app.py

import gradio as gr
import requests
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

# Initialize the CodeBERT model and tokenizer
tokenizer = AutoTokenizer.from_pretrained("microsoft/CodeBERT-base")
model = AutoModelForSeq2SeqLM.from_pretrained("microsoft/CodeBERT-base")

def fetch_repo_contents(repo_url):
    # Extract username and repo name from URL
    username, repo_name = repo_url.split("github.com/")[-1].split("/")
    
    # Fetch repo contents using GitHub API
    api_url = f"https://api.github.com/repos/{username}/{repo_name}/contents"
    response = requests.get(api_url)
    response.raise_for_status()
    return response.json()

def generate_chatbot_response(repo_url, question):
    # Fetch repository contents
    repo_contents = fetch_repo_contents(repo_url)

    # Generate a prompt based on the user's question and repository contents
    prompt = f"Answer the question about the repository {repo_url}: {question}\n\n"
    for item in repo_contents:
        prompt += f"{item['name']}:\n{item['download_url']}\n\n"

    # Tokenize the prompt and generate a response using the CodeBERT model
    inputs = tokenizer.encode(prompt, return_tensors="pt", max_length=1024, truncation=True)
    outputs = model.generate(inputs, max_length=150, num_return_sequences=1)
    response = tokenizer.decode(outputs[0], skip_special_tokens=True)

    return response

# Gradio UI
repo_url_input = gr.inputs.Textbox(lines=1, label="GitHub Repository URL")
question_input = gr.inputs.Textbox(lines=2, label="Question")
output_text = gr.outputs.Textbox(label="Answer")

gr.Interface(
    generate_chatbot_response,
    inputs=[repo_url_input, question_input],
    outputs=output_text,
    title="Create a Conversational AI Chatbot for Your Public GitHub Repository Codebase",
    theme="huggingface_dark",
).launch()