Delete backend.py

backend.py

@@ -1,214 +0,0 @@
-import os
-import tempfile
-import gc
-import logging
-from fastapi import FastAPI, UploadFile, File, HTTPException
-from pydantic import BaseModel
-import torch
-from dotenv import load_dotenv
-
-# ---------------- CORS IMPORT ----------------
-from fastapi.middleware.cors import CORSMiddleware  # ADD THIS
-# ---------------------------------------------
-
-# ---------------- Groq API ----------------
-from groq import Groq, APIError
-
-# ---------------- LangChain ----------------
-from langchain_community.document_loaders import PyPDFLoader
-from langchain_text_splitters import RecursiveCharacterTextSplitter
-from langchain_community.embeddings import HuggingFaceEmbeddings
-from langchain_community.vectorstores import Chroma
-
-# ---------------- CONFIGURATION ----------------
-logging.basicConfig(level=logging.INFO)
-load_dotenv()
-
-# Initialize FastAPI with CORS
-app = FastAPI()
-
-# ============== CORS MIDDLEWARE - MUST ADD THIS ==============
-app.add_middleware(
-    CORSMiddleware,
-    allow_origins=["*"],  # Allows all origins
-    allow_credentials=True,
-    allow_methods=["*"],  # Allows all methods
-    allow_headers=["*"],  # Allows all headers
-)
-# ==============================================================
-
-# API Configuration
-GROQ_API_KEY = os.environ.get("GROQ_API_KEY", st.secrets.get("GROQ_API_KEY") if 'st' in locals() else None)
-GROQ_MODEL = "llama-3.1-8b-instant"
-
-# Initialize Groq Client
-client = None
-if GROQ_API_KEY:
-    try:
-        client = Groq(api_key=GROQ_API_KEY)
-        logging.info("✅ Groq client initialized successfully.")
-    except Exception as e:
-        logging.error(f"❌ Failed to initialize Groq client: {e}")
-        client = None
-else:
-    logging.warning("⚠️ GROQ_API_KEY not found in environment.")
-
-# Global state
-retriever = None
-vectorstore = None
-
-# ---------------- Input Schema ----------------
-class Query(BaseModel):
-    question: str
-
-# ==================================================
-#  ROOT ENDPOINT (For health check)
-# ==================================================
-@app.get("/")
-async def root():
-    return {"message": "FastAPI Backend is running", "status": "healthy"}
-
-# ==================================================
-#  PDF Upload → Chunk → Embed → Vectorstore
-# ==================================================
-@app.post("/upload")
-async def upload_pdf(file: UploadFile = File(...)):
-    """Handles PDF upload and processing."""
-    global retriever, vectorstore
-
-    if not file.filename.endswith(".pdf"):
-        raise HTTPException(400, "Only PDF files allowed")
-
-    if not client:
-        raise HTTPException(500, "Groq API key is missing or invalid.")
-
-    path = None
-    try:
-        # 1. Save file temporarily
-        with tempfile.NamedTemporaryFile(delete=False, suffix=".pdf") as tmp:
-            tmp.write(await file.read())
-            path = tmp.name
-
-        logging.info(f"Processing PDF: {path}")
-
-        # 2. Load PDF
-        loader = PyPDFLoader(path)
-        docs = loader.load()
-
-        # 3. Split into chunks
-        splitter = RecursiveCharacterTextSplitter(
-            chunk_size=800,
-            chunk_overlap=50
-        )
-        chunks = splitter.split_documents(docs)
-
-        # 4. Create embeddings
-        embeddings = HuggingFaceEmbeddings(
-            model_name="sentence-transformers/all-MiniLM-L6-v2",
-            model_kwargs={"device": "cpu"},
-            encode_kwargs={"normalize_embeddings": True}
-        )
-
-        # 5. Clear previous vectorstore
-        if vectorstore:
-            del vectorstore
-            gc.collect()
-
-        # 6. Create Vectorstore and Retriever
-        vectorstore = Chroma.from_documents(chunks, embeddings)
-        retriever = vectorstore.as_retriever(search_kwargs={"k": 3})
-
-        logging.info(f"PDF processed. Chunks created: {len(chunks)}")
-
-        return {"message": "PDF processed", "chunks": len(chunks)}
-
-    except Exception as e:
-        logging.error(f"Error during PDF processing: {e}")
-        raise HTTPException(500, f"Error: {str(e)}")
-    finally:
-        # 7. Cleanup
-        if path and os.path.exists(path):
-            os.unlink(path)
-        gc.collect()
-
-# ==================================================
-#  ASK → RETRIEVE → GROQ → ANSWER
-# ==================================================
-@app.post("/ask")
-async def ask(req: Query):
-    global retriever
-
-    if client is None:
-        raise HTTPException(500, "Groq client is not initialized. Check API key setup.")
-
-    if retriever is None:
-        raise HTTPException(400, "Upload PDF first to initialize the knowledge base.")
-
-    try:
-        # 1. Retrieve relevant chunks
-        docs = retriever.invoke(req.question)
-        context = "\n\n".join(d.page_content for d in docs)
-
-        # 2. Build prompt
-        prompt = f"""
-You are a strict RAG Q&A assistant.
-Use ONLY the context provided. If the answer is not found, reply:
-"I cannot find this in the PDF."
-
----------------- CONTEXT ----------------
-{context}
------------------------------------------
-
-QUESTION: {req.question}
-
-FINAL ANSWER:
-"""
-
-        # 3. Call Groq
-        response = client.chat.completions.create(
-            model=GROQ_MODEL,
-            messages=[
-                {"role": "system",
-                 "content": "Use only the PDF content. If answer not found, say: 'I cannot find this in the PDF.'"},
-                {"role": "user", "content": prompt}
-            ],
-            temperature=0.0
-        )
-
-        answer = response.choices[0].message.content.strip()
-        return {"answer": answer, "sources": len(docs)}
-
-    except APIError as e:
-        logging.error(f"Groq API Error: {e}")
-        raise HTTPException(500, f"Groq API Error: {str(e)}")
-    except Exception as e:
-        logging.error(f"General error in /ask: {e}")
-        raise HTTPException(500, f"General error: {str(e)}")
-
-# ==================================================
-#  HEALTH & CLEAR
-# ==================================================
-@app.get("/health")
-async def health():
-    """Endpoint for checking service status."""
-    return {
-        "status": "running",
-        "pdf_loaded": retriever is not None,
-        "groq_client_ok": client is not None
-    }
-
-@app.post("/clear")
-async def clear():
-    """Clears the current RAG components from memory."""
-    global retriever, vectorstore
-
-    if vectorstore:
-        del vectorstore
-    retriever = None
-    vectorstore = None
-
-    gc.collect()
-    if torch.cuda.is_available():
-        torch.cuda.empty_cache()
-
-    return {"message": "Memory cleared. Upload a new PDF."}