app.py

## API Call
# main.py
## FASTAPI Main py file to access the POST body

from fastapi import FastAPI, Header, HTTPException
from pydantic import BaseModel
from typing import List, Optional
from utils.s3_utils import read_s3_file
from utils.embedding_utils import read_document, cumulative_semantic_chunking, embed_chunks
from utils.qdrant_utils import store_embeddings
import logging
import time
import os
from dotenv import load_dotenv

#load_dotenv()

# Retrieve the API key from the environment
API_KEY = os.getenv('X_API_KEY')
bucket_name = os.getenv('bucket_name')

# Initialize logging
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')

app = FastAPI()

class Metadata(BaseModel):
    mime_type: str
    file_size_bytes: str
    file_format: str

class DocumentIndexRequest(BaseModel):
    metadata: Metadata
    bucket_key: str
    user_id: str
    org_id: str
    file_id: int
    data_source_id: int
    
@app.get("/")
async def root():
    return {"message": "Welcome to the Document Indexing API!"}

@app.post("/api/document-index")
async def document_index(request: DocumentIndexRequest, x_api_key: str = Header(...)):
    logging.info(f"Received request: {request}")
    start_time = time.time()
    # Check if the API key provided in the header matches the one in the environment
    if x_api_key != API_KEY:
        logging.warning("Unauthorized access attempt with invalid API key.")
        raise HTTPException(status_code=401, detail="Unauthorized")
    
    try:
        #bucket_name = "document-ingestion-drive-dev"
        
        # Read file from S3 using the presigned URL
        content, metadata, file_format = read_s3_file(bucket_name, request.bucket_key)
        logging.info(f"File {request.bucket_key} retrieved from S3 with format {file_format}.")


        # Reading content using Simple Directory Reader
        text_content = read_document(content, file_id=request.file_id, file_format=file_format)
        logging.info(f"Text content extracted from file {request.bucket_key}.")
        #print('text_content',text_content)

        # Chunking text using semantic chunking
        chunks = cumulative_semantic_chunking(text_content, max_chunk_size=512, similarity_threshold=0.8)
        logging.info(f"Text content chunked into {len(chunks)} chunks.")
        # Embed chunks
        embeddings, total_tokens = embed_chunks(chunks)
        logging.info(f"Text content embedded into vectors. Total tokens used: {total_tokens}.")

        # Store embeddings in Qdrant
        store_embeddings(
            chunks=chunks,
            embeddings=embeddings,
            user_id=request.user_id,
            data_source_id=request.data_source_id,
            file_id=request.file_id,
            organization_id=request.org_id,
            s3_bucket_key=request.bucket_key,
            total_tokens=total_tokens     
        )
        
        logging.info(f"Embeddings for {request.bucket_key} stored successfully with metadata: {metadata}")
        logging.info(f"Embeddings for stored successfully with {total_tokens} tokens")
        time_taken = time.time() - start_time
        logging.info(f"Time taken to process and embed the document: {time_taken} seconds")


        return {"message": "Embeddings stored successfully"}
        
    
    except FileNotFoundError as e:
        logging.error(f"File not found: {str(e)}")
        raise HTTPException(status_code=404, detail=str(e))
    except PermissionError as e:
        logging.error(f"Permission error: {str(e)}")
        raise HTTPException(status_code=403, detail=str(e))
    except HTTPException as e:
        logging.error(f"HTTP error: {str(e.detail)}")
        raise
    except Exception as e:
        logging.error(f"Error processing file {request.bucket_key}: {str(e)}")
        raise HTTPException(status_code=500, detail=str(e))

# Run the FastAPI app
if __name__ == "__main__":
    import uvicorn
    uvicorn.run(app, host="0.0.0.0", port=8000)