router
+from src.main.ai.router.AIInternalAPIRouter import router as internal_router