app.py

import os
import gc
import tempfile
import uuid
import logging

import streamlit as st
from dotenv import load_dotenv

from gitingest import ingest
from llama_index.core import Settings, PromptTemplate, VectorStoreIndex, SimpleDirectoryReader
from llama_index.core.node_parser import MarkdownNodeParser
from llama_index.llms.sambanovasystems import SambaNovaCloud
from llama_index.embeddings.huggingface import HuggingFaceEmbedding

# Load environment variables from .env
load_dotenv()

# Configure logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

# Custom exception for application errors
class GitHubRAGError(Exception):
    """Custom exception for GitHub RAG application errors"""
    pass

# Fetch API key for SambaNova
SAMBANOVA_API_KEY = os.getenv("SAMBANOVA_API_KEY")
if not SAMBANOVA_API_KEY:
    raise ValueError("SAMBANOVA_API_KEY is not set in environment variables")

# Initialize Streamlit session state
if "id" not in st.session_state:
    st.session_state.id = uuid.uuid4()
    st.session_state.file_cache = {}
    st.session_state.messages = []

session_id = st.session_state.id

@st.cache_resource
def load_llm():
    """
    Load and cache the SambaNova LLM predictor
    """
    return SambaNovaCloud(
        api_key=SAMBANOVA_API_KEY,
        model="DeepSeek-R1-Distill-Llama-70B",
        temperature=0.1,
        top_p=0.1,
    )


def reset_chat():
    """Clear chat history and free resources"""
    st.session_state.messages = []
    gc.collect()


def process_with_gitingets(github_url: str):
    """Use gitingest to fetch and summarize the GitHub repository"""
    summary, tree, content = ingest(github_url)
    return summary, tree, content

# --- Sidebar: Load Repository ---
with st.sidebar:
    st.header("Add your GitHub repository!")
    github_url = st.text_input(
        "GitHub repo URL", placeholder="https://github.com/user/repo"
    )
    load_btn = st.button("Load Repository")

    if github_url and load_btn:
        try:
            repo_name = github_url.rstrip("/").split("/")[-1]
            cache_key = f"{session_id}-{repo_name}"

            # Only process if not cached
            if cache_key not in st.session_state.file_cache:
                with st.spinner("Processing repository..."):
                    summary, tree, content = process_with_gitingets(github_url)

                    with tempfile.TemporaryDirectory() as tmpdir:
                        md_path = os.path.join(tmpdir, f"{repo_name}.md")
                        with open(md_path, "w", encoding="utf-8") as f:
                            f.write(content)

                        loader = SimpleDirectoryReader(input_dir=tmpdir)
                        docs = loader.load_data()

                        embed_model = HuggingFaceEmbedding(
                        model_name="nomic-ai/nomic-embed-text-v2-moe",
                        trust_remote_code=True,
                    )
                        Settings.embed_model = embed_model

                        llm_predictor = load_llm()
                        Settings.llm = llm_predictor

                        node_parser = MarkdownNodeParser()
                        index = VectorStoreIndex.from_documents(
                            documents=docs,
                            transformations=[node_parser],
                            show_progress=True,
                        )

                        qa_prompt = PromptTemplate(
                            "You are an AI assistant specialized in analyzing GitHub repositories.\n"
                            "Repository structure:\n{tree}\n---\n"
                            "Context:\n{context_str}\n---\n"
                            "Question: {query_str}\nAnswer:"
                        )
                        query_engine = index.as_query_engine(streaming=True)
                        query_engine.update_prompts({
                            "response_synthesizer:text_qa_template": qa_prompt
                        })

                        st.session_state.file_cache[cache_key] = (query_engine, tree)
                st.success("Repository loaded and indexed. Ready to chat!")
            else:
                st.info("Repository already loaded.")
        except Exception as e:
            st.error(f"Error loading repository: {e}")
            logger.error(f"Load error: {e}")

# --- Main UI: Chat Interface ---
col1, col2 = st.columns([6, 1])
with col1:
    st.header("Chat with GitHub RAG")
with col2:
    st.button("Clear Chat ↺", on_click=reset_chat)

# Display chat history
for msg in st.session_state.messages:
    with st.chat_message(msg["role"]):
        st.markdown(msg["content"])

# Chat input box
if prompt := st.chat_input("Ask a question about the repository..."):
    st.session_state.messages.append({"role": "user", "content": prompt})
    with st.chat_message("user"):
        st.markdown(prompt)

    repo_name = github_url.rstrip("/").split("/")[-1]
    cache_key = f"{session_id}-{repo_name}"

    if cache_key not in st.session_state.file_cache:
        st.error("Please load a repository first!")
    else:
        query_engine, tree = st.session_state.file_cache[cache_key]
        with st.chat_message("assistant"):
            placeholder = st.empty()
            response_text = ""
            try:
                response = query_engine.query(prompt)
                if hasattr(response, 'response_gen'):
                    for chunk in response.response_gen:
                        response_text += chunk
                        placeholder.markdown(response_text + "▌")
                else:
                    response_text = str(response)
                    placeholder.markdown(response_text)
            except GitHubRAGError as e:
                st.error(str(e))
                logger.error(f"Error in chat processing: {e}")
                response_text = "Sorry, I couldn't process that request."
            except Exception as e:
                st.error("An unexpected error occurred while processing your query")
                logger.error(f"Unexpected error in chat: {e}")
                response_text = "Sorry, something went wrong."
            placeholder.markdown(response_text)
            st.session_state.messages.append({"role": "assistant", "content": response_text})