app.py

import os
from dotenv import load_dotenv
from fastapi import FastAPI, HTTPException, Request
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import JSONResponse
from pydantic import BaseModel
from typing import List, Optional

# Import RAG components
from langchain.embeddings import OpenAIEmbeddings
from langchain.vectorstores import Chroma
from langchain.text_splitter import CharacterTextSplitter
from langchain.chains import RetrievalQA
from langchain.llms import OpenAI
from langchain.document_loaders import TextLoader

# Load environment variables
load_dotenv()

# Initialize FastAPI app
app = FastAPI(title="Educational Research Methods Chatbot API")

# Add CORS middleware
app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],  # In production, replace with specific origins
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)

# Define request and response models
class ChatRequest(BaseModel):
    message: str
    conversation_history: Optional[List[dict]] = []

class ChatResponse(BaseModel):
    response: str
    citations: List[dict] = []

# Initialize RAG components
def initialize_rag():
    # Load research methods information
    loader = TextLoader("/home/ubuntu/research_methods_chatbot/research_methods_info.md")
    documents = loader.load()
    
    # Split documents into chunks
    text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
    texts = text_splitter.split_documents(documents)
    
    # Create embeddings
    embeddings = OpenAIEmbeddings()
    
    # Create vector store
    db = Chroma.from_documents(texts, embeddings)
    
    # Create retriever
    retriever = db.as_retriever(search_type="similarity", search_kwargs={"k": 3})
    
    # Create QA chain
    qa = RetrievalQA.from_chain_type(
        llm=OpenAI(),
        chain_type="stuff",
        retriever=retriever,
        return_source_documents=True,
        verbose=True,
    )
    
    return qa

# Initialize RAG pipeline
qa_chain = initialize_rag()

@app.post("/chat", response_model=ChatResponse)
async def chat(request: ChatRequest):
    try:
        # Prepare the query with conversation history context
        query = request.message
        if request.conversation_history:
            context = "\n".join([f"User: {msg['message']}\nAssistant: {msg['response']}" 
                               for msg in request.conversation_history[-3:]])
            query = f"Conversation history:\n{context}\n\nCurrent question: {query}"
        
        # Add instruction for APA7 citations
        query += "\nPlease include APA7 citations for any information provided."
        
        # Get response from RAG pipeline
        result = qa_chain({"query": query})
        
        # Extract citations from source documents
        citations = []
        if "source_documents" in result:
            for i, doc in enumerate(result["source_documents"]):
                if hasattr(doc, "metadata") and "source" in doc.metadata:
                    citations.append({
                        "id": i + 1,
                        "text": doc.metadata["source"],
                        "page": doc.metadata.get("page", "")
                    })
        
        return ChatResponse(
            response=result["result"],
            citations=citations
        )
    
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))

@app.get("/health")
async def health_check():
    return {"status": "healthy"}

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