app.py

import gradio as gr
import os
import aiohttp
import asyncio
from git import Repo, GitCommandError
from pathlib import Path
from datetime import datetime
import shutil
import json
import logging
import re
from typing import Dict, List, Optional, Tuple
import subprocess
import plotly.express as px
import plotly.graph_objects as go
from transformers import pipeline, AutoModelForCausalLM, AutoTokenizer
import threading
from http.server import HTTPServer, BaseHTTPRequestHandler
import speech_recognition as sr
from code_editor import code_editor
from functools import lru_cache
import hashlib
import markdown2
from concurrent.futures import ThreadPoolExecutor
from hdb_scan import HDBSCAN
import websockets
from websockets.exceptions import ConnectionClosed
from code_editor import code_editor, generate_code_snippet
# ========== Configuration ==========
WORKSPACE = Path("/tmp/issue_workspace")
WORKSPACE.mkdir(exist_ok=True)
GITHUB_API = "https://api.github.com/repos"
HF_INFERENCE_API = "https://api-inference.huggingface.co/models"
WEBHOOK_PORT = 8000
WS_PORT = 8001
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
executor = ThreadPoolExecutor(max_workers=4)

HF_MODELS = {
    "Mistral-8x7B": "mistralai/Mixtral-8x7B-Instruct-v0.1",
    "Llama-3-8B": "meta-llama/Meta-Llama-3-8B",
    "CodeLlama-34B": "codellama/CodeLlama-34b-Instruct-hf",
    "StarCoder2": "bigcode/starcoder2-15b"
}
DEFAULT_MODEL = "mistralai/Mixtral-8x7B-Instruct-v0.1"

# ========== Modern Theme ==========
theme = gr.themes.Soft(
    primary_hue="violet",
    secondary_hue="emerald",
    radius_size="xl",
    font=[gr.themes.GoogleFont("Inter"), "ui-sans-serif", "system-ui"]
).set(
    button_primary_background_fill="linear-gradient(90deg, #8B5CF6 0%, #EC4899 100%)",
    button_primary_text_color="white",
    button_primary_border_radius="12px",
    block_label_text_size="lg",
    block_label_text_weight="600",
    block_title_text_size="xl",
    block_title_text_weight="800",
    panel_background_fill="white",
    panel_border_radius="16px",
    block_shadow="*shadow_drop_lg",
)

# ========== Enhanced Webhook Handler ==========
class WebhookHandler(BaseHTTPRequestHandler):
    def do_POST(self):
        content_length = int(self.headers['Content-Length'])
        payload = json.loads(self.rfile.read(content_length).decode())
        event = self.headers.get('X-GitHub-Event')
        
        if event == 'issues':
            action = payload.get('action')
            if action in ['opened', 'reopened', 'closed', 'assigned']:
                asyncio.run_coroutine_threadsafe(
                    manager.handle_webhook_event(event, action, payload),
                    asyncio.get_event_loop()
                )
        
        self.send_response(200)
        self.end_headers()

# ========== AI-Powered Issue Manager ==========
class IssueManager:
    def __init__(self):
        self.issues: Dict[int, dict] = {}
        self.repo_url: Optional[str] = None
        self.repo: Optional[Repo] = None
        self.current_issue: Optional[int] = None
        self.github_token: Optional[str] = None
        self.hf_token: Optional[str] = None
        self.collaborators: Dict[str, dict] = {}
        self.points: int = 0
        self.severity_rules: Dict[str, List[str]] = {
            "Critical": ["critical", "urgent", "security", "crash"],
            "High": ["high", "important", "error", "regression"],
            "Medium": ["medium", "bug", "performance"],
            "Low": ["low", "documentation", "enhancement"]
        }
        self.issue_clusters: Dict[int, List[int]] = {}  # Store clusters
        self._init_local_models()
        self.ws_clients: List[websockets.WebSocketClientProtocol] = []
        self.code_editors: Dict[int, OTCodeEditor] = {}  # Store code editors for each issue
    
    def _init_local_models(self):
        self.code_model = pipeline(
            "text-generation",
            model="codellama/CodeLlama-7b-Instruct-hf",
            device_map="auto",
            torch_dtype="auto"
        )
        self.summarizer = pipeline(
            "summarization",
            model="philschmid/bart-large-cnn-samsum",
            device_map="auto"
        )
    
    @lru_cache(maxsize=100)
    async def cached_suggestion(self, issue_hash: str, model: str):
        return await self.suggest_resolution(issue_hash, model)

    async def handle_webhook_event(self, event: str, action: str, payload: dict):
        logger.info(f"Processing {event} {action} event")
        if action == 'closed':
            self.issues.pop(payload['issue']['number'], None)
        else:
            await self.crawl_issues(self.repo_url, self.github_token, self.hf_token)

    async def crawl_issues(self, repo_url: str, github_token: str, hf_token: str) -> Tuple[bool, str]:
        try:
            self.repo_url = repo_url
            self.github_token = github_token
            self.hf_token = hf_token
            self.repo = Repo.clone_from(repo_url, WORKSPACE / "repo")
            headers = {"Authorization": f"token {github_token}"}
            async with aiohttp.ClientSession(headers=headers) as session:
                async with session.get(f"{GITHUB_API}/{repo_url}/issues") as response:
                    issues = await response.json()
                    for issue in issues:
                        self.issues[issue['number']] = issue
            await self._cluster_similar_issues()
            return True, f"Found {len(self.issues)} issues (clustered into {len(self.issue_clusters)} groups)"
        except Exception as e:
            logger.error(f"Crawl error: {e}")
            return False, str(e)
    
    async def _cluster_similar_issues(self):
        embeddings = await self._generate_embeddings()
        # Use HDBSCAN for clustering
        clusterer = HDBSCAN(min_cluster_size=2, metric='cosine')
        clusters = clusterer.fit_predict(embeddings)
        self.issue_clusters = {}
        for i, cluster_id in enumerate(clusters):
            if cluster_id not in self.issue_clusters:
                self.issue_clusters[cluster_id] = []
            self.issue_clusters[cluster_id].append(i)

    async def _generate_embeddings(self):
        async with aiohttp.ClientSession() as session:
            texts = [f"{i['title']} {i['body']}" for i in self.issues.values()]
            response = await session.post(
                f"{HF_INFERENCE_API}/sentence-transformers/all-mpnet-base-v2",
                headers={"Authorization": f"Bearer {self.hf_token}"},
                json={"inputs": texts}
            )
            return await response.json()

    async def generate_code_patch(self, issue_number: int) -> dict:
        issue = self.issues[issue_number]
        context = await self._get_code_context(issue_number)
        prompt = f"""<issue>
{issue['title']}
{issue['body']}
</issue>

<code_context>
{context}
</code_context>

Generate a JSON patch file with specific changes needed to resolve this issue."""
        
        response = self.code_model(
            prompt,
            max_length=1024,
            temperature=0.2,
            return_full_text=False
        )
        try:
            return json.loads(response[0]['generated_text'])
        except json.JSONDecodeError:
            return {"error": "Failed to parse AI-generated patch"}

    async def _get_code_context(self, issue_number: int) -> str:
        repo_path = WORKSPACE / f"repo-{issue_number}"
        code_files = list(repo_path.glob('**/*.py')) + list(repo_path.glob('**/*.js'))
        return "\n".join(f.read_text()[:1000] for f in code_files[:5])

    async def suggest_resolution(self, issue_hash: str, model: str) -> str:
        issue = self.issues[int(issue_hash)]
        prompt = f"""
        ## Issue: {issue['title']}

        {issue['body']}

        Suggest a solution to this issue. 
        """
        
        async with aiohttp.ClientSession() as session:
            response = await session.post(
                f"{HF_INFERENCE_API}/{model}",
                headers={"Authorization": f"Bearer {self.hf_token}"},
                json={"inputs": prompt}
            )
            return await response.json()

    async def broadcast_collaboration_status(self):
        while True:
            try:
                await asyncio.sleep(1)
                # Send collaborator status to all connected clients
                await asyncio.gather(
                    *[client.send(json.dumps([{"name": name, "status": status} for name, status in self.collaborators.items()])) for client in self.ws_clients]
                )
            except ConnectionClosed:
                # Handle client disconnections
                pass

    async def handle_code_editor_update(self, issue_num: int, delta: str):
        if issue_num not in self.code_editors:
            return
        self.code_editors[issue_num].apply_delta(json.loads(delta))
        await asyncio.gather(
            *[client.send(json.dumps({"type": "code_update", "issue_num": issue_num, "delta": delta})) for client in self.ws_clients]
        )

# ========== Enhanced UI Components ==========
def create_ui():
    with gr.Blocks(theme=theme, title="🤖 AI Issue Resolver Pro", css=".gradio-container {max-width: 1200px !important}") as app:
        gr.Markdown("""
        # 🚀 AI Issue Resolver Pro
        *Next-generation issue resolution powered by AI collaboration*
        """)
        
        with gr.Row(variant="panel"):
            with gr.Column(scale=2):
                repo_url = gr.Textbox(label="GitHub Repo", placeholder="https://github.com/org/repo", info="Connect your repository")
                github_token = gr.Textbox(label="GitHub Token", type="password")
                hf_token = gr.Textbox(label="HF Token", type="password")
            
            with gr.Column(scale=1):
                model_select = gr.Dropdown(choices=list(HF_MODELS.keys()), value="Mistral-8x7B", 
                                        label="AI Model", info="Choose your resolution strategy")
                language_select = gr.Dropdown(choices=["python", "javascript", "java", "c", "cpp", "html", "css", "bash", "ruby", "go", "php", "rust", "typescript"], 
                                               value="python", label="Select Language", info="Choose the programming language for the code editor")
                crawl_btn = gr.Button("🚀 Scan Repository", variant="primary")

        with gr.Tabs():
            with gr.Tab("📋 Issue Board", id="board"):
                with gr.Row():
                    issue_list = gr.Dataframe(
                        headers=["ID", "Title", "Severity", "Cluster"],
                        datatype=["number", "str", "str", "number"],
                        interactive=True,
                        height=600
                    )
                    with gr.Column(scale=1):
                        stats_plot = gr.Plot()
                        collab_status = gr.HTML("<h3>👥 Active Collaborators</h3><div id='collab-list'></div>")

            with gr.Tab("💻 Resolution Studio", id="studio"):
                with gr.Row():
                    with gr.Column(scale=1):
                        issue_num = gr.Number(label="Issue #", precision=0)
                        issue_viz = gr.HTML()
                        ai_tools = gr.Accordion("🛠️ AI Tools")
                        with ai_tools:
                            suggest_btn = gr.Button("🧠 Suggest Resolution")
                            patch_btn = gr.Button("📝 Generate Patch")
                            test_btn = gr.Button("🧪 Create Tests")
                            impact_btn = gr.Button("📊 Impact Analysis")
                        
                    with gr.Column(scale=2):
                        with gr.Tabs():
                            with gr.Tab("Code Editor"):
                                code_edit = gr.HTML(elem_id="code-editor-container")
                            with gr.Tab("AI Chat"):
                                chat = gr.ChatInterface(
                                    self._ai_chat,
                                    additional_inputs=[issue_num]
                                )

            with gr.Tab("📈 Analytics", id="analytics"):
                with gr.Row():
                    gr.Markdown("### 📅 Resolution Timeline")
                    timeline = gr.Timeline()
                with gr.Row():
                    gr.Markdown("### 🏆 Achievement System")
                    badges = gr.HTML("<div class='badges'></div>")
        
        # Enhanced Event Handlers
        async def generate_patch(issue_num):
            patch = await manager.generate_code_patch(issue_num)
            return gr.JSON(value=patch)

        def update_code_editor(files):
            return code_editor(value=files, language=language_select.value)

        issue_list.select(
            fn=lambda evt: (evt[0], manager.issues[evt[0]]['body']),
            outputs=[issue_num, issue_body]
        ).then(
            fn=lambda num: generate_issue_preview(num),
            outputs=issue_viz
        )

        app.load(
            fn=init_collaboration,
            inputs=[],
            outputs=collab_status,
            _js=web_socket_js()
        )

        crawl_btn.click(
            fn=lambda repo, token, hf_token: manager.crawl_issues(repo, token, hf_token),
            inputs=[repo_url, github_token, hf_token],
            outputs=[issue_list, stats_plot]
        )

        suggest_btn.click(
            fn=lambda issue, model: manager.cached_suggestion(issue, model),
            inputs=[issue_num, model_select],
            outputs=chat
        )

        patch_btn.click(
            fn=generate_patch,
            inputs=[issue_num],
            outputs=chat
        )

        # Add more event handlers for other AI tools
        # ...

        issue_num.change(
            fn=lambda issue_num: create_code_editor(issue_num, language_select.value),
            inputs=[issue_num, language_select],
            outputs=code_edit
        )

    # Add real-time collaboration
    def web_socket_js():
        return """
        <script>
            const collabWs = new WebSocket('ws://localhost:8001');
            collabWs.onmessage = function(event) {
                const data = JSON.parse(event.data);
                if (data.type === 'code_update') {
                    const issueNum = data.issue_num;
                    const delta = data.delta;
                    const codeEditor = document.getElementById(`code-editor-${issueNum}`);
                    if (codeEditor) {
                        codeEditor.applyDelta(delta);
                    }
                } else if (data.type === 'collaboration_status') {
                    document.getElementById('collab-list').innerHTML = 
                        data.map(u => `<div class="collab-item">${u.name}: ${u.status}</div>`).join('');
                }
            };
        </script>
        """

    # Start enhanced webhook server
    webhook_server = HTTPServer(("", WEBHOOK_PORT), WebhookHandler)
    threading.Thread(target=webhook_server.serve_forever, daemon=True).start()

    # Start WebSocket server
    async def start_ws_server():
        async with websockets.serve(handle_ws_connection, "localhost", WS_PORT):
            await asyncio.Future()

    async def handle_ws_connection(websocket: websockets.WebSocketClientProtocol, path):
        manager.ws_clients.append(websocket)
        try:
            async for message in websocket:
                data = json.loads(message)
                if data.get("type") == "code_update":
                    await manager.handle_code_editor_update(data["issue_num"], data["delta"])
        finally:
            manager.ws_clients.remove(websocket)

    asyncio.run(start_ws_server())
    threading.Thread(target=lambda: asyncio.run(manager.broadcast_collaboration_status()), daemon=True).start()

    return app

# ========== Execution ==========
if __name__ == "__main__":
    manager = IssueManager()
    app = create_ui()
    app.launch(
        share=True,
        auth=("admin", os.getenv("APP_PASSWORD")),
        favicon_path="https://huggingface.co/front/assets/huggingface_logo-noborder.svg"
    )