app.py

from dotenv import load_dotenv
import os
from langchain_openai import ChatOpenAI, OpenAIEmbeddings
from langchain_astradb import AstraDBVectorStore
from langchain.agents import initialize_agent, AgentType
from langchain.tools.retriever import create_retriever_tool
from langchain.agents import AgentExecutor
from github import fetch_github_issues
import gradio as gr
from langchain import hub
from note import note_tool

load_dotenv()

def connet_to_vstore():
    embeddings = OpenAIEmbeddings()
    desire_namespace = os.getenv("ASTRA_DB_KEYSPACE")
    ASTRA_DB_APPLICATION_TOKEN = os.getenv("ASTRA_DB_APPLICATION_TOKEN")
    ASTRA_DB_API_ENDPOINT = os.getenv("ASTRA_DB_API_ENDPOINT")  
    if desire_namespace:
        ASTRA_DB_KEYSPACE = desire_namespace
    else:
        ASTRA_DB_KEYSPACE = None
    
    vstore = AstraDBVectorStore(
        embedding=embeddings,
        collection_name="github",
        namespace=ASTRA_DB_KEYSPACE,
        api_endpoint=ASTRA_DB_API_ENDPOINT,
        token=ASTRA_DB_APPLICATION_TOKEN,
    )
    return vstore

vstore = connet_to_vstore()
add_to_vectorstore = input("Do you want to update the issue? (yes/no): ").lower() in ["yes", "y"]
if add_to_vectorstore:
    owner = "Ini-design"
    repo = "register"
    issues = fetch_github_issues(owner, repo)
    try:
        vstore.delete_collection()
    except:
        pass
        
    vstore = connet_to_vstore()
    vstore.add_documents(issues)
    
   # results = vstore.similarity_search('flash message', k=3)
   # for res in results:
   #     print(f"*{res.page_content} {res.metadata}")
   
retriever = vstore.as_retriever(search_kwargs={"k":3})
retriever_tool = create_retriever_tool(
    retriever,
    "github_search",
    "search for information aabout github issues. For any question abour github issue, you must used this tools!"
)

prompt = hub.pull("hwchase17/openai-functions_agent")
llm = ChatOpenAI()

tools = [retriever_tool], note_tool
agent = initialize_agent(llm, tools, prompt)
agent_executor = AgentExecutor(agent=AgentType.OPENAI_FUNCTIONS, tools=tools, verbose=True)

def answer_question(question):
    """Function to process user question and return agent response"""
    if not question.strip():
        return "Please enter a question."
    
    try:
        result = agent_executor.invoke({"input": question})
        return result["output"]
    except Exception as e:
        return f"Error: {str(e)}"

# Create Gradio Interface
demo = gr.Interface(
    fn=answer_question,
    inputs=gr.Textbox(
        label="Ask about GitHub Issues",
        placeholder="Type your question here...",
        lines=3
    ),
    outputs=gr.Textbox(
        label="Response",
        lines=5
    ),
    title="GitHub Issues AI Agent",
    description="Ask questions about GitHub issues using AI-powered semantic search",
    examples=[
        ["What are the recent issues?"],
        ["Tell me about bug reports"],
        ["What features are being requested?"]
    ],
    allow_flagging="never"
)

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