new file: .gitignore

modified:   digitalseapi/app/infra/config.py
modified:   digitalseapi/app/main.py
modified:   digitalseapi/app/routers/__init__.py
modified:   digitalseapi/app/routers/documents.py
modified:   digitalseapi/requirements.txt
modified:   digitalseapi/tests/conftest.py
modified:   docker-compose.yml
new file:   ocrservice/.dockerignore
modified:   ocrservice/app.py
modified:   ocrservice/requirements.txt

.gitignore

@@ -0,0 +1,2 @@
+__pycache__
+venv

digitalseapi/app/infra/config.py

@@ -2,7 +2,7 @@
 from functools import lru_cache
 from pydantic_settings import BaseSettings  # Alteração aqui
 from minio import Minio
-
+from fastapi import Depends
 log = logging.getLogger("uvicorn")
 
 
@@ -16,9 +16,9 @@ class Settings(BaseSettings):
     MINIO_SECRET_KEY: str = "digitalse"
     MINIO_SECURE: bool = False
     SECRET_KEY: str = "digitalse"
-    ALGORITHM = "HS256"
-    ACCESS_TOKEN_EXPIRE_MINUTES = 30
-    REFRESH_TOKEN_EXPIRE_MINUTES = 600
+    #ALGORITHM: str = "HS256"
+    #ACCESS_TOKEN_EXPIRE_MINUTES: str = 30
+    #REFRESH_TOKEN_EXPIRE_MINUTES: str = 600
 
     class Config:
         env_file = ".env"

digitalseapi/app/main.py

@@ -3,18 +3,18 @@
 
 from app.infra.db import init_db
 
-from app.infra.config import get_settings
+from app.infra.config import settings
 from app.routers import main_router
 
 
 app = FastAPI(
     title="DigitalSE",
-    version="0.1.0",
+    version="0.1.15-rc0.2",
     description="Sistema de gestão documental com extração de dados por OCR",
 )
 allow_origin = ["*"]
 
-settings = get_settings()
+#settings = get_settings()
 
 
 app.add_middleware(

digitalseapi/app/routers/__init__.py

@@ -2,7 +2,7 @@
 
 from .documents import router as documents_router
 from .user import router as user_router
-from .auth import router as auth_router
+#from .auth import router as auth_router
 
 main_router = APIRouter()
 
@@ -14,6 +14,4 @@
                            prefix="/users",
                            tags=["Users"])
 
-main_router.include_router(auth_router,
-                           prefix="/auth",
-                           tags=["Auth"])
+#main_router.include_router(auth_router, prefix="/auth", tags=["Auth"])

digitalseapi/app/routers/documents.py

@@ -8,11 +8,44 @@
 from app.infra.db import get_session
 from app.infra.config import get_minio_client
 import json
+from motor.motor_asyncio import AsyncIOMotorClient
+from pydantic import BaseModel
+from typing import List
+
+class KeywordList(BaseModel):
+    keywords: List[str]
+
 
 router = APIRouter()
 
+#Banco mongo, refatorar posteriormente, inserindo na infra de banco de dados
+#mongo_client = AsyncIOMotorClient("mongodb://mongo:27017/")
+
+async def get_mongodb_client() -> AsyncIOMotorClient:
+    return AsyncIOMotorClient("mongodb://mongo:27017/")
+#db = mongo_client["document_db"]
+#collection = db["documents"]
+
+
+@router.post("/search")
+async def search_documents(keywords: KeywordList,
+                           mongodb_client: AsyncIOMotorClient = Depends(get_mongodb_client)):
+    try:
+        # Conecte-se à coleção no MongoDB
+        collection = mongodb_client["document_db"]["documents"]
+
+        # Consulta os documentos que contenham todas as palavras-chave fornecidas
+        cursor = collection.find({"_id": {"$in": keywords.keywords}})
+
+        # Transforma o cursor em uma lista de documentos
+        documents = await cursor.to_list(length=None)
+
+        # Retorna os documentos encontrados
+        return {"documents": documents}
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Error searching documents: {e}")
 
-@router.post("/upload/")
+@router.post("/")
 async def upload_image(*, input_images: List[UploadFile] = File(...),
                        title: Optional[str],
                        description: Optional[str],
@@ -41,7 +74,7 @@ async def upload_image(*, input_images: List[UploadFile] = File(...),
 
 
 # Endpoint para listar as imagens
-@router.get("/images/", response_model=List[GestaoArquivos])
+@router.get("/", response_model=List[GestaoArquivos])
 async def list_images(session: Session = Depends(get_session)):
     try:
         # Consulta o PostgreSQL para obter os caminhos das imagens

digitalseapi/requirements.txt

@@ -23,3 +23,5 @@ aio-pika==9.3.1
 setuptools==69.1.1
 #elastic-apm[fastapi]
 jose
+passlib
+motor==3.3.2

digitalseapi/tests/conftest.py

@@ -0,0 +1,24 @@
+
+import os
+
+import pytest
+from starlette.testclient import TestClient
+
+from app import main
+from app.config import get_settings, Settings
+
+
+def get_settings_override():
+    return Settings(testing=1, database_url=os.environ.get("DATABASE_TEST_URL"))
+
+
+@pytest.fixture(scope="module")
+def test_app():
+    # set up
+    main.app.dependency_overrides[get_settings] = get_settings_override
+    with TestClient(main.app) as test_client:
+
+        # testing
+        yield test_client
+
+    # tear down

docker-compose.yml

@@ -65,10 +65,21 @@ services:
       - rabbitmq
     networks:
       - digitalse-network
+  mongo:
+    image: mongo
+    ports:
+      - "27017:27017"
+    volumes:
+      - mongo_data:/data/db
+    networks:
+      - digitalse-network  
+
 networks:
   digitalse-network:
+
 volumes:
   data_postgres:
   minio_data:
   local-images:
   rabbitmq_data:
+  mongo_data:

ocrservice/.dockerignore

@@ -0,0 +1,3 @@
+venv
+.dockerignore
+Dockerfile

ocrservice/app.py

@@ -7,6 +7,8 @@
 import json
 from pydantic import BaseModel
 import re
+import asyncio
+from motor.motor_asyncio import AsyncIOMotorClient
 
 
 broker = RabbitBroker("amqp://guest:guest@rabbitmq:5672")
@@ -19,6 +21,11 @@
                       secret_key="digitalse",
                       secure=False)
 
+# Configuração do MongoDB
+mongo_client = AsyncIOMotorClient("mongodb://mongo:27017/")
+db = mongo_client["document_db"]
+collection = db["documents"]
+
 
 class UploadedFile(BaseModel):
     user_id: int
@@ -33,6 +40,24 @@ def clean_and_tokenize(text):
     tokens = cleaned_text.split()
     return tokens
 
+
+async def index_document(document_id, tokens):
+    for token in tokens:
+        await collection.update_one(
+            {"_id": token},
+            {"$addToSet": {"documents": document_id}},
+            upsert=True
+        )
+
+async def search_documents(keyword):
+    result = await collection.find_one({"_id": keyword})
+    if result:
+        return result.get("documents", [])
+    else:
+        return []
+
+
+
 @broker.subscriber("ocr")
 async def handle(message):
     print(message)
@@ -52,5 +77,6 @@ async def handle(message):
 
             # Imprimir o texto extraído
         print(tokens)
+        await index_document(uploaded_file.document_id, tokens)
     except Exception as e:
         print(f"Erro ao transformar a mensagem em uma instância Pydantic: {e}")

ocrservice/requirements.txt

@@ -2,4 +2,5 @@ faststream[rabbit]
 fast_depends
 minio 
 pytesseract
-pydantic
+pydantic
+motor==3.3.2