app.py

import gradio as gr
import os
from langchain.chains import ConversationalRetrievalChain
from langchain.memory import ConversationBufferMemory
from langchain_openai import ChatOpenAI, OpenAIEmbeddings
from langchain.prompts import PromptTemplate
from langchain_community.vectorstores import Chroma

def create_qa_chain():
    """
    Create the QA chain with the loaded vectorstore
    """
    # Initialize embeddings and load vectorstore
    embeddings = OpenAIEmbeddings()
    vectorstore = Chroma(
        persist_directory="./vectorstore",
        embedding_function=embeddings
    )
    
    # Set up retriever
    retriever = vectorstore.as_retriever(
        search_type="mmr",
        search_kwargs={
            "k": 6,
            "fetch_k": 20,
            "lambda_mult": 0.3,
        }
    )
    
    # Set up memory
    memory = ConversationBufferMemory(
        memory_key="chat_history",
        return_messages=True,
        output_key='answer'
    )
    
    # Create prompt template
    qa_prompt = PromptTemplate.from_template("""You are an expert technical writer specializing in API documentation. 
When describing API endpoints, structure your response in this exact format:

1. Start with the HTTP method and base URI structure
2. List all key parameters with:
   - Parameter name in bold (**parameter**)
   - Type and requirement status
   - Clear description
   - Example values where applicable
3. Show complete example requests with:
   - Basic example
   - Full example with all parameters
   - Headers included
4. Include any relevant response information
   
Use markdown formatting for:
- Code blocks with syntax highlighting
- Bold text for important terms
- Clear section separation

Context: {context}

Question: {question}

Technical answer (following the exact structure above):""")
    
    # Create the chain
    qa_chain = ConversationalRetrievalChain.from_llm(
        llm=ChatOpenAI(
            temperature=0.1,
            model_name="gpt-4-turbo-preview"
        ),
        retriever=retriever,
        memory=memory,
        return_source_documents=True,
        combine_docs_chain_kwargs={"prompt": qa_prompt},
        verbose=False
    )
    
    return qa_chain

def chat(message, history):
    """
    Process chat messages and return responses
    """
    # Get or create QA chain
    if not hasattr(chat, 'qa_chain'):
        chat.qa_chain = create_qa_chain()
    
    # Get response
    result = chat.qa_chain({"question": message})
    
    # Format sources
    sources = "\n\nSources:\n"
    seen_components = set()
    shown_sources = 0
    
    for doc in result["source_documents"]:
        component = doc.metadata.get('component', '')
        title = doc.metadata.get('title', '')
        combo = (component, title)
        
        if combo not in seen_components and shown_sources < 3:
            seen_components.add(combo)
            shown_sources += 1
            sources += f"\nSource {shown_sources}:\n"
            sources += f"Title: {title}\n"
            sources += f"Component: {component}\n"
            sources += f"Content: {doc.page_content[:300]}...\n"
    
    # Combine response with sources
    full_response = result["answer"] + sources
    
    return full_response

demo = gr.ChatInterface(
    chat,
    title="Apple Music API Documentation Assistant",
    description="Ask questions about the Apple Music API documentation.",
    examples=[
        "How to search for songs on Apple Music API?",
        "What are the required parameters for searching songs?",
        "Show me an example request with all parameters"
    ]
)

if __name__ == "__main__":
    demo.launch()