diff --git "a/app2.py" "b/app2.py"
--- "a/app2.py"
+++ "b/app2.py"
@@ -1,1366 +1,3798 @@
-import json
+import gradio as gr
import os
+import aiohttp
+import asyncio
+from git import Repo, GitCommandError, InvalidGitRepositoryError, NoSuchPathError
+from pathlib import Path
+from datetime import datetime, timedelta, timezone
+import shutil
+import json
+import logging
import re
+from typing import Dict, List, Optional, Tuple, Any
+import subprocess
+import plotly.express as px
+import plotly.graph_objects as go
import time
-import logging
-import mimetypes
-import zipfile
-import tempfile
-import chardet
-import io # Needed for processing CSV from string
-import csv # Needed for CSV
-import xml.etree.ElementTree as ET # Needed for XML
-from datetime import datetime
-from typing import List, Dict, Optional, Union, Tuple, Any # Added Any for extracted_data
-from pathlib import Path
-from urllib.parse import urlparse, urljoin
-import requests
-import validators
-import gradio as gr
-from diskcache import Cache
-from bs4 import BeautifulSoup
-from fake_useragent import UserAgent
-from cleantext import clean
-import qrcode
-from PIL import Image, ImageDraw, ImageFont
-import numpy as np
-import tarfile
-import gzip
-import math
-
-# Setup enhanced logging with more detailed formatting
-logging.basicConfig(
- level=logging.INFO,
- format='%(asctime)s - %(levelname)s - [%(filename)s:%(lineno)d] - %(message)s',
- handlers=[
- logging.StreamHandler(),
- logging.FileHandler('app.log', encoding='utf-8')
- ])
-logger = logging.getLogger(__name__) # logger is now defined here
-
-# Conditional imports for document processing
+import random
+import pandas as pd
+from collections import Counter
+import string
+from concurrent.futures import ThreadPoolExecutor
+from hdbscan import HDBSCAN
+# --- Required Imports ---
+import hashlib
+from functools import lru_cache
+import threading
+from http.server import HTTPServer, BaseHTTPRequestHandler
+import markdown2
+import websockets
+from websockets.server import WebSocketServerProtocol
+from websockets.exceptions import ConnectionClosed, ConnectionClosedOK, ConnectionAbortedError, ConnectionResetError, WebSocketException
+import signal # For graceful shutdown
+# ---------------------
+
+# Assuming code_editor is available, e.g., installed via pip or included locally
try:
- from PyPDF2 import PdfReader
- PDF_SUPPORT = True
+ from code_editor import code_editor
except ImportError:
- PDF_SUPPORT = False
- logger.warning("PyPDF2 not installed. PDF file processing will be limited.") # logger is available
-
-try:
- from docx import Document
- DOCX_SUPPORT = True
-except ImportError:
- DOCX_SUPPORT = False
- logger.warning("python-docx not installed. DOCX file processing will be limited.") # logger is available
-
-try:
- from pyth.plugins.rtf15.reader import Rtf15Reader
- from pyth.plugins.plaintext.writer import PlaintextWriter
- RTF_SUPPORT = True
-except ImportError:
- RTF_SUPPORT = False
- logger.warning("pyth not installed. RTF file processing will be limited.")
-
-try:
- from odf.opendocument import OpenDocumentText
- from odf import text as odftext
- ODT_SUPPORT = True
-except ImportError:
- ODT_SUPPORT = False
- logger.warning("odfpy not installed. ODT file processing will be limited.")
-
-
-# Setup enhanced logging with more detailed formatting
-logging.basicConfig(
- level=logging.INFO,
- format='%(asctime)s - %(levelname)s - [%(filename)s:%(lineno)d] - %(message)s',
- handlers=[
- logging.StreamHandler(),
- logging.FileHandler('app.log', encoding='utf-8')
- ])
+ logging.error("The 'code_editor' Gradio component is not installed or available.")
+ logging.error("Please install it, e.g., 'pip install gradio_code_editor'")
+ def code_editor(*args, **kwargs):
+ logging.warning("Using dummy code_editor. Code editing and collaboration will not function.")
+ # Create a dummy component that looks like a Textbox but is non-interactive
+ return gr.Textbox(label=kwargs.get('label', 'Code Editor (Unavailable)'), interactive=False, value="Error: Code editor component not found. Install 'gradio_code_editor'.", lines=10)
+
+
+# ========== Configuration ==========
+WORKSPACE = Path("./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 = int(os.environ.get("WEBHOOK_PORT", 8000))
+WS_PORT = int(os.environ.get("WS_PORT", 8001))
+GRADIO_PORT = int(os.environ.get("GRADIO_PORT", 7860))
+
+# Configure logging
+logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(name)s - %(levelname)s - %(message)s')
logger = logging.getLogger(__name__)
+# Set a higher logging level for libraries that are too verbose, e.g.:
+logging.getLogger("websockets").setLevel(logging.WARNING)
+logging.getLogger("aiohttp").setLevel(logging.WARNING)
+logging.getLogger("urllib3").setLevel(logging.WARNING) # Might be used by git or other libs
+logging.getLogger("git").setLevel(logging.WARNING)
+logging.getLogger("hdbscan").setLevel(logging.WARNING) # HDBSCAN can be chatty
+
+
+# Use ThreadPoolExecutor for any synchronous blocking operations if needed,
+# but most heavy lifting (API calls, git) is now async.
+executor = ThreadPoolExecutor(max_workers=4)
+
+
+# Example HF models (replace with your actual models)
+# Ensure these models are suitable for the tasks (text generation, embeddings)
+HF_MODELS = {
+ "Mistral-8x7B": "mistralai/Mixtral-8x7B-Instruct-v0.1",
+ "Llama-2-7B-chat": "huggingface/llama-2-7b-chat-hf",
+ "CodeLlama-34B": "codellama/CodeLlama-34b-Instruct-hf",
+ "Gemma-7B-it": "google/gemma-7b-it", # Added another option
+}
+# Embedding model ID - fixed, not user selectable
+HF_EMBEDDING_MODEL = "sentence-transformers/all-mpnet-base-v2"
+
+# Select default models defensively
+DEFAULT_MODEL_KEY = "Mistral-8x7B" if "Mistral-8x7B" in HF_MODELS else (list(HF_MODELS.keys())[0] if HF_MODELS else None)
+DEFAULT_MODEL_ID = HF_MODELS.get(DEFAULT_MODEL_KEY, None)
+
+DEFAULT_IDLE_MODEL_KEY = "Gemma-7B-it" if "Gemma-7B-it" in HF_MODELS else DEFAULT_MODEL_KEY # Prefer a smaller one if available
+DEFAULT_IDLE_MODEL_ID = HF_MODELS.get(DEFAULT_IDLE_MODEL_KEY, DEFAULT_MODEL_ID)
+
+if not HF_MODELS:
+ logger.critical("No HF models configured! AI features will be disabled.")
+elif DEFAULT_MODEL_ID is None:
+ logger.critical(f"Default model key '{DEFAULT_MODEL_KEY}' not found in configured models. AI features may be limited.")
+if DEFAULT_IDLE_MODEL_ID is None:
+ logger.warning(f"Idle model key '{DEFAULT_IDLE_MODEL_KEY}' not found or no models configured. Idle tasks may be disabled or use the default model if available.")
+
+
+# --- Idle State Configuration ---
+STALE_ISSUE_THRESHOLD_DAYS = 30
+MAX_SUMMARY_COMPUTATIONS_PER_CYCLE = 2
+MAX_CONTEXT_COMPUTATIONS_PER_CYCLE = 3
+MAX_MISSING_INFO_COMPUTATIONS_PER_CYCLE = 1
+MAX_ANALYSIS_COMPUTATIONS_PER_CYCLE = 1
+RECLUSTER_THRESHOLD = 5 # Number of significant webhook changes before re-clustering is flagged
+IDLE_PROCESSING_INTERVAL_SECONDS = 60.0 # How often the idle task loop runs
+
+# ========== Placeholder OTCodeEditor Class ==========
+# WARNING: This is a placeholder and DOES NOT implement Operational Transformation.
+# Concurrent edits WILL lead to data loss or inconsistencies.
+class OTCodeEditor:
+ """
+ Placeholder for an Operational Transformation (OT) enabled code editor backend.
+ This implementation is NOT thread-safe and does NOT handle concurrent edits correctly.
+ It merely logs received deltas and maintains a basic revision counter.
+ The actual document state is held client-side by the Gradio code_editor component.
+ A real collaborative editor would require a robust OT backend to manage
+ the authoritative document state and transform operations.
+ """
+ def __init__(self, initial_value: Dict[str, str]):
+ # In a real OT system, this would initialize the document state
+ # For this placeholder, we just store the initial files dict
+ self.files: Dict[str, str] = initial_value.copy()
+ self.revision = 0 # Basic revision counter, not used for OT logic
+ logger.debug(f"OTCodeEditor initialized with files: {list(self.files.keys())}")
+
+ def apply_delta(self, delta: Dict[str, Any]):
+ # VERY basic placeholder: This logs the delta but does NOT perform OT.
+ # It does NOT handle concurrent edits safely.
+ # In a real OT system, this would transform the delta against the current state
+ # and apply it, incrementing the revision based on successful application.
+ logger.warning(f"Placeholder apply_delta called. Delta: {str(delta)[:200]}. "
+ "WARNING: Full Operational Transformation is NOT implemented. Concurrent edits are UNSAFE.")
+ # The Gradio component holds the actual text state client-side.
+ # A real OT backend would use the delta to update its authoritative state.
+ # We only increment revision for basic tracking visibility if needed.
+ self.revision += 1
+
+ def get_content(self) -> Dict[str, str]:
+ # This is not used by the current Gradio code_editor integration,
+ # as the component holds the state client-side.
+ # In a real OT system, this would return the current authoritative document state.
+ return self.files.copy()
+
+# ========== Modern Theme ==========
+try:
+ theme = gr.themes.Soft(
+ primary_hue="violet",
+ secondary_hue="emerald",
+ radius_size="lg",
+ 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",
+ block_label_text_size="lg",
+ block_label_text_weight="600",
+ block_title_text_size="lg",
+ block_title_text_weight="800",
+ panel_background_fill="white",
+ block_shadow="*shadow_drop_lg",
+ # Add some more modern touches
+ input_background_fill="#f9fafb",
+ input_border_color="#e5e7eb",
+ input_border_radius="md",
+ button_secondary_background_fill="#f3f4f6",
+ button_secondary_text_color="#374151",
+ button_secondary_border_color="#d1d5db",
+ table_border_color="#e5e7eb",
+ table_row_background_even="#f9fafb",
+ table_row_background_odd="#ffffff",
+ # Use slightly softer colors for plots if default is too bright
+ # (This might need specific Plotly config instead of theme)
+ )
+except AttributeError as e:
+ logger.warning(f"Could not apply all theme settings (might be Gradio version difference): {e}. Using default Soft theme.")
+ theme = gr.themes.Soft()
+
+
+# ========== Enhanced Webhook Handler ==========
+class WebhookHandler(BaseHTTPRequestHandler):
+ manager_instance: Optional['IssueManager'] = None
+ main_loop: Optional[asyncio.AbstractEventLoop] = None # Store reference to the main asyncio loop
+
+ def do_POST(self):
+ content_length = int(self.headers.get('Content-Length', 0))
+ if content_length == 0:
+ self.send_response(400)
+ self.send_header("Content-type", "text/plain")
+ self.end_headers()
+ self.wfile.write(b"Empty payload")
+ logger.warning("Received empty webhook payload.")
+ return
-# Ensure output directories exist with modern structure
-OUTPUTS_DIR = Path('output')
-QR_CODES_DIR = OUTPUTS_DIR / 'qr_codes'
-TEMP_DIR = OUTPUTS_DIR / 'temp'
-for directory in [OUTPUTS_DIR, QR_CODES_DIR, TEMP_DIR]:
- directory.mkdir(parents=True, exist_ok=True)
-
-class EnhancedURLProcessor:
- """Advanced URL processing with enhanced content extraction"""
- def __init__(self):
- self.session = requests.Session()
- self.timeout = 15 # Extended timeout for larger content
- self.max_retries = 3
- self.user_agent = UserAgent()
-
- # Enhanced headers for better site compatibility
- self.session.headers.update({
- 'User-Agent': self.user_agent.random,
- 'Accept': 'text/html, application/json, application/xml, text/plain, */*', # Request common types
- 'Accept-Language': 'en-US,en;q=0.9',
- 'Accept-Encoding': 'gzip, deflate, br',
- 'Connection': 'keep-alive',
- 'Upgrade-Insecure-Requests': '1', # May be ignored for non-HTML
- 'Sec-Fetch-Dest': 'document',
- 'Sec-Fetch-Mode': 'navigate',
- 'Sec-Fetch-Site': 'none',
- 'Sec-Fetch-User': '?1',
- 'DNT': '1'
- })
-
- def validate_url(self, url: str) -> Dict:
- """Enhanced URL validation with detailed feedback"""
try:
- if not validators.url(url):
- return {'is_valid': False, 'message': 'Invalid URL format', 'details': 'URL must begin with http:// or https://'}
- parsed = urlparse(url)
- if not all([parsed.scheme, parsed.netloc]):
- return {'is_valid': False, 'message': 'Incomplete URL', 'details': 'Missing scheme or domain'}
- # Try HEAD request first to check accessibility
- try:
- head_response = self.session.head(url, timeout=5)
- head_response.raise_for_status()
- final_url = head_response.url # Capture potential redirects
- except requests.exceptions.RequestException:
- # If HEAD fails, try GET as some servers don't support HEAD
- response = self.session.get(url, timeout=self.timeout)
- response.raise_for_status()
- final_url = response.url # Capture potential redirects
-
- return {
- 'is_valid': True,
- 'message': 'URL is valid and accessible',
- 'details': {
- 'final_url': final_url,
- 'content_type': head_response.headers.get('Content-Type', 'unknown'),
- 'server': head_response.headers.get('Server', 'unknown'),
- 'size': head_response.headers.get('Content-Length', 'unknown')
- }
- }
+ payload_bytes = self.rfile.read(content_length)
+ payload = json.loads(payload_bytes.decode('utf-8'))
+ except json.JSONDecodeError:
+ logger.error(f"Invalid JSON payload received: {payload_bytes[:500]}")
+ self.send_response(400)
+ self.send_header("Content-type", "text/plain")
+ self.end_headers()
+ self.wfile.write(b"Invalid JSON payload")
+ return
except Exception as e:
- return {'is_valid': False, 'message': f'URL validation failed: {str(e)}', 'details': str(e)}
-
- def fetch_content(self, url: str, retry_count: int = 0) -> Optional[Dict]:
- """Enhanced content fetcher with retry mechanism and complete character extraction"""
- try:
- logger.info(f"Fetching content from URL: {url} (Attempt {retry_count + 1}/{self.max_retries})")
-
- # Update User-Agent randomly for each request
- self.session.headers.update({'User-Agent': self.user_agent.random})
+ logger.error(f"Error reading webhook payload: {e}")
+ self.send_response(500)
+ self.end_headers()
+ return
+
+ event = self.headers.get('X-GitHub-Event')
+ delivery_id = self.headers.get('X-GitHub-Delivery')
+ logger.info(f"Received GitHub webhook event: {event} (Delivery ID: {delivery_id})")
+
+ if event == 'issues' and WebhookHandler.manager_instance and WebhookHandler.main_loop:
+ action = payload.get('action')
+ logger.info(f"Issue action: {action}")
+ # Handle common actions that affect issue state or content
+ if action in ['opened', 'reopened', 'closed', 'assigned', 'unassigned', 'edited', 'labeled', 'unlabeled', 'milestoned', 'demilestoned']:
+ # Check if the loop is running before scheduling
+ if WebhookHandler.main_loop.is_running():
+ # Schedule the async handler to run in the main event loop
+ # Use run_coroutine_threadsafe as this is called from a different thread
+ asyncio.run_coroutine_threadsafe(
+ WebhookHandler.manager_instance.handle_webhook_event(event, action, payload),
+ WebhookHandler.main_loop
+ )
+ logger.debug(f"Scheduled webhook processing for action '{action}' in main loop.")
+ else:
+ logger.error("Asyncio event loop is not running in the target thread for webhook. Cannot process event.")
+ # Respond with an error as processing couldn't be scheduled
+ self.send_response(500)
+ self.send_header("Content-type", "text/plain")
+ self.end_headers()
+ self.wfile.write(b"Async processing loop not available.")
+ return
+ else:
+ logger.info(f"Webhook action '{action}' received but not actively handled by current logic.")
+ elif event == 'ping':
+ logger.info("Received GitHub webhook ping.")
+ else:
+ logger.warning(f"Unhandled event type: {event} or manager/loop not initialized.")
- response = self.session.get(url, timeout=self.timeout)
- response.raise_for_status()
- final_url = response.url # Capture potential redirects
+ # Always respond 200 OK for successful receipt, even if action is ignored
+ self.send_response(200)
+ self.send_header("Content-type", "text/plain")
+ self.end_headers()
+ self.wfile.write(b"OK")
- # Detect encoding
- if response.encoding is None or response.encoding == 'ISO-8859-1': # chardet often better than default response.encoding for text
- encoding_detection = chardet.detect(response.content)
- encoding = encoding_detection['encoding'] or 'utf-8'
- logger.debug(f"Detected encoding '{encoding}' with confidence {encoding_detection['confidence']:.2f} for {url}")
+# ========== AI-Powered Issue Manager ==========
+class IssueManager:
+ def __init__(self):
+ self.issues: Dict[int, dict] = {}
+ self.repo_url: Optional[str] = None
+ self.repo_owner: Optional[str] = None
+ self.repo_name: Optional[str] = None
+ self.repo_local_path: Optional[Path] = None
+ self.repo: Optional[Repo] = None
+ self.github_token: Optional[str] = None
+ self.hf_token: Optional[str] = None
+ self.collaborators: Dict[str, dict] = {} # {client_id: {name: str, status: str}}
+ self.points: int = 0 # Placeholder for potential gamification
+ self.severity_rules: Dict[str, List[str]] = {
+ "Critical": ["critical", "urgent", "security", "crash", "blocker", "p0", "s0"],
+ "High": ["high", "important", "error", "regression", "major", "p1", "s1"],
+ "Medium": ["medium", "bug", "performance", "minor", "p2", "s2"],
+ "Low": ["low", "documentation", "enhancement", "trivial", "feature", "p3", "s3", "chore", "refactor", "question", "help wanted"]
+ }
+ self.issue_clusters: Dict[int, List[int]] = {} # {cluster_id: [issue_index_in_list, ...]}
+ self.issue_list_for_clustering: List[dict] = [] # List of issue dicts used for the last clustering run
+ self.ws_clients: List[WebSocketServerProtocol] = []
+ self.code_editors: Dict[int, OTCodeEditor] = {} # {issue_id: OTCodeEditor_instance}
+ # Get or create the loop in the thread where the manager is initialized (main thread)
+ try:
+ self.main_loop = asyncio.get_running_loop()
+ logger.debug(f"IssueManager found running asyncio loop: {id(self.main_loop)}")
+ except RuntimeError:
+ self.main_loop = asyncio.new_event_loop()
+ asyncio.set_event_loop(self.main_loop)
+ logger.debug(f"IssueManager created and set new asyncio loop: {id(self.main_loop)}")
+
+ self.broadcast_task: Optional[asyncio.Task] = None
+ self.idle_task: Optional[asyncio.Task] = None
+
+ # --- State for Idle Processing Results ---
+ self.precomputed_context: Dict[int, Dict[str, Any]] = {} # {issue_id: {content: str, files: list, error: str, timestamp: float}}
+ self.precomputed_summaries: Dict[int, Dict[str, Any]] = {} # {issue_id: {summary: str, error: str, timestamp: float}}
+ self.precomputed_missing_info: Dict[int, Dict[str, Any]] = {} # {issue_id: {info_needed: str, error: str, timestamp: float}}
+ self.precomputed_analysis: Dict[int, Dict[str, Any]] = {} # {issue_id: {hypothesis: str, error: str, timestamp: float}}
+ # self.code_embeddings: Dict[str, List[float]] = {} # Not currently used after clustering
+ self.potential_duplicates: Dict[int, List[int]] = {} # {issue_id: [duplicate_issue_id, ...]}
+ self.stale_issues: List[int] = [] # [issue_id, ...]
+ self.high_priority_candidates: List[int] = [] # [issue_id, ...]
+ self.last_webhook_time: float = time.time() # Track last webhook for potential future use
+ self.needs_recluster: bool = False
+ self._webhook_change_count = 0
+
+ # --- Configuration for Idle Tasks ---
+ self.idle_processing_interval = IDLE_PROCESSING_INTERVAL_SECONDS
+ self.max_context_computations_per_cycle = MAX_CONTEXT_COMPUTATIONS_PER_CYCLE
+ self.max_summary_computations_per_cycle = MAX_SUMMARY_COMPUTATIONS_PER_CYCLE
+ self.max_missing_info_computations_per_cycle = MAX_MISSING_INFO_COMPUTATIONS_PER_CYCLE
+ self.max_analysis_computations_per_cycle = MAX_ANALYSIS_COMPUTATIONS_PER_CYCLE
+ self.stale_issue_threshold_days = STALE_ISSUE_THRESHOLD_DAYS
+ self.recluster_threshold = RECLUSTER_THRESHOLD
+
+ # Shutdown signals (placeholders, set by main execution block)
+ self.stop_ws_server = None
+ self.stop_webhook_server = None
+
+ def start_broadcast_loop(self):
+ """Starts the periodic broadcast task."""
+ # Ensure task is created in the correct loop
+ if not self.main_loop.is_running():
+ logger.error("Cannot start broadcast loop: Main event loop is not running.")
+ return
+
+ if not self.broadcast_task or self.broadcast_task.done():
+ self.broadcast_task = self.main_loop.create_task(self.broadcast_collaboration_status())
+ logger.info("Started collaboration status broadcast loop.")
+ else:
+ logger.debug("Broadcast loop already running.")
+
+
+ def stop_broadcast_loop(self):
+ """Stops the periodic broadcast task."""
+ if self.broadcast_task and not self.broadcast_task.done():
+ logger.info("Stopping collaboration status broadcast loop...")
+ self.broadcast_task.cancel()
+ # Await the task to finish cancellation in an async context if needed,
+ # but cancelling is sufficient to signal it to stop.
+ self.broadcast_task = None # Clear the reference
+
+
+ def _get_issue_hash(self, issue_data: Optional[dict]) -> str:
+ """Generates a hash based on key issue content for caching AI suggestions."""
+ if not issue_data:
+ return "empty_issue_hash" # Handle cases where issue_data is None
+
+ content = f"{issue_data.get('title', '')}{issue_data.get('body', '')}{','.join(issue_data.get('labels',[]))}"
+ return hashlib.md5(content.encode('utf-8', errors='ignore')).hexdigest() # Use utf-8 and ignore errors
+
+
+ @lru_cache(maxsize=100)
+ async def cached_suggestion(self, issue_hash: str, model_key: str) -> str:
+ """Retrieves or generates an AI suggestion, using an LRU cache based on issue content hash."""
+ logger.debug(f"Checking cache for suggestion: hash={issue_hash}, model={model_key}")
+ # The cache decorator handles the cache hit/miss logic.
+ # If it's a miss, the decorated function body is executed.
+
+ # Find the issue data corresponding to the hash
+ found_issue = None
+ # This linear scan is inefficient for many issues, but hashes are only for cache keys.
+ # A dict mapping hashes to issue IDs could be more efficient if this becomes a bottleneck.
+ # However, issues dict is not huge usually, so this is likely fine.
+ for issue in self.issues.values():
+ if self._get_issue_hash(issue) == issue_hash:
+ found_issue = issue
+ break
+
+ if not found_issue:
+ logger.error(f"Could not find issue data for hash {issue_hash} in current state. Suggestion cannot be generated.")
+ # Corrected: Return error message here if issue data is missing
+ return "Error: Issue data for this suggestion request (hash) not found in current state. The issue might have been updated or closed. Please re-select the issue."
+
+ if model_key not in HF_MODELS or HF_MODELS.get(model_key) is None:
+ logger.error(f"Invalid or unconfigured model key requested: {model_key}")
+ return f"Error: Invalid or unconfigured model key: {model_key}"
+
+ logger.info(f"Cache miss or first request for issue hash {issue_hash}. Requesting suggestion from {model_key}.")
+ # Call the actual suggestion generation function
+ return await self.suggest_resolution(found_issue, model_key)
+
+ async def handle_webhook_event(self, event: str, action: str, payload: dict):
+ """Processes incoming webhook events to update the issue state."""
+ logger.info(f"Processing webhook event: {event}, action: {action}")
+ issue_data = payload.get('issue')
+ repo_data = payload.get('repository')
+
+ if not issue_data or not repo_data:
+ logger.warning("Webhook payload missing 'issue' or 'repository' data.")
+ return
+
+ event_repo_url = repo_data.get('html_url')
+ # Only process events for the currently loaded repository
+ # Use .rstrip("/") on both sides for robust comparison
+ if self.repo_url is None or event_repo_url is None or event_repo_url.rstrip("/") != self.repo_url.rstrip("/"):
+ logger.info(f"Ignoring webhook event for different repository: {event_repo_url} (Current: {self.repo_url})")
+ return
+
+ issue_number = issue_data.get('number')
+ if issue_number is None: # Check explicitly for None
+ logger.warning("Webhook issue data missing 'number'.")
+ return
+
+ needs_ui_update = False
+ significant_change = False # Flag for changes affecting clustering/content/AI caches
+
+ if action == 'closed':
+ logger.info(f"Webhook: Removing closed issue {issue_number} from active list.")
+ if issue_number in self.issues:
+ self.issues.pop(issue_number)
+ needs_ui_update = True
+ significant_change = True # Closing is a significant change
+ # Clean up associated cached/computed data
+ self.precomputed_context.pop(issue_number, None)
+ self.precomputed_summaries.pop(issue_number, None)
+ self.precomputed_missing_info.pop(issue_number, None)
+ self.precomputed_analysis.pop(issue_number, None)
+ self.potential_duplicates.pop(issue_number, None)
+ # Remove from lists if present (use list comprehension for safe removal)
+ self.stale_issues = [i for i in self.stale_issues if i != issue_number]
+ self.high_priority_candidates = [i for i in self.high_priority_candidates if i != issue_number]
+ # Remove the code editor instance for the closed issue
+ self.code_editors.pop(issue_number, None)
+ logger.debug(f"Cleaned up state for closed issue {issue_number}.")
else:
- encoding = response.encoding
- logger.debug(f"Using response.encoding '{encoding}' for {url}")
+ logger.debug(f"Webhook: Issue {issue_number} closed, but not found in current active list. No state change needed.")
+
+
+ elif action in ['opened', 'reopened', 'edited', 'assigned', 'unassigned', 'labeled', 'unlabeled', 'milestoned', 'demilestoned']:
+ logger.info(f"Webhook: Adding/Updating issue {issue_number} (action: {action}).")
+ processed_data = self._process_issue_data(issue_data)
+
+ old_issue = self.issues.get(issue_number)
+ # Check for changes that impact AI suggestions or clustering
+ if not old_issue or \
+ old_issue.get('body') != processed_data.get('body') or \
+ old_issue.get('title') != processed_data.get('title') or \
+ set(old_issue.get('labels', [])) != set(processed_data.get('labels', [])):
+ significant_change = True
+ logger.info(f"Significant change detected for issue {issue_number} (content/labels).")
+ # Invalidate ALL precomputed AI state on significant edit
+ self.precomputed_context.pop(issue_number, None)
+ self.precomputed_summaries.pop(issue_number, None)
+ self.precomputed_missing_info.pop(issue_number, None)
+ self.precomputed_analysis.pop(issue_number, None)
+ # Clear the entire suggestion cache on significant change
+ self.cached_suggestion.cache_clear()
+ logger.debug("Cleared suggestion cache due to significant issue change.")
+
+ # Check if state-related fields changed (affecting idle processing lists)
+ # This check is for logging/debugging, the idle loop re-evaluates lists anyway
+ if not old_issue or \
+ old_issue.get('updated_at') != processed_data.get('updated_at') or \
+ old_issue.get('assignee') != processed_data.get('assignee') or \
+ set(old_issue.get('labels', [])) != set(processed_data.get('labels', [])) or \
+ old_issue.get('state') != processed_data.get('state'): # State change (open/reopened)
+ logger.debug(f"State-related change detected for issue {issue_number} (update time, assignee, labels, state). Idle loop will re-evaluate.")
+
+ self.issues[issue_number] = processed_data
+ needs_ui_update = True
+ else:
+ logger.info(f"Ignoring webhook action '{action}' for issue {issue_number} (already filtered).")
+
+ # --- Track changes for idle processing ---
+ if needs_ui_update:
+ self.last_webhook_time = time.time()
+ if significant_change:
+ self._increment_change_counter()
+ # Rebuild the list used for clustering immediately if a significant change occurred
+ # This list is a snapshot used by the async clustering task
+ self.issue_list_for_clustering = list(self.issues.values())
+ logger.info("Issue list for clustering updated due to significant webhook change.")
+ # Broadcast UI update notification
+ # Schedule this in the main loop using call_soon_threadsafe
+ if self.main_loop.is_running():
+ self.main_loop.call_soon_threadsafe(asyncio.create_task, self.broadcast_issue_update())
+ logger.debug("Scheduled issue update broadcast.")
+ else:
+ logger.warning("Main loop not running, cannot broadcast issue update.")
+
+
+ def _increment_change_counter(self):
+ """Increments change counter and sets recluster flag if threshold reached."""
+ self._webhook_change_count += 1
+ logger.debug(f"Significant change detected. Change count: {self._webhook_change_count}/{self.recluster_threshold}")
+ if self._webhook_change_count >= self.recluster_threshold:
+ self.needs_recluster = True
+ logger.info(f"Change threshold ({self.recluster_threshold}) reached. Flagging for re-clustering.")
+
+ def _process_issue_data(self, issue_data: dict) -> dict:
+ """Helper to structure issue data consistently."""
+ return {
+ "id": issue_data.get('number'), # Use .get for safety
+ "title": issue_data.get('title', 'No Title Provided'),
+ "body": issue_data.get('body', ''),
+ "state": issue_data.get('state', 'unknown'),
+ "labels": sorted([label.get('name', '') for label in issue_data.get('labels', []) if isinstance(label, dict) and label.get('name')]), # Ensure labels are dicts and have name
+ "assignee": issue_data.get('assignee', {}).get('login') if issue_data.get('assignee') and isinstance(issue_data.get('assignee'), dict) else None,
+ "url": issue_data.get('html_url', '#'),
+ "created_at": issue_data.get('created_at'),
+ "updated_at": issue_data.get('updated_at'),
+ }
+
+ async def crawl_issues(self, repo_url: str, github_token: Optional[str], hf_token: Optional[str]) -> Tuple[List[List], go.Figure, str, go.Figure]:
+ """
+ Crawls issues, resets state, clones repo, clusters, starts background tasks.
+ Returns dataframe data, stats plot, status message, and analytics plot.
+ """
+ # Strip whitespace from inputs
+ repo_url = repo_url.strip() if repo_url else None
+ github_token = github_token.strip() if github_token else None
+ hf_token = hf_token.strip() if hf_token else None
+
+ # Define a default empty plot for consistent return type
+ def get_empty_plot(title="Plot"):
+ fig = go.Figure()
+ fig.update_layout(title=title, xaxis={"visible": False}, yaxis={"visible": False},
+ annotations=[{"text": "Scan needed.", "xref": "paper", "yref": "paper", "showarrow": False, "font": {"size": 16}}],
+ plot_bgcolor='rgba(0,0,0,0)', paper_bgcolor='rgba(0,0,0,0)')
+ return fig
+
+ if not repo_url or not hf_token:
+ logger.error("Repository URL and Hugging Face Token are required.")
+ empty_fig = get_empty_plot("Issue Severity Distribution")
+ return [], empty_fig, "Error: Repository URL and Hugging Face Token are required.", empty_fig
+
+ logger.info(f"Starting new issue crawl and setup for {repo_url}")
+
+ # --- Reset Manager State ---
+ # Stop background tasks first
+ self.stop_idle_processing()
+ self.stop_broadcast_loop()
+
+ self.issues = {}
+ # Clear code_editors instances
+ self.code_editors = {}
+ self.issue_clusters = {}
+ self.issue_list_for_clustering = []
+ self.cached_suggestion.cache_clear() # Clear AI suggestion cache
+ self.precomputed_context = {}
+ self.precomputed_summaries = {}
+ self.precomputed_missing_info = {}
+ self.precomputed_analysis = {}
+ # self.code_embeddings = {} # Not used
+ self.potential_duplicates = {}
+ self.stale_issues = []
+ self.high_priority_candidates = []
+ self.needs_recluster = False
+ self._webhook_change_count = 0
+ self.last_webhook_time = time.time()
+ self.repo = None # Clear the repo object
+ self.repo_url = repo_url
+ self.github_token = github_token
+ self.hf_token = hf_token
+ logger.info("Internal state reset for new crawl.")
+
+ # --- Repository Cloning/Updating ---
+ match = re.match(r"https?://github\.com/([^/]+)/([^/]+)", self.repo_url)
+ if not match:
+ logger.error(f"Invalid GitHub URL format: {self.repo_url}")
+ empty_fig = get_empty_plot("Issue Severity Distribution")
+ return [], empty_fig, "Error: Invalid GitHub URL format. Use https://github.com/owner/repo", empty_fig
+ self.repo_owner, self.repo_name = match.groups()
+ self.repo_local_path = WORKSPACE / f"{self.repo_owner}_{self.repo_name}"
- # Decode content with fallback
- try:
- raw_content = response.content.decode(encoding, errors='replace')
- except (UnicodeDecodeError, LookupError):
- # Fallback to a more common encoding if the first attempt fails
+ try:
+ if self.repo_local_path.exists():
+ logger.info(f"Attempting to update existing repository clone at {self.repo_local_path}")
try:
- raw_content = response.content.decode('utf-8', errors='replace')
- encoding = 'utf-8 (fallback)'
- logger.warning(f"Decoding with {encoding} fallback for {url}")
- except Exception:
- raw_content = response.content.decode('latin-1', errors='replace') # Another common fallback
- encoding = 'latin-1 (fallback)'
- logger.warning(f"Decoding with {encoding} fallback for {url}")
-
-
- # Extract metadata
- metadata = {
- 'original_url': url,
- 'final_url': final_url,
- 'timestamp': datetime.now().isoformat(),
- 'detected_encoding': encoding,
- 'content_type': response.headers.get('Content-Type', ''),
- 'content_length': len(response.content),
- 'headers': dict(response.headers),
- 'status_code': response.status_code
- }
-
- # Process based on content type
- processed_extraction = self._process_web_content(raw_content, metadata['content_type'], final_url)
+ self.repo = Repo(self.repo_local_path)
+ # Ensure the origin remote matches the requested URL
+ if not self.repo.remotes or 'origin' not in self.repo.remotes:
+ logger.warning(f"Existing repo at {self.repo_local_path} has no 'origin' remote. Re-cloning.")
+ if self.repo_local_path.exists(): shutil.rmtree(self.repo_local_path)
+ self.repo = Repo.clone_from(self.repo_url, self.repo_local_path, progress=lambda op, cur, tot, msg: logger.debug(f"Clone progress: {msg}"))
+ else:
+ origin = self.repo.remotes.origin
+ remote_url = next((u for u in origin.urls), None) # Get first URL
+ expected_urls = {self.repo_url, self.repo_url + ".git"}
+ if remote_url not in expected_urls:
+ logger.warning(f"Existing repo path {self.repo_local_path} has different remote URL ('{remote_url}' vs '{self.repo_url}'). Removing and re-cloning.")
+ # Remove the directory entirely before re-cloning
+ if self.repo_local_path.exists(): shutil.rmtree(self.repo_local_path)
+ self.repo = Repo.clone_from(self.repo_url, self.repo_local_path, progress=lambda op, cur, tot, msg: logger.debug(f"Clone progress: {msg}"))
+ else:
+ logger.info("Pulling latest changes...")
+ # Use a timeout for pull operations
+ try:
+ # Fetch first to get latest refs
+ origin.fetch(progress=lambda op, cur, tot, msg: logger.debug(f"Fetch progress: {msg}"), timeout=120)
+ # Then pull
+ origin.pull(progress=lambda op, cur, tot, msg: logger.debug(f"Pull progress: {msg}"), timeout=120)
+ # Unshallow if necessary
+ if self.repo.git.rev_parse('--is-shallow-repository').strip() == 'true':
+ logger.info("Repository is shallow, unshallowing...")
+ # Use a timeout for unshallow
+ self.repo.git.fetch('--unshallow', timeout=300)
+ except GitCommandError as pull_err:
+ logger.error(f"Git pull/fetch error: {pull_err}. Proceeding with potentially stale local copy.")
+ except Exception as pull_err:
+ logger.exception(f"Unexpected error during git pull/fetch: {pull_err}. Proceeding with potentially stale local copy.")
+
+
+ except (InvalidGitRepositoryError, NoSuchPathError):
+ logger.warning(f"Invalid or missing Git repository at {self.repo_local_path}. Re-cloning.")
+ # Ensure directory is clean before re-cloning
+ if self.repo_local_path.exists(): shutil.rmtree(self.repo_local_path)
+ self.repo = Repo.clone_from(self.repo_url, self.repo_local_path, progress=lambda op, cur, tot, msg: logger.debug(f"Clone progress: {msg}"))
+ except GitCommandError as git_err:
+ logger.error(f"Git operation error during update: {git_err}. Trying to proceed with existing copy, but it might be stale.")
+ if not self.repo: # If repo object wasn't successfully created before the error
+ try: self.repo = Repo(self.repo_local_path)
+ except Exception: logger.error("Failed to even load existing repo after git error.")
+ except Exception as e:
+ logger.exception(f"An unexpected error occurred during repository update check: {e}")
+ # If repo object wasn't successfully created before the error
+ if not self.repo:
+ try: self.repo = Repo(self.repo_local_path)
+ except Exception: logger.error("Failed to even load existing repo after update error.")
- return {
- 'source': 'url',
- 'url': url, # Keep original URL as identifier
- 'raw_content': raw_content,
- 'metadata': metadata,
- 'extracted_data': processed_extraction['data'],
- 'processing_notes': processed_extraction['notes']
- }
- except requests.exceptions.RequestException as e:
- if retry_count < self.max_retries - 1:
- logger.warning(f"Retry {retry_count + 1}/{self.max_retries} for URL: {url}")
- time.sleep(2 ** retry_count) # Exponential backoff
- return self.fetch_content(url, retry_count + 1)
- logger.error(f"Failed to fetch content after {self.max_retries} attempts from {url}: {e}")
- return {
- 'source': 'url',
- 'url': url,
- 'raw_content': None,
- 'metadata': {'original_url': url, 'timestamp': datetime.now().isoformat()},
- 'extracted_data': None,
- 'processing_notes': f"Failed to fetch content: {str(e)}"
- }
- except Exception as e:
- logger.error(f"Unexpected error while fetching or processing URL {url}: {e}")
- return {
- 'source': 'url',
- 'url': url,
- 'raw_content': raw_content if 'raw_content' in locals() else None,
- 'metadata': metadata if 'metadata' in locals() else {'original_url': url, 'timestamp': datetime.now().isoformat()},
- 'extracted_data': None,
- 'processing_notes': f"Unexpected processing error: {str(e)}"
- }
+ else:
+ logger.info(f"Cloning repository {self.repo_url} to {self.repo_local_path}")
+ # Use a timeout for the initial clone
+ self.repo = Repo.clone_from(self.repo_url, self.repo_local_path, progress=lambda op, cur, tot, msg: logger.debug(f"Clone progress: {msg}"), timeout=300)
- def _process_web_content(self, content: str, content_type: str, base_url: str) -> Dict[str, Any]:
- """Process content based on detected content type"""
- lower_content_type = content_type.lower()
- notes = []
- extracted_data: Any = None # Use Any to allow different types
- try:
- if 'text/html' in lower_content_type:
- logger.debug(f"Processing HTML content from {base_url}")
- extracted_data = self._process_html_content_enhanced(content, base_url)
- notes.append("Processed as HTML")
- elif 'application/json' in lower_content_type or 'text/json' in lower_content_type:
- logger.debug(f"Processing JSON content from {base_url}")
- try:
- extracted_data = json.loads(content)
- notes.append("Parsed as JSON")
- except json.JSONDecodeError as e:
- extracted_data = content # Keep raw text if invalid JSON
- notes.append(f"Failed to parse as JSON: {e}")
- logger.warning(f"Failed to parse JSON from {base_url}: {e}")
- except Exception as e:
- extracted_data = content
- notes.append(f"Error processing JSON: {e}")
- logger.error(f"Error processing JSON from {base_url}: {e}")
- elif 'application/xml' in lower_content_type or 'text/xml' in lower_content_type or lower_content_type.endswith('+xml'):
- logger.debug(f"Processing XML content from {base_url}")
- try:
- # Try parsing XML. Convert to a string or a dict representation if needed.
- # For simplicity, we'll convert to a readable string representation of the tree.
- root = ET.fromstring(content)
- # A simple way to represent XML as text
- xml_text = ET.tostring(root, encoding='unicode', method='xml')
- extracted_data = xml_text # Store as string for now
- notes.append("Parsed as XML (text representation)")
- except ET.ParseError as e:
- extracted_data = content
- notes.append(f"Failed to parse as XML: {e}")
- logger.warning(f"Failed to parse XML from {base_url}: {e}")
- except Exception as e:
- extracted_data = content
- notes.append(f"Error processing XML: {e}")
- logger.error(f"Error processing XML from {base_url}: {e}")
- elif 'text/plain' in lower_content_type or 'text/' in lower_content_type: # Catch other text types
- logger.debug(f"Processing Plain Text content from {base_url}")
- extracted_data = content
- notes.append("Processed as Plain Text")
- else:
- logger.debug(f"Unknown content type '{content_type}' from {base_url}. Storing raw content.")
- extracted_data = content # Store raw content for unknown types
- notes.append(f"Unknown content type '{content_type}'. Stored raw text.")
+ logger.info("Repository clone/update process finished.")
+ if not self.repo:
+ raise Exception("Repository object could not be initialized after cloning/update.")
+ except GitCommandError as e:
+ logger.error(f"Failed to clone/update repository: {e}")
+ empty_fig = get_empty_plot("Issue Severity Distribution")
+ empty_fig.update_layout(annotations=[{"text": "Repo Error.", "xref": "paper", "yref": "paper", "showarrow": False, "font": {"size": 16}}])
+ return [], empty_fig, f"Error cloning/updating repository: {e}. Check URL, permissions, and network.", empty_fig
except Exception as e:
- logger.error(f"Unexpected error in _process_web_content for {base_url} ({content_type}): {e}")
- extracted_data = content # Fallback to raw content on error
- notes.append(f"Unexpected processing error: {e}. Stored raw text.")
+ logger.exception(f"An unexpected error occurred during repository handling: {e}")
+ empty_fig = get_empty_plot("Issue Severity Distribution")
+ empty_fig.update_layout(annotations=[{"text": "Repo Error.", "xref": "paper", "yref": "paper", "showarrow": False, "font": {"size": 16}}])
+ return [], empty_fig, f"An unexpected error occurred during repo setup: {e}", empty_fig
- return {'data': extracted_data, 'notes': notes}
+ # --- Issue Fetching ---
+ api_url = f"{GITHUB_API}/{self.repo_owner}/{self.repo_name}/issues?state=open&per_page=100"
+ headers = {"Accept": "application/vnd.github.v3+json"}
+ if github_token:
+ headers["Authorization"] = f"token {github_token}"
- def _process_html_content_enhanced(self, content: str, base_url: str) -> Dict[str, Any]:
- """Process HTML content, preserving text, and extracting metadata."""
- extracted: Dict[str, Any] = {
- 'title': None,
- 'meta_description': None, # Add extraction for meta description
- 'full_text': "",
- 'links': [] # Add extraction for links
- }
try:
- soup = BeautifulSoup(content, 'html.parser')
+ all_issues_data = []
+ page = 1
+ logger.info(f"Fetching open issues from GitHub API (repo: {self.repo_owner}/{self.repo_name})...")
+ async with aiohttp.ClientSession(headers=headers) as session:
+ while True:
+ paginated_url = api_url
+ logger.debug(f"Fetching URL: {paginated_url}")
+ # Use a timeout for API requests
+ try:
+ async with session.get(paginated_url, timeout=30) as response:
+ rate_limit_remaining = response.headers.get('X-RateLimit-Remaining')
+ rate_limit_reset = response.headers.get('X-RateLimit-Reset')
+ logger.debug(f"GitHub API Response Status: {response.status}, RateLimit Remaining: {rate_limit_remaining}, Reset: {rate_limit_reset}")
+
+ if response.status == 403 and rate_limit_remaining == '0':
+ reset_time = int(rate_limit_reset) if rate_limit_reset else time.time() + 60
+ wait_time = max(reset_time - time.time() + 5, 0) # Wait until reset time + a buffer
+ logger.warning(f"GitHub API rate limit exceeded. Waiting until {datetime.fromtimestamp(reset_time).strftime('%H:%M:%S')} ({wait_time:.0f}s)")
+ await asyncio.sleep(wait_time)
+ continue # Retry the same page
+
+ response.raise_for_status() # Raise for other 4xx/5xx errors
+
+ issues_page_data = await response.json()
+ if not issues_page_data: break # No more issues on this page
+
+ logger.info(f"Fetched page {page} with {len(issues_page_data)} items.")
+ all_issues_data.extend(issues_page_data)
+
+ link_header = response.headers.get('Link')
+ if link_header and 'rel="next"' in link_header:
+ # Simple parsing for the next link, more robust parsing might be needed for complex headers
+ next_url_match = re.search(r'<([^>]+)>;\s*rel="next"', link_header)
+ if next_url_match:
+ # The next URL is provided directly, use it
+ api_url = next_url_match.group(1)
+ page += 1 # Increment page counter for logging, though not strictly needed for the loop logic now
+ logger.debug(f"Found next page link: {api_url}")
+ else:
+ logger.warning("Link header contains 'rel=\"next\"' but could not parse the URL. Stopping pagination.")
+ break
+ else:
+ logger.debug("No 'next' link found in Link header. Assuming last page.")
+ break
+ await asyncio.sleep(0.1) # Small delay between requests
+
+ except asyncio.TimeoutError:
+ logger.warning(f"GitHub API request timed out for page {page}. Stopping pagination early.")
+ break # Stop pagination on timeout
+ except aiohttp.ClientResponseError as e:
+ # Re-raise client response errors so the outer handler catches them
+ raise e
+ except Exception as e:
+ logger.exception(f"An unexpected error occurred during GitHub API pagination for page {page}. Stopping pagination.")
+ break # Stop pagination on unexpected error
- # Extract Title
- if soup.title and soup.title.string:
- extracted['title'] = soup.title.string.strip()
- # Extract Meta Description
- meta_desc = soup.find('meta', attrs={'name': 'description'})
- if meta_desc and meta_desc.get('content'):
- extracted['meta_description'] = meta_desc['content'].strip()
+ logger.info(f"Total items fetched (including potential PRs): {len(all_issues_data)}")
+ # Filter out pull requests (issues with 'pull_request' key)
+ self.issues = {
+ issue_data['number']: self._process_issue_data(issue_data)
+ for issue_data in all_issues_data
+ if 'pull_request' not in issue_data and issue_data.get('number') is not None # Ensure number exists
+ }
- # Extract and process links (convert relative to absolute)
- for a_tag in soup.find_all('a', href=True):
- href = a_tag['href']
- text = a_tag.get_text().strip()
- try:
- absolute_url = urljoin(base_url, href)
- extracted['links'].append({'text': text, 'url': absolute_url})
- except Exception:
- extracted['links'].append({'text': text, 'url': href}) # Keep relative if join fails
+ logger.info(f"Filtered out pull requests, {len(self.issues)} actual open issues remaining.")
+
+ empty_fig = get_empty_plot("Issue Severity Distribution")
+ if not self.issues:
+ logger.warning("No open issues found for this repository.")
+ empty_fig.update_layout(annotations=[{"text": "No issues found.", "xref": "paper", "yref": "paper", "showarrow": False, "font": {"size": 16}}])
+ return [], empty_fig, "No open issues found in the repository.", empty_fig
+
+ # --- Clustering and UI Data Prep ---
+ self.issue_list_for_clustering = list(self.issues.values())
+ logger.info("Clustering issues...")
+ await self._cluster_similar_issues()
+
+ # --- Initial Idle Task Prep (Run synchronously after load) ---
+ logger.info("Identifying potential duplicates based on initial clusters...")
+ self._identify_potential_duplicates()
+ logger.info("Identifying potentially stale issues...")
+ self._identify_stale_issues()
+ logger.info("Identifying high priority candidates...")
+ self._identify_high_priority_candidates()
+
+ # --- Prepare Dataframe Output & Stats ---
+ dataframe_data = []
+ severity_counts = {"Critical": 0, "High": 0, "Medium": 0, "Low": 0, "Unknown": 0}
+ index_to_cluster_id: Dict[int, int] = {}
+ # Map issue index in the clustering list back to its cluster ID
+ for cluster_id, indices in self.issue_clusters.items():
+ for index in indices:
+ if 0 <= index < len(self.issue_list_for_clustering):
+ index_to_cluster_id[index] = cluster_id
+ else:
+ logger.warning(f"Clustering returned invalid index {index} for list of length {len(self.issue_list_for_clustering)}")
+
+ for i, issue in enumerate(self.issue_list_for_clustering):
+ severity = self._determine_severity(issue.get('labels', [])) # Use get for safety
+ severity_counts[severity] += 1
+ # Get cluster ID using the index from the clustering list
+ cluster_id = index_to_cluster_id.get(i, -1)
+ dataframe_data.append([
+ issue.get('id', 'N/A'), # Use get for safety
+ issue.get('title', 'No Title'), # Use get for safety
+ severity,
+ cluster_id if cluster_id != -1 else "N/A" # Display "N/A" for noise (-1)
+ ])
+
+ logger.info("Generating statistics plot...")
+ stats_fig = self._generate_stats_plot(severity_counts)
+
+ # --- Start Background Tasks ---
+ # Ensure tasks are created in the manager's loop
+ self.start_broadcast_loop()
+ self.start_idle_processing()
+
+ success_msg = f"Found {len(self.issues)} open issues. Clustered into {len(self.issue_clusters)} groups. Repo ready. Background analysis started."
+ logger.info(success_msg)
+ # Return both plots (stats and analytics severity are the same initially)
+ return dataframe_data, stats_fig, success_msg, stats_fig
+
+ except aiohttp.ClientResponseError as e:
+ logger.error(f"GitHub API request failed: Status={e.status}, Message='{e.message}', URL='{e.request_info.url}'")
+ error_msg = f"Error fetching issues: {e.status} - {e.message}. Check token/URL."
+ if e.status == 404: error_msg = f"Error: Repository not found at {self.repo_url}."
+ elif e.status == 401: error_msg = "Error: Invalid GitHub token or insufficient permissions for this repository."
+ elif e.status == 403:
+ rate_limit_remaining = e.headers.get('X-RateLimit-Remaining') # FIX: Access rate_limit_remaining from error headers
+ rate_limit_reset = e.headers.get('X-RateLimit-Reset')
+ reset_time_str = "unknown"
+ if rate_limit_reset:
+ try: reset_time_str = datetime.fromtimestamp(int(rate_limit_reset), timezone.utc).strftime('%Y-%m-%d %H:%M:%S %Z')
+ except ValueError: pass
+ error_msg = f"Error: GitHub API rate limit likely exceeded or access forbidden (Remaining: {rate_limit_remaining}). Reset time: {reset_time_str}. Check token or wait."
+ self.stop_idle_processing()
+ self.stop_broadcast_loop()
+ empty_fig = get_empty_plot("Issue Severity Distribution")
+ empty_fig.update_layout(annotations=[{"text": "API Error.", "xref": "paper", "yref": "paper", "showarrow": False, "font": {"size": 16}}])
+ return [], empty_fig, error_msg, empty_fig
+ except asyncio.TimeoutError:
+ logger.error("GitHub API request timed out.")
+ self.stop_idle_processing()
+ self.stop_broadcast_loop()
+ empty_fig = get_empty_plot("Issue Severity Distribution")
+ empty_fig.update_layout(annotations=[{"text": "API Timeout.", "xref": "paper", "yref": "paper", "showarrow": False, "font": {"size": 16}}])
+ return [], empty_fig, "Error: GitHub API request timed out.", empty_fig
+ except Exception as e:
+ self.stop_idle_processing()
+ self.stop_broadcast_loop()
+ logger.exception(f"An unexpected error occurred during issue crawl: {e}")
+ empty_fig = get_empty_plot("Issue Severity Distribution")
+ empty_fig.update_layout(annotations=[{"text": "Unexpected Error.", "xref": "paper", "yref": "paper", "showarrow": False, "font": {"size": 16}}])
+ return [], empty_fig, f"An unexpected error occurred: {e}", empty_fig
+
+ def _determine_severity(self, labels: List[str]) -> str:
+ """Determines issue severity based on labels using predefined rules."""
+ labels_lower = {label.lower().strip() for label in labels}
+ for severity, keywords in self.severity_rules.items():
+ if any(keyword in label for label in labels_lower for keyword in keywords):
+ return severity
+ return "Unknown"
+
+ def _generate_stats_plot(self, severity_counts: Dict[str, int]) -> go.Figure:
+ """Generates a Plotly bar chart for issue severity distribution."""
+ filtered_counts = {k: v for k, v in severity_counts.items() if v > 0}
+ if not filtered_counts:
+ fig = go.Figure()
+ fig.update_layout(title="Issue Severity Distribution", xaxis={"visible": False}, yaxis={"visible": False},
+ # FIX: Corrected dictionary syntax for font size
+ annotations=[{"text": "No issues to display.", "xref": "paper", "yref": "paper", "showarrow": False, "font": {"size": 16}}],
+ plot_bgcolor='rgba(0,0,0,0)', paper_bgcolor='rgba(0,0,0,0)')
+ return fig # Return the empty figure here
+
+ severities = list(filtered_counts.keys())
+ counts = list(filtered_counts.values())
+ order = ['Critical', 'High', 'Medium', 'Low', 'Unknown']
+ # Sort severities based on the predefined order
+ severities_sorted = sorted(severities, key=lambda x: order.index(x) if x in order else len(order))
+ counts_sorted = [filtered_counts[s] for s in severities_sorted]
+
+ fig = px.bar(x=severities_sorted, y=counts_sorted, title="Issue Severity Distribution",
+ labels={'x': 'Severity', 'y': 'Number of Issues'}, color=severities_sorted,
+ color_discrete_map={'Critical': '#DC2626', 'High': '#F97316', 'Medium': '#FACC15', 'Low': '#84CC16', 'Unknown': '#6B7280'},
+ text=counts_sorted)
+ fig.update_layout(xaxis_title=None, yaxis_title="Number of Issues", plot_bgcolor='rgba(0,0,0,0)',
+ paper_bgcolor='rgba(0,0,0,0)', showlegend=False,
+ xaxis={'categoryorder':'array', 'categoryarray': order},
+ yaxis={'rangemode': 'tozero'}) # Ensure y-axis starts at 0
+ fig.update_traces(textposition='outside')
+ return fig
+
+ async def _cluster_similar_issues(self):
+ """Generates embeddings and clusters issues using HDBSCAN. Uses self.issue_list_for_clustering."""
+ if not self.issue_list_for_clustering:
+ logger.warning("Cannot cluster issues: No issues loaded or list is empty.")
+ self.issue_clusters = {}
+ self._webhook_change_count = 0 # Reset on empty list
+ self.needs_recluster = False
+ return
+ if not self.hf_token or HF_EMBEDDING_MODEL is None:
+ logger.error("Cannot cluster issues: Hugging Face token or embedding model missing.")
+ self.issue_clusters = {}
+ self._webhook_change_count = 0 # Reset on missing token/model
+ self.needs_recluster = False
+ return
+
+ num_issues = len(self.issue_list_for_clustering)
+ logger.info(f"Generating embeddings for {num_issues} issues for clustering using {HF_EMBEDDING_MODEL}...")
+ try:
+ # Use title + a snippet of the body for embedding
+ texts_to_embed = [
+ f"Title: {i.get('title','')} Body: {i.get('body','')[:500]}" # Limit body length
+ for i in self.issue_list_for_clustering
+ ]
+ embeddings = await self._generate_embeddings(texts_to_embed)
+
+ if embeddings is None or not isinstance(embeddings, list) or len(embeddings) != num_issues:
+ logger.error(f"Failed to generate valid embeddings for clustering. Expected {num_issues}, got {type(embeddings)} len {len(embeddings) if embeddings else 'N/A'}.")
+ self.issue_clusters = {}
+ self._webhook_change_count = 0 # Reset on embedding failure
+ self.needs_recluster = False
+ return
+
+ logger.info(f"Generated {len(embeddings)} embeddings. Running HDBSCAN clustering...")
+ # Adjust min_cluster_size dynamically based on issue count?
+ min_cluster_size = max(2, min(5, num_issues // 10)) # Example: min 2, max 5, or 10% of issues
+ clusterer = HDBSCAN(min_cluster_size=min_cluster_size, metric='cosine', allow_single_cluster=True, gen_min_span_tree=True)
+ clusters = clusterer.fit_predict(embeddings)
+
+ new_issue_clusters: Dict[int, List[int]] = {}
+ noise_count = 0
+ for i, cluster_id in enumerate(clusters):
+ cluster_id_int = int(cluster_id)
+ if cluster_id_int == -1:
+ noise_count += 1
+ continue
+ if cluster_id_int not in new_issue_clusters:
+ new_issue_clusters[cluster_id_int] = []
+ new_issue_clusters[cluster_id_int].append(i)
+
+ self.issue_clusters = new_issue_clusters
+ logger.info(f"Clustering complete. Found {len(self.issue_clusters)} clusters (min size {min_cluster_size}) with {noise_count} noise points.")
+
+ # Reset the change counter and flag after successful clustering
+ self._webhook_change_count = 0
+ self.needs_recluster = False
+ logger.debug("Reset webhook change counter and recluster flag after clustering.")
+ except Exception as e:
+ logger.exception(f"Error during issue clustering: {e}")
+ self.issue_clusters = {}
+ self._webhook_change_count = 0 # Reset on clustering failure
+ self.needs_recluster = False
- # Extract all text content (similar to stripped_strings but ensures order)
- text_parts = []
- # Use a more robust way to get visible text, including handling script/style tags
- for script_or_style in soup(["script", "style"]):
- script_or_style.extract() # Remove script and style tags
- text = soup.get_text(separator='\n') # Get text with newlines
- # Clean up whitespace and empty lines
- lines = text.splitlines()
- cleaned_lines = [line.strip() for line in lines if line.strip()]
- extracted['full_text'] = '\n'.join(cleaned_lines)
+ def _identify_potential_duplicates(self):
+ """Populates self.potential_duplicates based on self.issue_clusters and self.issue_list_for_clustering."""
+ self.potential_duplicates = {}
+ if not self.issue_clusters or not self.issue_list_for_clustering:
+ logger.debug("Skipping duplicate identification: No clusters or issue list.")
+ return
+ index_to_id = {}
+ try:
+ for i, issue in enumerate(self.issue_list_for_clustering):
+ issue_id = issue.get('id')
+ if issue_id is None:
+ logger.warning(f"Issue at index {i} in clustering list is missing an ID.")
+ continue
+ index_to_id[i] = issue_id
except Exception as e:
- logger.error(f"Enhanced HTML processing error for {base_url}: {e}")
- extracted['full_text'] = content # Fallback to raw content
- extracted['processing_error'] = f"Enhanced HTML processing failed: {e}"
-
- return extracted
-
-class EnhancedFileProcessor:
- """Advanced file processing with enhanced content extraction"""
- def __init__(self, max_file_size: int = 5 * 1024 * 1024 * 1024): # 5GB default
- self.max_file_size = max_file_size
- # Expanded supported extensions to include common docs and structured formats
- self.supported_extensions = {
- '.txt', '.md', '.csv', '.json', '.xml', '.html', '.htm',
- '.log', '.yml', '.yaml', '.ini', '.conf', '.cfg',
- '.pdf', '.doc', '.docx', '.rtf', '.odt',
- # Archives are handled separately but listed for context
- '.zip', '.tar', '.gz', '.bz2', '.7z', '.rar',
- }
- self.archive_extensions = {'.zip', '.tar', '.gz', '.bz2', '.7z', '.rar'}
+ logger.error(f"Error creating index-to-ID map for duplicate check: {e}. Issue list might be inconsistent.")
+ return
+
+ for cluster_id, indices in self.issue_clusters.items():
+ if len(indices) > 1:
+ # Get issue IDs for indices in this cluster, skipping any invalid indices
+ cluster_issue_ids = [index_to_id[i] for i in indices if i in index_to_id]
+ if len(cluster_issue_ids) > 1: # Ensure there's more than one valid issue ID in the cluster
+ for issue_id in cluster_issue_ids:
+ # For each issue in the cluster, list all *other* issues in the same cluster as potential duplicates
+ self.potential_duplicates[issue_id] = [other_id for other_id in cluster_issue_ids if other_id != issue_id]
+
+ logger.info(f"Identified potential duplicates for {len(self.potential_duplicates)} issues based on clustering.")
+
+ async def _generate_embeddings(self, texts: List[str]):
+ """Generates sentence embeddings using Hugging Face Inference API."""
+ if not self.hf_token:
+ logger.error("Hugging Face token is not set. Cannot generate embeddings.")
+ return None
+ if not texts:
+ logger.warning("Embedding generation requested with empty text list.")
+ return []
+ if HF_EMBEDDING_MODEL is None:
+ logger.error("HF Embedding model is not configured.")
+ return None
+
+ model_id = HF_EMBEDDING_MODEL # Use the fixed embedding model
+ api_url = f"{HF_INFERENCE_API}/{model_id}"
+ headers = {"Authorization": f"Bearer {self.hf_token}"}
+ timeout = aiohttp.ClientTimeout(total=180) # Increased timeout for embedding large batches
+
+ logger.info(f"Requesting embeddings from {api_url} for {len(texts)} texts.")
+ # HF Inference API has a limit on the number of inputs per request (often 512 or 1024)
+ # Batching is recommended for large lists of texts.
+ batch_size = 500 # Example batch size, adjust based on model limits if known
+ all_embeddings = []
+
+ for i in range(0, len(texts), batch_size):
+ batch_texts = texts[i:i + batch_size]
+ payload = {"inputs": batch_texts, "options": {"wait_for_model": True}}
+ logger.debug(f"Processing embedding batch {i//batch_size + 1}/{(len(texts)-1)//batch_size + 1} ({len(batch_texts)} texts)")
+
+ # Implement retry logic for batches
+ retries = 3
+ for attempt in range(retries):
+ try:
+ async with aiohttp.ClientSession(headers=headers, timeout=timeout) as session:
+ async with session.post(api_url, json=payload) as response:
+ rate_limit_remaining = response.headers.get('X-Ratelimit-Remaining')
+ logger.debug(f"HF Embedding API Response Status: {response.status}, RateLimit Remaining: {rate_limit_remaining}")
+
+ if response.status == 429: # Too Many Requests
+ retry_after = int(response.headers.get('Retry-After', 10))
+ logger.warning(f"HF Embedding API rate limited. Waiting for {retry_after} seconds before retry {attempt + 1}/{retries}.")
+ await asyncio.sleep(retry_after)
+ continue # Retry the same batch
+ response.raise_for_status() # Raise for other 4xx/5xx errors
+
+ result = await response.json()
+
+ if isinstance(result, list) and all(isinstance(emb, list) and all(isinstance(f, float) for f in emb) for emb in result):
+ if len(result) == len(batch_texts):
+ all_embeddings.extend(result)
+ logger.debug(f"Successfully received {len(result)} embeddings for batch.")
+ break # Batch successful, move to next batch
+ else:
+ logger.error(f"HF Embedding API returned wrong number of embeddings for batch: Got {len(result)}, expected {len(batch_texts)}.")
+ return None # Indicate failure
+ elif isinstance(result, dict) and 'error' in result:
+ error_msg = result['error']
+ estimated_time = result.get('estimated_time')
+ logger.error(f"HF Inference API embedding error on batch: {error_msg}" + (f" (Estimated time: {estimated_time}s)" if estimated_time else ""))
+ return None # Indicate failure
+ else:
+ logger.error(f"Unexpected embedding format received on batch: Type={type(result)}. Response: {str(result)[:500]}")
+ return None # Indicate failure
+ except asyncio.TimeoutError:
+ logger.warning(f"HF Inference API embedding request timed out after {timeout.total} seconds for batch. Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1:
+ await asyncio.sleep(5) # Wait a bit before retrying timeout
+ continue
+ else:
+ logger.error("Max retries reached for embedding batch timeout.")
+ return None # Indicate failure
+ except aiohttp.ClientResponseError as e:
+ error_body = await e.text()
+ logger.error(f"HF Inference API embedding request failed on batch: Status={e.status}, Message='{e.message}'. Body: {error_body[:500]}. Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1 and e.status in [500, 502, 503, 504]: # Retry on server errors
+ await asyncio.sleep(5)
+ continue
+ else:
+ logger.error("Max retries reached or non-retryable error for embedding batch.")
+ return None # Indicate failure
+ except Exception as e:
+ logger.exception(f"Unexpected error during embedding generation on batch: {e}. Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1:
+ await asyncio.sleep(5)
+ continue
+ else:
+ logger.error("Max retries reached for unexpected error during embedding batch.")
+ return None # Indicate failure
+ else: # This else block executes if the inner loop completes without a 'break' (i.e., all retries failed)
+ logger.error(f"Failed to process embedding batch after {retries} retries.")
+ return None # Indicate failure
- def process_file(self, file) -> List[Dict]:
- """Process uploaded file with enhanced error handling and complete extraction"""
- if not file or not hasattr(file, 'name'):
- logger.warning("Received invalid file object.")
- return []
+ await asyncio.sleep(0.1) # Small delay between batches
- dataset = []
- file_path = Path(file.name) # Use Path object for easier handling
+ if len(all_embeddings) == len(texts):
+ logger.info(f"Successfully generated embeddings for all {len(all_embeddings)} texts.")
+ return all_embeddings
+ else:
+ logger.error(f"Embedding generation failed partway through. Expected {len(texts)}, got {len(all_embeddings)}.")
+ return None # Indicate overall failure
+
+
+ async def generate_code_patch(self, issue_number: int, model_key: str) -> dict:
+ """Generates a code patch suggestion using a selected AI model."""
+ if issue_number not in self.issues:
+ return {"error": f"Issue {issue_number} not found."}
+ if not self.hf_token:
+ return {"error": "Hugging Face token not set."}
+ if model_key not in HF_MODELS or HF_MODELS.get(model_key) is None:
+ return {"error": f"Invalid or unconfigured model key: {model_key}"}
+ if not self.repo_local_path or not self.repo:
+ return {"error": "Repository not cloned/available locally. Please scan the repository first."}
+
+ issue = self.issues[issue_number]
+ model_id = HF_MODELS[model_key]
+ logger.info(f"Generating patch for issue {issue_number} ('{issue.get('title', 'N/A')[:50]}...') using model {model_id}")
+
+ # --- Context Gathering ---
+ context_str = "Context gathering failed or not available."
+ context_source = "Error"
+ start_time_context = time.time()
+ context_data = self.precomputed_context.get(issue_number) # Use .get for safety
+
+ if context_data:
+ timestamp = context_data.get('timestamp', 0)
+ timestamp_str = datetime.fromtimestamp(timestamp).strftime('%Y-%m-%d %H:%M:%S')
+ if context_data.get("error"):
+ context_str = f"Pre-computed context retrieval failed: {context_data['error']}"
+ context_source = f"Pre-computed (Failed @ {timestamp_str})"
+ elif context_data.get("content"):
+ context_str = context_data["content"]
+ num_files = len(context_data.get('files',[]))
+ context_source = f"Pre-computed ({num_files} files @ {timestamp_str})"
+ else:
+ context_str = "Pre-computed context was empty or unavailable."
+ context_source = f"Pre-computed (Empty @ {timestamp_str})"
+ logger.info(f"Using pre-computed context for issue {issue_number} (Source: {context_source})")
+ else:
+ logger.info(f"No pre-computed context found for issue {issue_number}, computing now.")
+ context_source = "Computed On-Demand"
+ # Compute context on demand and store it
+ context_result = await self._get_code_context(issue)
+ self.precomputed_context[issue_number] = {
+ "content": context_result.get("content"),
+ "files": context_result.get("files", []),
+ "error": context_result.get("error"),
+ "timestamp": time.time()
+ }
+ if "error" in context_result and context_result["error"]:
+ context_str = f"Error retrieving context: {context_result['error']}"
+ context_source += " (Error)"
+ else:
+ context_str = context_result.get("content", "No specific context found.")
+ context_source += f" ({len(context_result.get('files',[]))} files)"
+
+ context_load_duration = time.time() - start_time_context
+ logger.info(f"Computed context on-demand in {context_load_duration:.2f}s. Source: {context_source}")
+
+ # --- Get Pre-computed Info ---
+ summary_text = self._get_precomputed_text(issue_number, self.precomputed_summaries, "summary", "Summary")
+ missing_info_text = self._get_precomputed_text(issue_number, self.precomputed_missing_info, "info_needed", "Missing Info Analysis")
+ analysis_text = self._get_precomputed_text(issue_number, self.precomputed_analysis, "hypothesis", "Preliminary Analysis")
+ duplicate_info = self._get_duplicate_info_text(issue_number)
+
+ # --- Enhanced Prompt ---
+ # Added clear delimiters and instructions for the AI
+ prompt = f"""You are an expert software engineer AI assistant generating a minimal `diff` code patch to fix a GitHub issue.
+
+
+Issue ID: {issue.get('id', 'N/A')}
+Title: {issue.get('title', 'N/A')}
+Labels: {', '.join(issue.get('labels', []))}
+Body:
+{issue.get('body', 'N/A')}
+
+
+
+Summary: {summary_text}
+Missing Information Analysis: {missing_info_text}
+Preliminary Hypothesis: {analysis_text}
+{duplicate_info}
+
+
+
+{context_str}
+
+
+
+1. **Analyze:** Carefully review all provided information (issue details, AI analysis, code context).
+2. **Identify Changes:** Determine the minimal code modifications *strictly within the provided * to address the issue.
+3. **Format:** Generate a standard `diff` patch (--- a/..., +++ b/..., @@ ..., +/-, space for context). Use exact relative paths from the `` section. If the context is insufficient to generate a meaningful patch, state this clearly in the explanation.
+4. **Output Structure:** Provide a concise explanation *before* the ```diff ... ``` block. The explanation should summarize the proposed fix.
+5. **Constraints:**
+ * If the provided `` is insufficient or irrelevant, state "Insufficient context to generate patch." in your explanation and **do not generate a ```diff``` block.** Explain *why* the context is insufficient.
+ * Do not invent code, file paths, or line numbers not present or implied by the context.
+ * Generate *only* the explanation and the diff block (if applicable). Do not include any other conversational text or markdown sections.
+
+
+Patch Suggestion:
+"""
+ # --- Call Inference API ---
+ api_url = f"{HF_INFERENCE_API}/{model_id}"
+ headers = {"Authorization": f"Bearer {self.hf_token}"}
+ # Increased max_new_tokens for potentially larger patches/explanations
+ payload = {
+ "inputs": prompt,
+ "parameters": {"max_new_tokens": 3072, "temperature": 0.1, "return_full_text": False, "do_sample": False},
+ "options": {"wait_for_model": True}
+ }
+ timeout = aiohttp.ClientTimeout(total=240) # Increased timeout
- try:
- file_size = file_path.stat().st_size
- if file_size > self.max_file_size:
- logger.warning(f"File '{file_path.name}' size ({file_size} bytes) exceeds maximum allowed size ({self.max_file_size} bytes).")
- return [{
- 'source': 'file',
- 'filename': file_path.name,
- 'file_size': file_size,
- 'extracted_data': None,
- 'processing_notes': 'File size exceeds limit.'
- }]
-
- with tempfile.TemporaryDirectory() as temp_dir:
- temp_dir_path = Path(temp_dir)
-
- # Decide processing strategy
- if file_path.suffix.lower() in self.archive_extensions:
- dataset.extend(self._process_archive(file_path, temp_dir_path))
- elif file_path.suffix.lower() in self.supported_extensions:
- # Pass the path to the single file processor
- dataset.extend(self._process_single_file(file_path))
+ retries = 3
+ for attempt in range(retries):
+ try:
+ start_time_api = time.time()
+ async with aiohttp.ClientSession(headers=headers, timeout=timeout) as session:
+ async with session.post(api_url, json=payload) as response:
+ api_duration = time.time() - start_time_api
+ rate_limit_remaining = response.headers.get('X-Ratelimit-Remaining')
+ logger.debug(f"HF Patch API Response Status: {response.status}, Duration: {api_duration:.2f}s, RateLimit Remaining: {rate_limit_remaining}")
+
+ if response.status == 429: # Too Many Requests
+ retry_after = int(response.headers.get('Retry-After', 10))
+ logger.warning(f"HF Patch API rate limited. Waiting for {retry_after} seconds before retry {attempt + 1}/{retries}.")
+ await asyncio.sleep(retry_after)
+ continue # Retry the same request
+ response.raise_for_status() # Raise for other 4xx/5xx errors
+
+ result = await response.json()
+
+ if result and isinstance(result, list) and 'generated_text' in result[0]:
+ generated_text = result[0].get('generated_text', '').strip()
+ logger.info(f"Received patch suggestion from {model_id} ({len(generated_text)} chars).")
+
+ # Robustly find the diff block
+ diff_match = re.search(r"```diff\n(.*?)```", generated_text, re.DOTALL | re.IGNORECASE)
+
+ explanation = generated_text # Default explanation is the whole text
+ patch_content = None
+
+ if diff_match:
+ patch_content = diff_match.group(1).strip()
+ # Split text before and after the first diff block
+ parts = generated_text.split("```diff", 1)
+ explanation = parts[0].strip()
+ # Check if the patch content looks like a diff (contains ---, +++, @@)
+ if not re.search(r'^(--- |\+\+\+ |@@ )', patch_content, re.MULTILINE):
+ logger.warning(f"Generated text contains ```diff``` block but content doesn't look like a standard diff for issue {issue_number}. Treating as no valid patch.")
+ # Treat as no valid patch generated
+ patch_content = None
+ explanation = generated_text # Fallback to entire text as explanation
+
+ if not explanation: explanation = "(No explanation provided by AI.)"
+
+ if patch_content:
+ logger.info(f"Successfully parsed diff block for issue {issue_number}.")
+ return {"explanation": explanation, "patch": patch_content, "model_used": model_id, "status": "Patch Generated"}
+ else:
+ # Check if the explanation indicates insufficient context or inability to generate
+ if re.search(r"(insufficient context|cannot generate|unable to create patch|context required)", explanation, re.IGNORECASE):
+ logger.warning(f"AI indicated insufficient context for issue {issue_number}.")
+ return {"explanation": explanation, "patch": None, "model_used": model_id, "status": "Insufficient Context"}
+ else:
+ logger.warning(f"No valid diff block found in patch suggestion response for issue {issue_number}.")
+ return {"explanation": f"(AI response did not contain a valid ```diff block. Full response below)\n---\n{generated_text}", "patch": None, "model_used": model_id, "status": "No Diff Generated"}
+
+ elif isinstance(result, dict) and 'error' in result:
+ error_msg = result['error']
+ estimated_time = result.get('estimated_time')
+ logger.error(f"HF Inference API patch error for issue {issue_number}: {error_msg}" + (f" (Est: {estimated_time}s)" if estimated_time else "") + f". Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1:
+ await asyncio.sleep(5) # Wait before retrying API error
+ continue
+ else:
+ logger.error("Max retries reached for API error during patch generation.")
+ return {"error": f"AI model error after retries: {error_msg}"}
+ else:
+ logger.error(f"Unexpected patch response format from {model_id} for issue {issue_number}: {str(result)[:500]}. Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1:
+ await asyncio.sleep(5)
+ continue
+ else:
+ logger.error("Max retries reached for unexpected API response during patch generation.")
+ return {"error": "Unexpected response format from AI model after retries."}
+
+ except asyncio.TimeoutError:
+ logger.warning(f"HF Inference API patch request timed out ({timeout.total}s) for issue {issue_number}. Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1:
+ await asyncio.sleep(5) # Wait before retrying timeout
+ continue
else:
- logger.warning(f"Unsupported file type for processing: '{file_path.name}'")
- # Optionally process as raw text even if extension is unsupported
- try:
- # Read as text with error replacement
- content_bytes = file_path.read_bytes()
- encoding_detection = chardet.detect(content_bytes)
- encoding = encoding_detection['encoding'] or 'utf-8'
- raw_content = content_bytes.decode(encoding, errors='replace')
- dataset.append({
- 'source': 'file',
- 'filename': file_path.name,
- 'file_size': file_size,
- 'mime_type': mimetypes.guess_type(file_path.name)[0] or 'unknown/unknown',
- 'extracted_data': {'plain_text': raw_content}, # Store raw text under a key
- 'processing_notes': 'Processed as plain text (unsupported extension).'
- })
- except Exception as e:
- logger.error(f"Error reading or processing unsupported file '{file_path.name}' as text: {e}")
- dataset.append({
- 'source': 'file',
- 'filename': file_path.name,
- 'file_size': file_size,
- 'mime_type': mimetypes.guess_type(file_path.name)[0] or 'unknown/unknown',
- 'extracted_data': None,
- 'processing_notes': f'Unsupported file type and failed to read as text: {e}'
- })
+ logger.error("Max retries reached for patch generation timeout.")
+ return {"error": "AI model request timed out after retries."}
+ except aiohttp.ClientResponseError as e:
+ error_body = await e.response.text()
+ logger.error(f"HF Inference API patch request failed for issue {issue_number}: Status={e.status}, Message='{e.message}'. Body: {error_body[:500]}. Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1 and e.status in [500, 502, 503, 504]: # Retry on server errors
+ await asyncio.sleep(5)
+ continue
+ else:
+ logger.error("Max retries reached or non-retryable error for patch generation.")
+ return {"error": f"AI model request failed ({e.status}) after retries. Details: {error_body[:200]}..."}
+ except Exception as e:
+ logger.exception(f"Unexpected error during code patch generation for issue {issue_number}: {e}. Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1:
+ await asyncio.sleep(5)
+ continue
+ else:
+ logger.error("Max retries reached for unexpected error during patch generation.")
+ return {"error": f"An unexpected error occurred during patch generation after retries: {e}"}
+
+ # This part is reached if all retries fail
+ logger.error(f"Patch generation failed for issue {issue_number} after {retries} retries.")
+ return {"error": f"Patch generation failed after {retries} retries."}
+
+
+ async def _get_code_context(self, issue: dict) -> dict:
+ """Retrieves relevant code context based on file paths mentioned in the issue."""
+ if not self.repo or not self.repo_local_path:
+ return {"content": "Local repository not available.", "files": [], "error": "Local repository not available."}
+
+ issue_id = issue.get('id', 'N/A')
+ issue_body = issue.get('body', '') or ""
+ issue_title = issue.get('title', '') or ""
+ text_to_search = f"{issue_title}\n{issue_body}"
+
+ if not text_to_search.strip():
+ return {"content": "No issue title or body provided to search for file paths.", "files": [], "error": None}
+
+ # Enhanced path pattern:
+ # - Looks for paths starting with /, ./, or just a filename/path segment
+ # - Allows common path characters (letters, numbers, ., /, _, -)
+ # - Requires a file extension from a common list
+ # - Uses word boundaries or punctuation/whitespace lookarounds to avoid partial matches within words
+ # - Allows paths enclosed in quotes or backticks
+ path_pattern = r"""
+ (?]|$) # Ensure it's followed by whitespace, punctuation, or end of string
+ """
+ potential_files = set(re.findall(path_pattern, text_to_search, re.VERBOSE | re.IGNORECASE))
+
+ if not potential_files:
+ return {"content": "No file paths matching common patterns found in the issue title or body.", "files": [], "error": None}
+
+ logger.info(f"Found {len(potential_files)} potential file references in issue {issue_id}: {potential_files}")
+ context_content = ""
+ max_context_length = 6000 # Character limit for total context fed to the model
+ files_included = []
+ files_not_found = []
+ files_read_error = []
+ files_skipped_length = []
+ file_snippet_length = 1000 # Characters per file snippet
+
+ for file_path_str in potential_files:
+ # Clean up quotes/backticks and normalize separators
+ normalized_path_str = file_path_str.strip('\'"`').replace('\\', '/')
+ # Resolve relative paths if the repo root is the current directory context
+ relative_path = Path(normalized_path_str)
+ try:
+ # Construct the full path relative to the repo root
+ full_path = self.repo_local_path / relative_path
+ # Resolve the path to handle '..' etc. and get the absolute path
+ resolved_full_path = full_path.resolve()
- except Exception as e:
- logger.error(f"Error processing file '{file_path.name}': {str(e)}")
- dataset.append({
- 'source': 'file',
- 'filename': file_path.name,
- 'file_size': file_size if 'file_size' in locals() else None,
- 'extracted_data': None,
- 'processing_notes': f'Overall file processing error: {str(e)}'
- })
- return dataset
+ # Ensure the resolved path is still within the repository directory
+ if not resolved_full_path.is_relative_to(self.repo_local_path.resolve()):
+ logger.warning(f"Skipping potentially malicious path outside repo: {file_path_str} resolved to {resolved_full_path}")
+ files_read_error.append(file_path_str + " (outside repo)")
+ continue
- def _is_archive(self, filepath: Union[str, Path]) -> bool:
- """Check if file is an archive"""
- p = Path(filepath) if isinstance(filepath, str) else filepath
- return p.suffix.lower() in self.archive_extensions
+ # Get the clean relative path string from the resolved path
+ relative_path_str = str(resolved_full_path.relative_to(self.repo_local_path.resolve()))
- def _process_single_file(self, file_path: Path) -> List[Dict]:
- """Process a single file with enhanced character extraction and format-specific handling"""
- dataset_entries = []
- filename = file_path.name
- file_size = file_path.stat().st_size
- mime_type, _ = mimetypes.guess_type(file_path)
- mime_type = mime_type or 'unknown/unknown'
- file_extension = file_path.suffix.lower()
+ except Exception as e:
+ logger.warning(f"Could not resolve or validate path {file_path_str} for issue {issue_id}: {e}")
+ files_read_error.append(file_path_str + " (path error)")
+ continue
- logger.info(f"Processing single file: '{filename}' ({mime_type}, {file_size} bytes)")
- raw_content: Optional[str] = None
- extracted_data: Any = None
- processing_notes = []
+ if resolved_full_path.is_file():
+ try:
+ file_content = resolved_full_path.read_text(encoding='utf-8', errors='ignore')
+ content_snippet = f"---\nFile: {relative_path_str}\n---\n{file_content[:file_snippet_length]}{' [...]' if len(file_content) > file_snippet_length else ''}\n\n"
- try:
- # Read content efficiently
- content_bytes = file_path.read_bytes()
- encoding_detection = chardet.detect(content_bytes)
- encoding = encoding_detection['encoding'] or 'utf-8'
- raw_content = content_bytes.decode(encoding, errors='replace')
+ if len(context_content) + len(content_snippet) <= max_context_length:
+ context_content += content_snippet
+ files_included.append(relative_path_str)
+ else:
+ logger.warning(f"Skipping file {relative_path_str} for context in issue {issue_id} due to total length limit ({max_context_length} chars).")
+ files_skipped_length.append(relative_path_str)
+ except OSError as e:
+ logger.warning(f"Could not read file {resolved_full_path} for issue {issue_id}: {e}")
+ files_read_error.append(relative_path_str)
+ except Exception as e:
+ logger.warning(f"Unexpected error reading file {resolved_full_path} for issue {issue_id}: {e}")
+ files_read_error.append(relative_path_str)
+ else:
+ logger.info(f"Potential path '{relative_path_str}' (from '{file_path_str}') not found or not a file in local repo for issue {issue_id}.")
+ files_not_found.append(relative_path_str)
+
+ final_content = ""
+ error_status = None
+ if files_included:
+ final_content = context_content.strip()
+ logger.info(f"Included context from {len(files_included)} files for issue {issue_id}: {files_included}")
+ else:
+ final_content = "No content could be retrieved from the potential file paths found."
+ logger.warning(f"Context generation for issue {issue_id} resulted in no included files.")
+ if potential_files: # If paths were found but none included
+ error_status = "No readable or found files among potential paths."
+
+ status_notes = []
+ if files_not_found:
+ status_notes.append(f"Files mentioned but not found: {files_not_found}")
+ logger.info(f"Files mentioned but not found for issue {issue_id}: {files_not_found}")
+ if files_read_error:
+ status_notes.append(f"Files failed to read: {files_read_error}")
+ logger.warning(f"Files mentioned but failed to read for issue {issue_id}: {files_read_error}")
+ if files_skipped_length:
+ status_notes.append(f"File content skipped due to length limit: {files_skipped_length}")
+ logger.warning(f"File content skipped due to length limit for issue {issue_id}: {files_skipped_length}")
+
+ if status_notes:
+ final_content += "\n\n--- Context Notes ---\n" + "\n".join(status_notes)
+ if error_status is None and (files_not_found or files_read_error):
+ error_status = "Some mentioned files were not found or could not be read."
+
+ return {"content": final_content.strip(), "files": files_included, "error": error_status}
+
+
+ async def suggest_resolution(self, issue: dict, model_key: str) -> str:
+ """Suggests a resolution description using a selected AI model."""
+ if not self.hf_token:
+ return "Error: Hugging Face token not set."
+ if model_key not in HF_MODELS or HF_MODELS.get(model_key) is None:
+ return f"Error: Invalid or unconfigured model key: {model_key}"
+
+ model_id = HF_MODELS[model_key]
+ issue_id = issue.get('id','N/A')
+ logger.info(f"Requesting resolution suggestion for issue {issue_id} ('{issue.get('title', 'N/A')[:50]}...') using {model_id}")
+
+ # --- Get Pre-computed Info ---
+ summary_text = self._get_precomputed_text(issue_id, self.precomputed_summaries, "summary", "Summary")
+ missing_info_text = self._get_precomputed_text(issue_id, self.precomputed_missing_info, "info_needed", "Missing Info Analysis")
+ analysis_text = self._get_precomputed_text(issue_id, self.precomputed_analysis, "hypothesis", "Preliminary Analysis")
+ duplicate_info = self._get_duplicate_info_text(issue_id)
+
+ # Enhanced Prompt - More structured and clear instructions
+ prompt = f"""You are a helpful AI assistant acting as a senior software engineer. Analyze the following GitHub issue and provide concise, actionable, step-by-step suggestions on how to approach its resolution.
+
+
+Issue ID: {issue.get('id', 'N/A')}
+Title: {issue.get('title', 'N/A')}
+Labels: {', '.join(issue.get('labels', []))}
+Body:
+{issue.get('body', 'N/A')[:2000]} # Limit body length for suggestion prompt
+
+
+
+Summary: {summary_text}
+Missing Information Analysis: {missing_info_text}
+Preliminary Hypothesis: {analysis_text}
+{duplicate_info}
+
+
+
+Based on *all* the information provided above, outline a potential plan. Use clear headings for each step.
+1. **Address Missing Information (If Any):** If the 'Missing Information Analysis' identified gaps, state concrete steps to get that information (e.g., "Request logs from the user," "Ask for steps to reproduce," "Check environment details"). If none needed, state "Issue description seems reasonably complete."
+2. **Refine Understanding / Verify Hypothesis:** Based on the 'Preliminary Hypothesis' and issue details, state the most likely root cause or the main goal of the issue. Mention any ambiguities or areas needing confirmation.
+3. **Identify Relevant Code Areas (Hypothesize):** Based on keywords, error messages, paths mentioned, or common conventions, list specific files, functions, or classes likely needing investigation. Explicitly state if this is speculative.
+4. **Propose Implementation / Investigation Steps:** Describe the core logic changes, debugging steps, configuration updates, or API interactions needed. Break it down into logical, actionable steps.
+5. **Testing Recommendations:** Suggest specific unit, integration, or manual tests crucial for verifying the fix or implementation. Mention key scenarios to cover.
+6. **Next Steps (Standard Workflow):** Briefly mention the standard development workflow (e.g., branching, coding, testing, committing, PR).
+
+
+Suggested Resolution Approach:
+"""
+ api_url = f"{HF_INFERENCE_API}/{model_id}"
+ headers = {"Authorization": f"Bearer {self.hf_token}"}
+ # Adjusted parameters for more creative but still focused output
+ payload = {
+ "inputs": prompt,
+ "parameters": {"max_new_tokens": 1500, "temperature": 0.7, "return_full_text": False, "do_sample": True, "top_p": 0.95},
+ "options": {"wait_for_model": True}
+ }
+ timeout = aiohttp.ClientTimeout(total=120) # Increased timeout
- # --- Attempt format-specific parsing ---
+ retries = 3
+ for attempt in range(retries):
+ try:
+ start_time_api = time.time()
+ async with aiohttp.ClientSession(headers=headers, timeout=timeout) as session:
+ async with session.post(api_url, json=payload) as response:
+ api_duration = time.time() - start_time_api
+ rate_limit_remaining = response.headers.get('X-Ratelimit-Remaining')
+ logger.debug(f"HF Suggestion API Response Status: {response.status}, Duration: {api_duration:.2f}s, RateLimit Remaining: {rate_limit_remaining}")
+
+ if response.status == 429: # Too Many Requests
+ retry_after = int(response.headers.get('Retry-After', 10))
+ logger.warning(f"HF Suggestion API rate limited. Waiting for {retry_after} seconds before retry {attempt + 1}/{retries}.")
+ await asyncio.sleep(retry_after)
+ continue # Retry the same request
+ response.raise_for_status() # Raise for other 4xx/5xx errors
+
+ result = await response.json()
+
+ if result and isinstance(result, list) and 'generated_text' in result[0]:
+ suggestion = result[0].get('generated_text', 'AI Error: No suggestion text generated.').strip()
+ logger.info(f"Received suggestion from {model_id} for issue {issue_id} ({len(suggestion)} chars).")
+ return suggestion
+ elif isinstance(result, dict) and 'error' in result:
+ error_msg = result['error']
+ estimated_time = result.get('estimated_time')
+ logger.error(f"HF Inference API suggestion error for issue {issue_id}: {error_msg}" + (f" (Est: {estimated_time}s)" if estimated_time else "") + f". Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1:
+ await asyncio.sleep(5) # Wait before retrying API error
+ continue
+ else:
+ logger.error("Max retries reached for API error during suggestion generation.")
+ return f"Error: AI model returned an error after retries: {error_msg}"
+ else:
+ logger.error(f"Unexpected suggestion response format from {model_id} for issue {issue_id}: {str(result)[:500]}. Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1:
+ await asyncio.sleep(5)
+ continue
+ else:
+ logger.error("Max retries reached for unexpected API response during suggestion generation.")
+ return "Error: Received unexpected response format from AI model after retries."
+
+ except asyncio.TimeoutError:
+ logger.warning(f"HF Inference API suggestion request timed out ({timeout.total}s) for issue {issue_id}. Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1:
+ await asyncio.sleep(5) # Wait before retrying timeout
+ continue
+ else:
+ logger.error("Max retries reached for suggestion generation timeout.")
+ return "Error: AI model request timed out after retries. The model might be busy."
+ except aiohttp.ClientResponseError as e:
+ error_body = await e.response.text()
+ logger.error(f"HF Inference API suggestion request failed for issue {issue_id}: Status={e.status}, Message='{e.message}'. Body: {error_body[:500]}. Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1 and e.status in [500, 502, 503, 504]: # Retry on server errors
+ await asyncio.sleep(5)
+ continue
+ else:
+ logger.error("Max retries reached or non-retryable error for suggestion generation.")
+ return f"Error: AI model request failed ({e.status}) after retries. Check model/token/API status. Details: {error_body[:200]}..."
+ except Exception as e:
+ logger.exception(f"Error suggesting resolution for issue {issue_id}: {e}. Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1:
+ await asyncio.sleep(5)
+ continue
+ else:
+ logger.error("Max retries reached for unexpected error during suggestion generation.")
+ return f"An unexpected error occurred during suggestion generation after retries: {e}"
+
+ # This part is reached if all retries fail
+ logger.error(f"Suggestion generation failed for issue {issue_id} after {retries} retries.")
+ return f"Suggestion generation failed after {retries} retries."
+
+
+ def _get_precomputed_text(self, issue_id: int, data_dict: dict, key: str, name: str) -> str:
+ """Safely retrieves precomputed text, handling errors and pending states."""
+ if issue_id in data_dict:
+ entry = data_dict[issue_id]
+ timestamp = entry.get("timestamp", 0)
+ # Consider it "recent" if computed within the last 2 idle cycles
+ is_recent = time.time() - timestamp < self.idle_processing_interval * 2
+
+ if entry.get("error"):
+ # Sanitize error message to prevent HTML injection in preview
+ sanitized_error = gr.Textbox.sanitize_html(str(entry['error']))
+ return f"❌ {name} Error (at {datetime.fromtimestamp(timestamp).strftime('%H:%M:%S')}): {sanitized_error}"
+ elif key in entry and entry[key] is not None: # Check key exists and is not None
+ # Sanitize the text content
+ sanitized_text = gr.Textbox.sanitize_html(str(entry[key]))
+ return sanitized_text
+ else: # No error, but key might be missing or None
+ if is_recent:
+ return f"⏳ {name} computation pending..."
+ else:
+ return f"(ℹ️ {name} not computed or result was empty)"
+ else:
+ return f"(⏳ {name} not computed yet)"
+
+ def _get_duplicate_info_text(self, issue_id: int) -> str:
+ """Formats duplicate info text."""
+ if issue_id in self.potential_duplicates:
+ dup_ids = self.potential_duplicates[issue_id]
+ if dup_ids:
+ # Sort duplicate IDs for consistent display
+ sorted_dup_ids = sorted(dup_ids)
+ # Use the JS selectIssue function to make links clickable in the preview HTML
+ dup_links = ", ".join([f"#{dup_id}" for dup_id in sorted_dup_ids])
+ return f"\n### ⚠️ Potential Duplicate Issues:\nIssue IDs: {dup_links}"
+ return ""
- # 1. Attempt JSON parsing (explicit .json or application/json, OR if content looks like JSON)
- is_explicit_json = mime_type == 'application/json' or file_extension == '.json'
- looks_like_json = raw_content.strip().startswith('{') or raw_content.strip().startswith('[')
+ async def broadcast_collaboration_status(self):
+ """Periodically sends collaborator status to all connected clients."""
+ while True:
+ try:
+ await asyncio.sleep(5)
+ if not self.ws_clients: continue
+
+ # Clean up collaborators dict - remove entries for clients no longer in ws_clients
+ active_client_ids = {getattr(client, 'client_id', None) for client in self.ws_clients}
+ collaborators_to_remove = [cid for cid in self.collaborators if cid not in active_client_ids]
+ for cid in collaborators_to_remove:
+ logger.debug(f"Removing stale collaborator entry for client ID {cid}")
+ del self.collaborators[cid]
+
+ status_payload = json.dumps({"type": "collaboration_status", "collaborators": self.collaborators})
+
+ active_clients_snapshot = list(self.ws_clients) # Take a snapshot
+ tasks = []
+ # No need for disconnected_clients list here, remove_ws_client handles it
+
+ for client in active_clients_snapshot:
+ # Check if client is closed using the standard websockets property
+ if client.closed:
+ # remove_ws_client will be called by the exception handler or finally block
+ continue
+ try:
+ tasks.append(client.send(status_payload))
+ except (ConnectionClosed, ConnectionAbortedError, ConnectionResetError, WebSocketException) as e:
+ # These exceptions indicate the connection is likely already closed or broken
+ logger.warning(f"Client {getattr(client, 'client_id', client.remote_address)} seems disconnected before send: {type(e).__name__}. Attempting removal.")
+ # Schedule removal in the main loop if not already in it
+ if self.main_loop.is_running():
+ self.main_loop.call_soon_threadsafe(self.remove_ws_client, client)
+ else:
+ logger.error("Main loop not running, cannot schedule client removal.")
+ continue # Skip to next client
+ except Exception as e:
+ logger.error(f"Unexpected error preparing send to client {getattr(client, 'client_id', client.remote_address)}: {e}. Attempting removal.")
+ if self.main_loop.is_running():
+ self.main_loop.call_soon_threadsafe(self.remove_ws_client, client)
+ else:
+ logger.warning("Main loop not running, cannot schedule client removal.")
+ continue # Skip to next client
- if is_explicit_json or looks_like_json:
- try:
- extracted_data = json.loads(raw_content)
- processing_notes.append("Parsed as JSON.")
- if not is_explicit_json:
- processing_notes.append("Note: Content looked like JSON despite extension/mime.")
- logger.warning(f"File '{filename}' identified as JSON content despite extension/mime.")
- mime_type = 'application/json' # Update mime_type if successfully parsed as JSON
- except json.JSONDecodeError as e:
- processing_notes.append(f"Failed to parse as JSON: {e}.")
- if is_explicit_json:
- logger.error(f"Explicit JSON file '{filename}' has invalid format: {e}")
- else:
- logger.warning(f"Content of '{filename}' looks like JSON but failed to parse: {e}")
- except Exception as e:
- processing_notes.append(f"Error processing JSON: {e}.")
- logger.error(f"Error processing JSON in '{filename}': {e}")
- # 2. Attempt XML parsing (if not already parsed as JSON, and looks like XML)
- # Add check if extracted_data is still None (meaning JSON parsing failed or wasn't attempted/relevant)
- looks_like_xml = extracted_data is None and raw_content.strip().startswith('<') and raw_content.strip().endswith('>') # Simple heuristic
- is_explicit_xml = extracted_data is None and (mime_type in ('application/xml', 'text/xml') or mime_type.endswith('+xml') or file_extension in ('.xml', '.xsd'))
+ # No need to await tasks here if remove_ws_client is scheduled via call_soon_threadsafe
- if extracted_data is None and (is_explicit_xml or looks_like_xml):
- try:
- root = ET.fromstring(raw_content)
- # Convert XML element tree to a structured dictionary or string
- # Simple string representation for QR code suitability
- extracted_data = ET.tostring(root, encoding='unicode', method='xml')
- processing_notes.append("Parsed as XML (text representation).")
- if not is_explicit_xml:
- processing_notes.append("Note: Content looked like XML despite extension/mime.")
- # Update mime_type if successfully parsed as XML
- if 'xml' not in mime_type: mime_type = 'application/xml'
- except ET.ParseError as e:
- processing_notes.append(f"Failed to parse as XML: {e}.")
- if is_explicit_xml:
- logger.error(f"Explicit XML file '{filename}' has invalid format: {e}")
- else:
- logger.warning(f"Content of '{filename}' looks like XML but failed to parse: {e}")
- except Exception as e:
- processing_notes.append(f"Error processing XML: {e}.")
- logger.error(f"Error processing XML in '{filename}': {e}")
+ except asyncio.CancelledError:
+ logger.info("Broadcast loop cancelled.")
+ break
+ except Exception as e:
+ logger.exception(f"Error in broadcast loop: {e}")
+ await asyncio.sleep(10) # Wait before next attempt after unexpected error
+ async def handle_code_editor_update(self, issue_num: int, delta_str: str, sender_client_id: str):
+ """
+ Applies a delta from one client and broadcasts it to others.
+ WARNING: Lacks Operational Transformation - concurrent edits are UNSAFE.
+ """
+ # Ensure issue exists and is currently loaded/tracked
+ if issue_num is None or issue_num not in self.issues:
+ logger.warning(f"Received code update for non-existent issue {issue_num} in manager. Ignoring.")
+ return
+ # Ensure an editor instance exists for this issue
+ if issue_num not in self.code_editors:
+ logger.warning(f"Received code update for issue {issue_num} but no editor instance found. Ignoring.")
+ # This state indicates the UI might be out of sync with the backend state.
+ # Ideally, the client only sends updates after handle_issue_select has run.
+ return
- # 3. Attempt CSV parsing (if not already parsed, and looks like CSV or is explicit CSV)
- is_explicit_csv = extracted_data is None and (mime_type == 'text/csv' or file_extension == '.csv')
- # Heuristic: check for commas/semicolons and multiple lines
- looks_like_csv = extracted_data is None and (',' in raw_content or ';' in raw_content) and ('\n' in raw_content or len(raw_content.splitlines()) > 1)
- if extracted_data is None and (is_explicit_csv or looks_like_csv):
- try:
- # Use Sniffer to guess dialect for better compatibility
- dialect = 'excel' # Default dialect
- try:
- # Look at first few lines to guess dialect
- sample = '\n'.join(raw_content.splitlines()[:10])
- if sample:
- dialect = csv.Sniffer().sniff(sample).name
- logger.debug(f"Sniffer detected CSV dialect: {dialect} for '{filename}'")
- except csv.Error:
- logger.debug(f"Sniffer failed to detect dialect for '{filename}', using 'excel'.")
- dialect = 'excel' # Fallback
-
- # Read using the guessed or default dialect
- csv_reader = csv.reader(io.StringIO(raw_content), dialect=dialect)
- rows = list(csv_reader)
-
- if rows:
- # Limit the number of rows included for potentially huge CSVs
- max_rows_preview = 100
- extracted_data = {
- 'headers': rows[0] if rows[0] else None, # Assume first row is header
- 'rows': rows[1:max_rows_preview+1] # Get up to max_rows_preview data rows
- }
- if len(rows) > max_rows_preview + 1:
- processing_notes.append(f"CSV truncated to {max_rows_preview} data rows.")
- processing_notes.append("Parsed as CSV.")
- if not is_explicit_csv:
- processing_notes.append("Note: Content looked like CSV despite extension/mime.")
- mime_type = 'text/csv' # Update mime_type
+ logger.debug(f"Handling code editor update for issue {issue_num} from {sender_client_id}. "
+ "WARNING: NO OT IMPLEMENTED - Last write wins / potential conflicts.")
- else:
- extracted_data = "Empty CSV"
- processing_notes.append("Parsed as empty CSV.")
- if not is_explicit_csv:
- processing_notes.append("Note: Content looked like CSV but was empty.")
+ try:
+ delta_obj = json.loads(delta_str)
+ # In a real OT system, apply_delta would update the authoritative state.
+ # Here, we just log and increment revision for the placeholder.
+ # The actual text state is managed client-side by the Gradio component.
+ self.code_editors[issue_num].apply_delta(delta_obj)
+ logger.debug(f"Applied delta for issue {issue_num} from client {sender_client_id} (Placeholder OT Logic - Revision {self.code_editors[issue_num].revision})")
+
+ update_payload = json.dumps({
+ "type": "code_update",
+ "issue_num": issue_num,
+ "delta": delta_str, # Send the original delta string
+ "senderId": sender_client_id
+ })
- except Exception as e:
- processing_notes.append(f"Failed to parse as CSV: {e}.")
- logger.warning(f"Failed to parse CSV from '{filename}': {e}")
+ active_clients_snapshot = list(self.ws_clients) # Take a snapshot
+ tasks = []
+ # No need for disconnected_clients list here, remove_ws_client handles it
+ for client in active_clients_snapshot:
+ client_ws_id = getattr(client, 'client_id', None)
+ # Only broadcast to other clients, and only if the client is open
+ if client_ws_id != sender_client_id and not client.closed:
+ try:
+ tasks.append(client.send(update_payload))
+ except (ConnectionClosed, ConnectionAbortedError, ConnectionResetError, WebSocketException) as e:
+ logger.warning(f"Client {getattr(client, 'client_id', client.remote_address)} seems disconnected during broadcast: {type(e).__name__}. Attempting removal.")
+ # Schedule removal in the main loop if not already in it
+ if self.main_loop.is_running():
+ self.main_loop.call_soon_threadsafe(self.remove_ws_client, client)
+ else:
+ logger.warning("Main loop not running, cannot schedule client removal.")
+ continue # Skip to next client
+ except Exception as e:
+ logger.error(f"Unexpected error preparing broadcast to client {getattr(client, 'client_id', client.remote_address)}: {e}. Attempting removal.")
+ if self.main_loop.is_running():
+ self.main_loop.call_soon_threadsafe(self.remove_ws_client, client)
+ else:
+ logger.warning("Main loop not running, cannot schedule client removal.")
+ continue # Skip to next client
+
+
+ if tasks:
+ logger.debug(f"Broadcasting code update for issue {issue_num} to {len(tasks)} other clients.")
+ # No need to await tasks here if remove_ws_client is scheduled via call_soon_threadsafe
+
+
+ except json.JSONDecodeError:
+ logger.error(f"Received invalid JSON delta for issue {issue_num} from {sender_client_id}: {delta_str[:200]}")
+ except Exception as e:
+ logger.exception(f"Error handling code editor update for issue {issue_num} from {sender_client_id}: {e}")
+
+ async def broadcast_issue_update(self):
+ """Notifies clients that the issue list/data has changed (e.g., due to webhook)."""
+ if not self.ws_clients: return
+
+ logger.info("Broadcasting 'issues_updated' notification to all clients.")
+ update_payload = json.dumps({"type": "issues_updated"})
+ active_clients_snapshot = list(self.ws_clients)
+ tasks = []
+ # No need for disconnected_clients list here, remove_ws_client handles it
+
+ for client in active_clients_snapshot:
+ if client.closed:
+ continue
+ try:
+ tasks.append(client.send(update_payload))
+ except (ConnectionClosed, ConnectionAbortedError, ConnectionResetError, WebSocketException) as e:
+ logger.warning(f"Client {getattr(client, 'client_id', client.remote_address)} seems disconnected during issue update broadcast: {type(e).__name__}. Attempting removal.")
+ if self.main_loop.is_running():
+ self.main_loop.call_soon_threadsafe(self.remove_ws_client, client)
+ else:
+ logger.warning("Main loop not running, cannot schedule client removal.")
+ continue
+ except Exception as e:
+ logger.error(f"Unexpected error preparing issue update broadcast to client {getattr(client, 'client_id', client.remote_address)}: {e}. Attempting removal.")
+ if self.main_loop.is_running():
+ self.main_loop.call_soon_threadsafe(self.remove_ws_client, client)
+ else:
+ logger.warning("Main loop not running, cannot schedule client removal.")
+ continue
+
+ if tasks:
+ logger.debug(f"Broadcasting issue update to {len(tasks)} clients.")
+ # No need to await tasks here if remove_ws_client is scheduled via call_soon_threadsafe
+
+
+ def remove_ws_client(self, client_to_remove: WebSocketServerProtocol):
+ """Safely removes a client from the list and collaborator dict."""
+ # Use the stored client_id or remote_address for logging
+ client_id = getattr(client_to_remove, 'client_id', None)
+ client_addr = client_to_remove.remote_address if hasattr(client_to_remove, 'remote_address') else 'N/A'
+ client_desc = f"{client_id or 'Unknown ID'} ({client_addr})"
+ removed_from_list = False
+ removed_from_collab = False
+
+ # This method should ideally be called from the main async loop thread
+ # if it modifies shared state like self.ws_clients and self.collaborators.
+ # If called from elsewhere (e.g., a thread's error handler), use call_soon_threadsafe.
+ # Assuming it's called from the WS async handler or scheduled into the main loop.
- # 4. Attempt Document Text Extraction (if not already parsed)
- if extracted_data is None:
- try:
- extracted_text = None
- if file_extension == '.pdf' and PDF_SUPPORT:
- with tempfile.NamedTemporaryFile(delete=False, suffix='.pdf') as tmp_file:
- tmp_file.write(content_bytes) # Write bytes to temp file
- temp_path = Path(tmp_file.name)
- try:
- reader = PdfReader(temp_path)
- text_content = "".join(page.extract_text() or "" for page in reader.pages)
- extracted_text = text_content
- processing_notes.append("Extracted text from PDF.")
- finally:
- temp_path.unlink() # Clean up temp file
- elif file_extension == '.docx' and DOCX_SUPPORT:
- with tempfile.NamedTemporaryFile(delete=False, suffix='.docx') as tmp_file:
- tmp_file.write(content_bytes) # Write bytes to temp file
- temp_path = Path(tmp_file.name)
- try:
- document = Document(temp_path)
- text_content = "\n".join(paragraph.text for paragraph in document.paragraphs)
- extracted_text = text_content
- processing_notes.append("Extracted text from DOCX.")
- finally:
- temp_path.unlink() # Clean up temp file
- elif file_extension == '.rtf' and RTF_SUPPORT:
- # pyth can read directly from file-like object or string
- try:
- doc = Rtf15Reader.read(io.StringIO(raw_content))
- text_content = PlaintextWriter.write(doc).getvalue()
- extracted_text = text_content
- processing_notes.append("Extracted text from RTF.")
- except Exception as e:
- processing_notes.append(f"RTF extraction error: {e}")
- logger.warning(f"Failed to extract RTF text from '{filename}': {e}")
- elif file_extension == '.odt' and ODT_SUPPORT:
- with tempfile.NamedTemporaryFile(delete=False, suffix='.odt') as tmp_file:
- tmp_file.write(content_bytes) # Write bytes to temp file
- temp_path = Path(tmp_file.name)
- try:
- text_doc = OpenDocumentText(temp_path)
- paragraphs = text_doc.getElementsByType(odftext.P)
- text_content = "\n".join("".join(node.text for node in p.childNodes) for p in paragraphs)
- extracted_text = text_content
- processing_notes.append("Extracted text from ODT.")
- finally:
- temp_path.unlink() # Clean up temp file
- elif file_extension in ['.doc', '.ppt', '.pptx', '.xls', '.xlsx']:
- # These require more complex or platform-specific libraries (e.g. antiword, pandoc, COM objects on Windows)
- processing_notes.append(f"Automatic text extraction for {file_extension.upper()} not fully implemented.")
- logger.warning(f"Automatic text extraction for {file_extension.upper()} not fully implemented for '{filename}'.")
-
- if extracted_text is not None:
- # Limit extracted text size
- max_extracted_text_size = 10000 # Limit text preview
- extracted_data = {'text': extracted_text[:max_extracted_text_size]}
- if len(extracted_text) > max_extracted_text_size:
- extracted_data['text'] += "..."
- processing_notes.append("Extracted text truncated.")
-
- except ImportError as e:
- processing_notes.append(f"Missing dependency for document type ({e}). Cannot extract text.")
- except Exception as e:
- processing_notes.append(f"Error during document text extraction: {e}")
- logger.warning(f"Error during document text extraction for '{filename}': {e}")
+ try:
+ # Check if the client is actually in the list before attempting removal
+ if client_to_remove in self.ws_clients:
+ self.ws_clients.remove(client_to_remove)
+ removed_from_list = True
+ logger.info(f"Removed WebSocket client from list: {client_desc} (Remaining: {len(self.ws_clients)})")
+ else:
+ logger.debug(f"Client {client_desc} not found in list during removal attempt.")
+
+ except ValueError:
+ logger.debug(f"Client {client_desc} already removed from list or not found (ValueError).")
+ pass # Already removed
+
+ if client_id and client_id in self.collaborators:
+ del self.collaborators[client_id]
+ removed_from_collab = True
+ logger.info(f"Removed collaborator entry for {client_id}.")
+
+ # Schedule an immediate status broadcast if anything was removed
+ if (removed_from_list or removed_from_collab):
+ if self.main_loop.is_running():
+ # Use call_soon_threadsafe to schedule the async coroutine in the main loop
+ self.main_loop.call_soon_threadsafe(asyncio.create_task, self.broadcast_collaboration_status_once())
+ logger.debug(f"Scheduled immediate status broadcast after removing client {client_desc}.")
+ else:
+ logger.warning("Main loop not running, cannot schedule immediate status broadcast after client removal.")
- # 5. Fallback to Plain Text (if no specific extraction succeeded)
- if extracted_data is None:
- extracted_data = {'plain_text': raw_content}
- processing_notes.append("Stored as plain text.")
- # Re-guess mime type if it was something specific like application/octet-stream and we just got text
- if mime_type in ['unknown/unknown', 'application/octet-stream']:
- guessed_text_mime, _ = mimetypes.guess_type('dummy.txt') # Use a dummy file name to guess plain text
- if guessed_text_mime: mime_type = guessed_text_mime
+ async def broadcast_collaboration_status_once(self):
+ """Sends a single collaboration status update immediately."""
+ if not self.ws_clients: return
+ status_payload = json.dumps({"type": "collaboration_status", "collaborators": self.collaborators})
+ active_clients_snapshot = list(self.ws_clients)
+ tasks = []
+ # No need for disconnected_clients list here, remove_ws_client handles it
+ for client in active_clients_snapshot:
+ # Check if client is closed using the standard websockets property
+ if client.closed:
+ continue
+ try:
+ tasks.append(client.send(status_payload))
+ except (ConnectionClosed, ConnectionAbortedError, ConnectionResetError, WebSocketException) as e:
+ logger.warning(f"Client {getattr(client, 'client_id', client.remote_address)} seems disconnected during single broadcast: {type(e).__name__}. Attempting removal.")
+ if self.main_loop.is_running():
+ self.main_loop.call_soon_threadsafe(self.remove_ws_client, client)
+ else:
+ logger.warning("Main loop not running, cannot schedule client removal.")
+ continue
+ except Exception as e:
+ logger.error(f"Unexpected error preparing single broadcast to client {getattr(client, 'client_id', client.remote_address)}: {e}. Attempting removal.")
+ if self.main_loop.is_running():
+ self.main_loop.call_soon_threadsafe(self.remove_ws_client, client)
+ else:
+ logger.warning("Main loop not running, cannot schedule client removal.")
+ continue
- except Exception as e:
- # Catch errors during initial read or other unexpected issues
- logger.error(f"Fatal error processing single file '{filename}': {e}")
- processing_notes.append(f"Fatal processing error: {e}")
- raw_content = None # Ensure raw_content is None if reading failed
- extracted_data = None
-
-
- # Add file info to the entry
- entry = {
- 'source': 'file',
- 'filename': filename,
- 'file_size': file_size,
- 'mime_type': mime_type,
- 'created': datetime.fromtimestamp(file_path.stat().st_ctime).isoformat() if file_path.exists() else None,
- 'modified': datetime.fromtimestamp(file_path.stat().st_mtime).isoformat() if file_path.exists() else None,
- 'raw_content': raw_content, # Always include raw content if readable
- 'extracted_data': extracted_data, # Include the structured/extracted data
- 'processing_notes': processing_notes # Include any notes/errors encountered
- }
+ if tasks:
+ logger.debug(f"Broadcasting single status update to {len(tasks)} clients.")
+ # No need to await tasks here if remove_ws_client is scheduled via call_soon_threadsafe
- dataset_entries.append(entry)
- return dataset_entries
+ def _identify_stale_issues(self):
+ """Identifies issues not updated recently based on 'updated_at'."""
+ self.stale_issues = []
+ threshold = timedelta(days=self.stale_issue_threshold_days)
+ now_aware = datetime.now(timezone.utc)
- def _process_archive(self, archive_path: Path, extract_to: Path) -> List[Dict]:
- """Process an archive file with enhanced extraction"""
- dataset = []
- archive_extension = archive_path.suffix.lower()
- logger.info(f"Processing archive: '{archive_path.name}'")
+ for issue_id, issue_data in self.issues.items():
+ updated_at_str = issue_data.get("updated_at")
+ if updated_at_str:
+ try:
+ # Handle potential different ISO 8601 formats (with/without Z, different offsets)
+ # datetime.fromisoformat handles Z and +/-HH:MM offsets
+ updated_at = datetime.fromisoformat(updated_at_str.replace("Z", "+00:00"))
+ # Ensure both datetimes are timezone-aware for comparison
+ if updated_at.tzinfo is None:
+ # Assume UTC if no timezone info (common for older systems)
+ updated_at = updated_at.replace(tzinfo=timezone.utc)
+
+ if now_aware - updated_at > threshold:
+ self.stale_issues.append(issue_id)
+ except (ValueError, TypeError) as e:
+ logger.warning(f"Could not parse 'updated_at' ('{updated_at_str}') for issue {issue_id}: {e}")
+ else:
+ logger.debug(f"Issue {issue_id} missing 'updated_at' field.")
+
+ logger.info(f"Identified {len(self.stale_issues)} potentially stale issues (updated > {self.stale_issue_threshold_days} days ago).")
+
+ def _identify_high_priority_candidates(self):
+ """Identifies high-priority issues (e.g., Critical/High severity)."""
+ self.high_priority_candidates = []
+ for issue_id, issue_data in self.issues.items():
+ severity = self._determine_severity(issue_data.get('labels', []))
+ if severity in ["Critical", "High"]:
+ self.high_priority_candidates.append(issue_id)
+ logger.info(f"Identified {len(self.high_priority_candidates)} high-priority candidates (Critical/High severity).")
+
+ async def _compute_and_store_summary(self, issue_id: int):
+ """Generates and stores a summary for a given issue using an LLM (Idle Task)."""
+ # Re-check if issue exists just before starting computation
+ if issue_id not in self.issues:
+ logger.warning(f"Skipping summary generation for issue {issue_id}: Issue no longer exists.")
+ return
+ if not self.hf_token:
+ self.precomputed_summaries[issue_id] = {"error": "HF token not set", "timestamp": time.time()}
+ return
+ if DEFAULT_IDLE_MODEL_ID is None:
+ self.precomputed_summaries[issue_id] = {"error": "Idle model not configured", "timestamp": time.time()}
+ return
try:
- if archive_extension == '.zip':
- if zipfile.is_zipfile(archive_path):
- with zipfile.ZipFile(archive_path, 'r') as zip_ref:
- for file_info in zip_ref.infolist():
- if file_info.file_size > 0 and not file_info.filename.endswith('/'):
- try:
- zip_ref.extract(file_info, path=extract_to)
- extracted_file_path = extract_to / file_info.filename
- # Recursively process the extracted file if it's supported and not an archive itself
- if extracted_file_path.suffix.lower() in self.supported_extensions and not self._is_archive(extracted_file_path):
- dataset.extend(self._process_single_file(extracted_file_path))
- elif extracted_file_path.suffix.lower() in self.archive_extensions:
- # Recursively process nested archives (careful with depth!)
- logger.info(f"Found nested archive '{file_info.filename}', processing recursively.")
- dataset.extend(self._process_archive(extracted_file_path, extract_to))
- else:
- logger.debug(f"Skipping unsupported file in archive: '{file_info.filename}'")
- except Exception as e:
- logger.warning(f"Error extracting/processing file '{file_info.filename}' from zip '{archive_path.name}': {e}")
- else:
- logger.error(f"'{archive_path.name}' is not a valid zip file.")
+ issue = self.issues[issue_id]
+ model_id = DEFAULT_IDLE_MODEL_ID # Use designated idle model
+ logger.info(f"Idle Task: Generating summary for issue {issue_id} using {model_id}")
+ start_time = time.time()
+ prompt = f"""Concisely summarize the following GitHub issue in 1-2 sentences. Focus on the core problem or request reported by the user.
+Issue Title: {issue.get('title', 'N/A')}
+Issue Body (first 1000 chars):
+{issue.get('body', 'N/A')[:1000]}
+Summary:"""
+ api_url = f"{HF_INFERENCE_API}/{model_id}"
+ headers = {"Authorization": f"Bearer {self.hf_token}"}
+ payload = {
+ "inputs": prompt,
+ "parameters": {"max_new_tokens": 128, "temperature": 0.2, "return_full_text": False, "do_sample": False},
+ "options": {"wait_for_model": True}
+ }
+ timeout = aiohttp.ClientTimeout(total=60)
- elif archive_extension in ('.tar', '.gz', '.tgz'):
+ retries = 2 # Fewer retries for idle tasks
+ for attempt in range(retries):
try:
- # Determine mode: 'r' for tar, 'r:gz' for tar.gz, 'r:bz2' for tar.bz2 (bz2 not fully supported yet)
- mode = 'r'
- if archive_extension in ('.tar.gz', '.tgz'): mode = 'r:gz'
- # elif archive_extension == '.tar.bz2': mode = 'r:bz2' # Needs bz2 support
- # Note: 'r:*' attempts to guess compression, safer to be explicit
-
- with tarfile.open(archive_path, mode) as tar_ref:
- for member in tar_ref.getmembers():
- if member.isfile():
- try:
- tar_ref.extract(member, path=extract_to)
- extracted_file_path = extract_to / member.name
- # Recursively process extracted file
- if extracted_file_path.suffix.lower() in self.supported_extensions and not self._is_archive(extracted_file_path):
- dataset.extend(self._process_single_file(extracted_file_path))
- elif extracted_file_path.suffix.lower() in self.archive_extensions:
- logger.info(f"Found nested archive '{member.name}', processing recursively.")
- dataset.extend(self._process_archive(extracted_file_path, extract_to))
- else:
- logger.debug(f"Skipping unsupported file in archive: '{member.name}'")
- except Exception as e:
- logger.warning(f"Error extracting/processing file '{member.name}' from tar '{archive_path.name}': {e}")
- except tarfile.TarError as e:
- logger.error(f"Error processing TAR archive '{archive_path.name}': {e}")
-
- elif archive_extension == '.gz':
- # GZIP archives typically contain a single file. Extract it and process.
- extracted_name = archive_path.stem # Get name without .gz
- extracted_path = extract_to / extracted_name
- try:
- with gzip.open(archive_path, 'rb') as gz_file, open(extracted_path, 'wb') as outfile:
- outfile.write(gz_file.read())
- # Process the extracted file if supported
- if extracted_path.suffix.lower() in self.supported_extensions and not self._is_archive(extracted_path):
- dataset.extend(self._process_single_file(extracted_path))
- elif extracted_path.suffix.lower() in self.archive_extensions:
- logger.info(f"Found nested archive '{extracted_name}', processing recursively.")
- dataset.extend(self._process_archive(extracted_path, extract_to))
+ async with aiohttp.ClientSession(headers=headers, timeout=timeout) as session:
+ async with session.post(api_url, json=payload) as response:
+ if response.status == 429:
+ retry_after = int(response.headers.get('Retry-After', 5))
+ logger.warning(f"Idle Summary API rate limited. Waiting {retry_after}s before retry {attempt + 1}/{retries}.")
+ await asyncio.sleep(retry_after)
+ continue
+ response.raise_for_status()
+ result = await response.json()
+ duration = time.time() - start_time
+
+ if result and isinstance(result, list) and 'generated_text' in result[0]:
+ summary = result[0].get('generated_text', '').strip() or "(AI generated empty summary)"
+ self.precomputed_summaries[issue_id] = {"summary": summary, "error": None, "timestamp": time.time()}
+ logger.info(f"Stored summary for issue {issue_id} (took {duration:.2f}s).")
+ break # Success
+ elif isinstance(result, dict) and 'error' in result:
+ raise ValueError(f"API Error: {result['error']}")
+ else:
+ raise ValueError(f"Unexpected API response format: {str(result)[:200]}")
+ except (asyncio.TimeoutError, aiohttp.ClientResponseError, ValueError) as e:
+ logger.warning(f"Idle Summary API error for issue {issue_id}: {e}. Retry {attempt + 1}/{retries}.", exc_info=False)
+ if attempt < retries - 1:
+ await asyncio.sleep(5)
+ continue
else:
- logger.debug(f"Skipping unsupported file (from gz): '{extracted_name}'")
-
- except gzip.GzipFile as e:
- logger.error(f"Error processing GZIP file '{archive_path.name}': {e}")
- except Exception as e:
- logger.error(f"Error extracting/processing from GZIP '{archive_path.name}': {e}")
- finally:
- if extracted_path.exists(): extracted_path.unlink() # Clean up extracted file
+ raise e # Re-raise after max retries
+ except Exception as e:
+ logger.exception(f"Unexpected error in Idle Summary task for issue {issue_id}. Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1:
+ await asyncio.sleep(5)
+ continue
+ else:
+ raise e # Re-raise after max retries
+ else: # Executed if loop completes without break
+ raise Exception(f"Max retries ({retries}) failed for issue {issue_id}")
- # TODO: Add support for other archive types (.bz2, .7z, .rar)
- elif archive_extension in ('.bz2', '.7z', '.rar'):
- logger.warning(f"Support for {archive_extension} archives is not yet fully implemented and requires external tools/libraries.")
except Exception as e:
- logger.error(f"Overall archive processing error for '{archive_path.name}': {e}")
+ err_msg = f"Failed summary: {e}"
+ logger.error(f"Failed to generate summary for issue {issue_id}: {e}", exc_info=False) # Keep log cleaner
+ self.precomputed_summaries[issue_id] = {"error": err_msg, "summary": None, "timestamp": time.time()}
+
+ async def _compute_and_store_missing_info(self, issue_id: int):
+ """Idle Task: Use LLM to identify missing information needed for an issue."""
+ # Re-check if issue exists just before starting computation
+ if issue_id not in self.issues: return
+ if not self.hf_token:
+ self.precomputed_missing_info[issue_id] = {"error": "HF token not set", "timestamp": time.time()}
+ return
+ if DEFAULT_IDLE_MODEL_ID is None:
+ self.precomputed_missing_info[issue_id] = {"error": "Idle model not configured", "timestamp": time.time()}
+ return
- # Clean up extracted files in temp_dir after processing
- # Handled by context manager 'with tempfile.TemporaryDirectory()'
+ try:
+ issue = self.issues[issue_id]
+ model_id = DEFAULT_IDLE_MODEL_ID # Use cheap model
+ logger.info(f"Idle Task: Identifying missing info for issue {issue_id} using {model_id}")
+ start_time = time.time()
+
+ prompt = f"""Analyze the following GitHub issue description. Identify critical information potentially missing for effective debugging or resolution. List the missing items concisely (e.g., "Steps to reproduce", "Error logs", "Expected vs. Actual behavior", "Environment details"). If the description seems reasonably complete, respond with ONLY the word "None".
+Issue Title: {issue.get('title', 'N/A')}
+Issue Body:
+{issue.get('body', 'N/A')[:1500]}
+Missing Information:"""
+
+ api_url = f"{HF_INFERENCE_API}/{model_id}"
+ headers = {"Authorization": f"Bearer {self.hf_token}"}
+ payload = {
+ "inputs": prompt,
+ "parameters": {"max_new_tokens": 64, "temperature": 0.1, "return_full_text": False, "do_sample": False},
+ "options": {"wait_for_model": True}
+ }
+ timeout = aiohttp.ClientTimeout(total=45)
+ retries = 2
+ for attempt in range(retries):
+ try:
+ async with aiohttp.ClientSession(headers=headers, timeout=timeout) as session:
+ async with session.post(api_url, json=payload) as response:
+ if response.status == 429:
+ retry_after = int(response.headers.get('Retry-After', 5))
+ logger.warning(f"Idle Missing Info API rate limited. Waiting {retry_after}s before retry {attempt + 1}/{retries}.")
+ await asyncio.sleep(retry_after)
+ continue
+ response.raise_for_status()
+ result = await response.json()
+ duration = time.time() - start_time
+
+ if result and isinstance(result, list) and 'generated_text' in result[0]:
+ info_needed = result[0].get('generated_text', '').strip()
+ if info_needed.lower() == "none" or not info_needed:
+ info_needed = "None needed."
+ self.precomputed_missing_info[issue_id] = {"info_needed": info_needed, "error": None, "timestamp": time.time()}
+ logger.info(f"Stored missing info analysis for issue {issue_id} (took {duration:.2f}s): '{info_needed[:50]}...'")
+ break # Success
+ elif isinstance(result, dict) and 'error' in result:
+ raise ValueError(f"API Error: {result['error']}")
+ else:
+ raise ValueError(f"Unexpected API response format: {str(result)[:200]}")
+ except (asyncio.TimeoutError, aiohttp.ClientResponseError, ValueError) as e:
+ logger.warning(f"Idle Missing Info API error for issue {issue_id}: {e}. Retry {attempt + 1}/{retries}.", exc_info=False)
+ if attempt < retries - 1:
+ await asyncio.sleep(5)
+ continue
+ else:
+ raise e # Re-raise after max retries
+ except Exception as e:
+ logger.exception(f"Unexpected error in Idle Missing Info task for issue {issue_id}. Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1:
+ await asyncio.sleep(5)
+ continue
+ else:
+ raise e # Re-raise after max retries
+ else: # Executed if loop completes without break
+ raise Exception(f"Max retries ({retries}) failed for issue {issue_id}")
- return dataset
+ except Exception as e:
+ err_msg = f"Failed missing info analysis: {e}"
+ logger.error(f"Failed missing info analysis for issue {issue_id}: {e}", exc_info=False)
+ self.precomputed_missing_info[issue_id] = {"error": err_msg, "info_needed": None, "timestamp": time.time()}
+
+ async def _compute_and_store_preliminary_analysis(self, issue_id: int):
+ """Idle Task: Use LLM for a very concise preliminary hypothesis."""
+ # Re-check if issue exists just before starting computation
+ if issue_id not in self.issues: return
+ if not self.hf_token:
+ self.precomputed_analysis[issue_id] = {"error": "HF token not set", "timestamp": time.time()}
+ return
+ if DEFAULT_IDLE_MODEL_ID is None:
+ self.precomputed_analysis[issue_id] = {"error": "Idle model not configured", "timestamp": time.time()}
+ return
- def chunk_data(self, data: Union[Dict, List], max_size: int = 2953) -> List[Dict]:
- """Enhanced data chunking with sequence metadata"""
try:
- # Convert data to JSON string
- # Use separators=(',', ':') to remove unnecessary whitespace for maximum data density in QR code
- json_str = json.dumps(data, ensure_ascii=False, separators=(',', ':'))
- total_length = len(json_str)
-
- # Calculate overhead for metadata
- # Metadata structure: {"idx":0,"tc":1,"tl":XXX,"hash":"YYYY","data":"..."}, shortened keys
- metadata_template = {
- "idx": 0, # chunk_index
- "tc": 1, # total_chunks
- "tl": total_length, # total_length
- "hash": "", # chunk_hash
- "data": "" # chunk_data
+ issue = self.issues[issue_id]
+ model_id = DEFAULT_IDLE_MODEL_ID # Use cheap model
+ logger.info(f"Idle Task: Generating preliminary analysis for issue {issue_id} using {model_id}")
+ start_time = time.time()
+
+ prompt = f"""Analyze the GitHub issue below. Provide a single, concise sentence hypothesizing the root cause OR the main goal. Start with "Hypothesis:". If unsure, respond ONLY with "Hypothesis: Further investigation needed.".
+Issue Title: {issue.get('title', 'N/A')}
+Issue Body:
+{issue.get('body', 'N/A')[:1500]}
+Response:"""
+
+ api_url = f"{HF_INFERENCE_API}/{model_id}"
+ headers = {"Authorization": f"Bearer {self.hf_token}"}
+ payload = {
+ "inputs": prompt,
+ "parameters": {"max_new_tokens": 80, "temperature": 0.3, "return_full_text": False, "do_sample": False},
+ "options": {"wait_for_model": True}
}
- # Estimate overhead more accurately by dumping a sample metadata structure
- # and adding some safety margin. Shortened keys reduce overhead.
- overhead_estimate = len(json.dumps(metadata_template, separators=(',', ':'))) + 50 # Extra padding
-
- # Calculate effective chunk size
- effective_chunk_size = max_size - overhead_estimate
-
- if effective_chunk_size <= 0:
- logger.error(f"Max QR size ({max_size}) is too small for metadata overhead ({overhead_estimate}). Cannot chunk.")
- return []
-
- if total_length <= effective_chunk_size:
- # Data fits in one chunk
- chunk_data = json_str # Use the full string
-
- chunk = {
- "idx": 0,
- "tc": 1,
- "tl": total_length,
- "hash": hash(chunk_data) & 0xFFFFFFFF, # 32-bit hash
- "data": chunk_data
- }
- return [chunk]
-
- # Calculate number of chunks needed
- num_chunks = -(-total_length // effective_chunk_size) # Ceiling division
- # Adjust chunk_size slightly to distribute evenly, maybe not strictly necessary
- # chunk_size = -(-total_length // num_chunks) # Use this if perfect distribution is needed
-
- chunks = []
- current_pos = 0
- for i in range(num_chunks):
- # Find the end of the current chunk. Avoid splitting in the middle of escaped characters or surrogate pairs if possible,
- # but simple slicing is usually okay for standard text that's already been errors='replace'.
- # We'll use basic slicing for simplicity, as the JSON string is just text.
- end_pos = min(current_pos + effective_chunk_size, total_length)
-
- # Basic attempt to not break in the middle of a UTF-8 character if slicing bytes,
- # but since we are slicing a *decoded string*, this is less of an issue.
- # However, slicing in the middle of JSON structure is bad.
- # For simplicity and robustness with arbitrary JSON structures, slicing the raw string is the easiest.
- chunk_data_str = json_str[current_pos:end_pos]
-
- chunk = {
- "idx": i,
- "tc": num_chunks,
- "tl": total_length,
- "hash": hash(chunk_data_str) & 0xFFFFFFFF,
- "data": chunk_data_str
- }
- chunks.append(chunk)
- current_pos = end_pos
+ timeout = aiohttp.ClientTimeout(total=45)
- # Final check: Ensure all data was chunked
- if current_pos < total_length:
- # This shouldn't happen with correct ceiling division and min()
- logger.error(f"Chunking logic error: Only processed {current_pos} of {total_length} characters.")
- return [] # Indicate failure
-
-
- logger.info(f"Chunked data into {num_chunks} chunks for QR codes.")
- return chunks
+ retries = 2
+ for attempt in range(retries):
+ try:
+ async with aiohttp.ClientSession(headers=headers, timeout=timeout) as session:
+ async with session.post(api_url, json=payload) as response:
+ if response.status == 429:
+ retry_after = int(response.headers.get('Retry-After', 5))
+ logger.warning(f"Idle Analysis API rate limited. Waiting {retry_after}s before retry {attempt + 1}/{retries}.")
+ await asyncio.sleep(retry_after)
+ continue
+ response.raise_for_status()
+ result = await response.json()
+ duration = time.time() - start_time
+
+ if result and isinstance(result, list) and 'generated_text' in result[0]:
+ hypothesis = result[0].get('generated_text', '').strip()
+ if not hypothesis.lower().startswith("hypothesis:"):
+ hypothesis = "Hypothesis: (Analysis unclear)"
+ elif len(hypothesis) < 15:
+ hypothesis = "Hypothesis: (Analysis failed or too short)"
+
+ self.precomputed_analysis[issue_id] = {"hypothesis": hypothesis, "error": None, "timestamp": time.time()}
+ logger.info(f"Stored preliminary analysis for issue {issue_id} (took {duration:.2f}s): '{hypothesis[:60]}...'")
+ break # Success
+ elif isinstance(result, dict) and 'error' in result:
+ raise ValueError(f"API Error: {result['error']}")
+ else:
+ raise ValueError(f"Unexpected API response format: {str(result)[:200]}")
+ except (asyncio.TimeoutError, aiohttp.ClientResponseError, ValueError) as e:
+ logger.warning(f"Idle Analysis API error for issue {issue_id}: {e}. Retry {attempt + 1}/{retries}.", exc_info=False)
+ if attempt < retries - 1:
+ await asyncio.sleep(5)
+ continue
+ else:
+ raise e # Re-raise after max retries
+ except Exception as e:
+ logger.exception(f"Unexpected error in Idle Analysis task for issue {issue_id}. Retry {attempt + 1}/{retries}.")
+ if attempt < retries - 1:
+ await asyncio.sleep(5)
+ continue
+ else:
+ raise e # Re-raise after max retries
+ else: # Executed if loop completes without break
+ raise Exception(f"Max retries ({retries}) failed for issue {issue_id}")
except Exception as e:
- logger.error(f"Error chunking data: {e}")
- return []
-
-def generate_stylish_qr(data: Union[str, Dict],
- filename: str,
- size: int = 10,
- border: int = 4,
- fill_color: str = "#000000",
- back_color: str = "#FFFFFF") -> str:
- """Generate a stylish QR code with enhanced visual appeal"""
- try:
- qr = qrcode.QRCode(
- version=None,
- error_correction=qrcode.constants.ERROR_CORRECT_M, # Increased error correction
- box_size=size,
- border=border
- )
-
- # Add data to QR code
- if isinstance(data, dict):
- # Use compact JSON representation
- qr.add_data(json.dumps(data, ensure_ascii=False, separators=(',', ':')))
+ err_msg = f"Failed preliminary analysis: {e}"
+ logger.error(f"Failed preliminary analysis for issue {issue_id}: {e}", exc_info=False)
+ self.precomputed_analysis[issue_id] = {"error": err_msg, "hypothesis": None, "timestamp": time.time()}
+
+ def start_idle_processing(self):
+ """Starts the background idle processing task if conditions are met."""
+ # Ensure task is created in the correct loop
+ if not self.main_loop.is_running():
+ logger.error("Cannot start idle processing: Main event loop is not running.")
+ return
+
+ if not self.idle_task or self.idle_task.done():
+ if self.hf_token and self.repo: # HF token and repo are minimum requirements
+ self.idle_task = self.main_loop.create_task(self.run_idle_processing())
+ logger.info(f"Started background idle processing task (interval: {self.idle_processing_interval}s).")
+ else:
+ missing = []
+ if not self.hf_token: missing.append("HF Token")
+ if not self.repo: missing.append("Repository")
+ logger.warning(f"Cannot start idle processing: Missing {', '.join(missing)}.")
else:
- qr.add_data(str(data)) # Ensure it's a string
+ logger.debug("Idle processing task already running.")
+
+ def stop_idle_processing(self):
+ """Stops the background idle processing task."""
+ if self.idle_task and not self.idle_task.done():
+ logger.info("Stopping background idle processing task...")
+ self.idle_task.cancel()
+ # Await the task to finish cancellation in an async context if needed
+ self.idle_task = None # Clear the reference
+
+ async def run_idle_processing(self):
+ """Main loop for background analysis, including new idle thinking tasks."""
+ logger.info("Idle processing loop starting.")
+ try:
+ while True:
+ await asyncio.sleep(self.idle_processing_interval)
+
+ # Re-check conditions inside the loop
+ if not self.issues or not self.repo or not self.hf_token:
+ logger.debug("Idle processing skipped: No issues, repo, or HF token.")
+ continue # Continue the loop, just skip the processing cycle
+
+ logger.info(f"--- Starting idle processing cycle ---")
+ cycle_tasks = []
+ start_time_cycle = time.time()
+
+ # 1. Re-clustering (if needed)
+ if self.needs_recluster:
+ logger.info("Idle Task: Scheduling re-clustering and duplicate identification...")
+ # Capture current issues list *before* the async task potentially runs
+ # This snapshot ensures consistency for the clustering process
+ self.issue_list_for_clustering = list(self.issues.values())
+ cycle_tasks.append(self._run_clustering_and_duplicates_async())
+ else:
+ logger.debug("No re-clustering needed in this cycle.")
- qr.make(fit=True)
+ # 2. Identify Stale & High Priority Issues (Sync)
+ # These are quick operations, run synchronously
+ try:
+ self._identify_stale_issues()
+ self._identify_high_priority_candidates()
+ except Exception as e:
+ logger.error(f"Error during synchronous stale/priority identification: {e}")
+
+ # --- Identify Issues Needing Work ---
+ # Get current list of issue IDs
+ current_issue_ids = list(self.issues.keys())
+ # Shuffle to process issues more evenly over time, not just oldest/newest
+ random.shuffle(current_issue_ids)
+
+ # Filter issues that need specific computations
+ # Check if the issue still exists in self.issues before adding to needing list
+ issues_needing_context = [i for i in current_issue_ids if i in self.issues and i not in self.precomputed_context]
+ issues_needing_summary = [i for i in current_issue_ids if i in self.issues and i not in self.precomputed_summaries]
+ issues_needing_missing_info = [i for i in current_issue_ids if i in self.issues and i not in self.precomputed_missing_info]
+ issues_needing_analysis = [i for i in current_issue_ids if i in self.issues and i not in self.precomputed_analysis]
+
+
+ # Prioritize high-priority and stale issues for analysis tasks
+ # Ensure candidates are still in the 'needing' list before adding to ordered lists
+ priority_summary_candidates = [i for i in self.high_priority_candidates if i in issues_needing_summary]
+ stale_summary_candidates = [i for i in self.stale_issues if i in issues_needing_summary and i not in priority_summary_candidates]
+ other_summary_candidates = [i for i in issues_needing_summary if i not in priority_summary_candidates and i not in stale_summary_candidates]
+ ordered_summary_candidates = priority_summary_candidates + stale_summary_candidates + other_summary_candidates
+
+ priority_analysis_candidates = [i for i in self.high_priority_candidates if i in issues_needing_analysis]
+ stale_analysis_candidates = [i for i in self.stale_issues if i in issues_needing_analysis and i not in priority_analysis_candidates]
+ other_analysis_candidates = [i for i in issues_needing_analysis if i not in priority_analysis_candidates and i not in stale_analysis_candidates]
+ ordered_analysis_candidates = priority_analysis_candidates + stale_analysis_candidates + other_analysis_candidates
+
+ priority_info_candidates = [i for i in self.high_priority_candidates if i in issues_needing_missing_info]
+ stale_info_candidates = [i for i in self.stale_issues if i in issues_needing_missing_info and i not in priority_info_candidates]
+ other_info_candidates = [i for i in issues_needing_missing_info if i not in priority_info_candidates and i not in stale_info_candidates]
+ ordered_info_candidates = priority_info_candidates + stale_info_candidates + other_info_candidates
+
+ # Context is needed for patch generation, prioritize high-priority/stale issues
+ priority_context_candidates = [i for i in self.high_priority_candidates if i in issues_needing_context]
+ stale_context_candidates = [i for i in self.stale_issues if i in issues_needing_context and i not in priority_context_candidates]
+ other_context_candidates = [i for i in issues_needing_context if i not in priority_context_candidates and i not in stale_context_candidates]
+ ordered_context_candidates = priority_context_candidates + stale_context_candidates + other_context_candidates
+
+
+ logger.debug(f"Idle candidates: Ctx:{len(issues_needing_context)} ({len(ordered_context_candidates)} ordered), Sum:{len(issues_needing_summary)} ({len(ordered_summary_candidates)} ordered), "
+ f"Info:{len(issues_needing_missing_info)} ({len(ordered_info_candidates)} ordered), Anl:{len(issues_needing_analysis)} ({len(ordered_analysis_candidates)} ordered)")
+
+
+ # 3. Schedule Context Pre-computation (I/O bound)
+ context_computed_count = 0
+ for issue_id in ordered_context_candidates:
+ if context_computed_count < self.max_context_computations_per_cycle:
+ # Check if issue still exists before scheduling
+ if issue_id in self.issues:
+ cycle_tasks.append(self._compute_and_store_context(issue_id))
+ context_computed_count += 1
+ else: break
+ if context_computed_count > 0: logger.info(f"Scheduled {context_computed_count} context computations.")
+
+ # 4. Schedule Summary Generation (LLM - Medium Cost)
+ summary_computed_count = 0
+ for issue_id in ordered_summary_candidates:
+ if summary_computed_count < self.max_summary_computations_per_cycle:
+ if issue_id in self.issues:
+ cycle_tasks.append(self._compute_and_store_summary(issue_id))
+ summary_computed_count += 1
+ else: break
+ if summary_computed_count > 0: logger.info(f"Scheduled {summary_computed_count} summary computations.")
+
+ # 5. Schedule Missing Info Analysis (LLM - Low Cost)
+ missing_info_count = 0
+ for issue_id in ordered_info_candidates:
+ if missing_info_count < self.max_missing_info_computations_per_cycle:
+ if issue_id in self.issues:
+ cycle_tasks.append(self._compute_and_store_missing_info(issue_id))
+ missing_info_count += 1
+ else: break
+ if missing_info_count > 0: logger.info(f"Scheduled {missing_info_count} missing info analyses.")
+
+ # 6. Schedule Preliminary Analysis (LLM - Low Cost)
+ analysis_count = 0
+ for issue_id in ordered_analysis_candidates:
+ if analysis_count < self.max_analysis_computations_per_cycle:
+ if issue_id in self.issues:
+ cycle_tasks.append(self._compute_and_store_preliminary_analysis(issue_id))
+ analysis_count += 1
+ else: break
+ if analysis_count > 0: logger.info(f"Scheduled {analysis_count} preliminary analyses.")
+
+ # --- Execute Scheduled Async Tasks ---
+ if cycle_tasks:
+ logger.info(f"Executing {len(cycle_tasks)} async idle tasks for this cycle...")
+ # Run tasks concurrently
+ # Use return_exceptions=True to ensure all tasks are attempted even if one fails
+ results = await asyncio.gather(*cycle_tasks, return_exceptions=True)
+ num_errors = 0
+ for i, result in enumerate(results):
+ if isinstance(result, Exception):
+ num_errors += 1
+ # Log the error, but gather ensures other tasks complete
+ logger.error(f"Error encountered in background idle task {i+1}/{len(cycle_tasks)}: {result}", exc_info=False) # Keep log cleaner
+ cycle_duration = time.time() - start_time_cycle
+ logger.info(f"Idle processing cycle finished in {cycle_duration:.2f} seconds. {len(results)} tasks processed ({num_errors} errors).")
+ else:
+ logger.info("No async idle tasks to perform in this cycle.")
+ logger.info(f"--- Finished idle processing cycle ---")
+
+ except asyncio.CancelledError:
+ logger.info("Idle processing loop cancelled.")
+ except Exception as e:
+ logger.exception(f"Critical error in idle processing loop: {e}")
+ # Wait longer before potentially restarting after a critical error
+ await asyncio.sleep(self.idle_processing_interval * 5)
+ finally:
+ logger.info("Idle processing loop finished.")
+
+ async def _compute_and_store_context(self, issue_id: int):
+ """Helper async task to compute and store context for one issue during idle time."""
+ # Re-check if issue exists just before starting computation
+ if issue_id not in self.issues:
+ logger.warning(f"Skipping context computation for issue {issue_id}: Issue no longer exists.")
+ return
+ try:
+ logger.debug(f"Starting background context computation for issue {issue_id}...")
+ start_time = time.time()
+ issue_data = self.issues[issue_id] # Get latest issue data
+ context_result = await self._get_code_context(issue_data)
+ duration = time.time() - start_time
+
+ computed_data = {
+ "content": context_result.get("content"),
+ "files": context_result.get("files", []),
+ "error": context_result.get("error"),
+ "timestamp": time.time()
+ }
+ # Store the result even if there was an error during computation
+ self.precomputed_context[issue_id] = computed_data
- # Create QR code image with custom colors
- qr_image = qr.make_image(fill_color=fill_color, back_color=back_color)
+ log_msg = f"Stored context result for issue {issue_id} (found {len(computed_data['files'])} files, took {duration:.2f}s)."
+ if computed_data['error']:
+ log_msg += f" Error: {computed_data['error']}"
+ logger.info(log_msg)
- # Convert to RGBA for transparency support and potential overlays
- qr_image = qr_image.convert('RGBA')
+ except Exception as e:
+ logger.exception(f"Failed to compute context for issue {issue_id} in background task: {e}")
+ # Store an error state in precomputed_context
+ self.precomputed_context[issue_id] = {"error": f"Unexpected computation error: {e}", "timestamp": time.time(), "content": None, "files": []}
- # Optional: Add a small logo or icon in the center (requires design)
- # logo = Image.open("logo.png").convert("RGBA")
- # logo = logo.resize((logo.width // 4, logo.height // 4)) # Resize logo
- # logo_pos = ((qr_image.width - logo.width) // 2, (qr_image.height - logo.height) // 2)
- # qr_image.paste(logo, logo_pos, logo)
- # Add subtle gradient overlay (optional visual enhancement)
+ async def _run_clustering_and_duplicates_async(self):
+ """Runs clustering and then identifies duplicates as a single background task unit."""
try:
- gradient = Image.new('RGBA', qr_image.size, (0, 0, 0, 0))
- draw = ImageDraw.Draw(gradient)
- # Horizontal gradient for subtle effect
- for i in range(qr_image.width):
- # Fades from left (alpha=0) to right (max_alpha)
- alpha = int(255 * (i/qr_image.width) * 0.05) # e.g., 5% maximum opacity fade-in
- draw.line([(i, 0), (i, qr_image.height)], fill=(0, 0, 0, alpha))
- # Combine images
- final_image = Image.alpha_composite(qr_image, gradient)
+ logger.info("Background Task: Starting re-clustering issues...")
+ start_time = time.time()
+ # _cluster_similar_issues uses self.issue_list_for_clustering (snapshot taken before scheduling)
+ await self._cluster_similar_issues() # This resets flags on success internally
+ cluster_duration = time.time() - start_time
+ if self.issue_clusters:
+ logger.info(f"Background Task: Clustering finished in {cluster_duration:.2f}s. Identifying duplicates...")
+ start_time_dup = time.time()
+ # _identify_potential_duplicates uses the results of _cluster_similar_issues
+ self._identify_potential_duplicates()
+ dup_duration = time.time() - start_time_dup
+ logger.info(f"Background Task: Duplicate identification finished in {dup_duration:.2f}s.")
+ else:
+ logger.warning("Background Task: Clustering did not produce results. Skipping duplicate identification.")
+ # Flags are already reset by _cluster_similar_issues on failure/empty
+
except Exception as e:
- logger.warning(f"Failed to add gradient overlay to QR code: {e}. Using plain QR.")
- final_image = qr_image
+ logger.error(f"Error during background clustering/duplicate identification task: {e}", exc_info=True)
- # Save the image
- output_path = QR_CODES_DIR / filename
- final_image.save(output_path, quality=90) # Save with slightly lower quality for smaller file size
+# ========== Gradio UI Definition ==========
+def create_ui(manager: IssueManager) -> gr.Blocks:
+ """Creates the Gradio interface."""
- return str(output_path)
- except Exception as e:
- logger.error(f"QR generation error: {e}")
- return ""
+ # --- Helper Functions for UI ---
+ def generate_issue_preview(issue_num: Optional[int]) -> str:
+ """Generates HTML preview including new idle thinking results."""
+ if issue_num is None or issue_num not in manager.issues:
+ return "Select an issue from the 'Issue Board' tab.
"
-def generate_qr_codes(data: Union[str, Dict, List], combined: bool = True) -> List[str]:
- """Generate QR codes with enhanced visual appeal and metadata"""
- # Assume 'data' here is the list of dictionaries produced by process_inputs
- if not isinstance(data, list):
- logger.error("generate_qr_codes received data that is not a list.")
- return []
+ try:
+ issue = manager.issues[issue_num]
+ # Use markdown2 for rendering issue body
+ html_body = markdown2.markdown(
+ issue.get('body', '*No description provided.*') or '*No description provided.*',
+ extras=["fenced-code-blocks", "tables", "strike", "task_list", "code-friendly", "html-classes", "nofollow", "spoiler"]
+ )
- try:
- file_processor = EnhancedFileProcessor() # Use the enhanced processor for chunking
- paths = []
-
- if combined:
- # Process combined data
- chunks = file_processor.chunk_data(data) # chunk_data works on the list of dicts
- if not chunks:
- logger.warning("No chunks generated for combined data.")
- return []
- for i, chunk in enumerate(chunks):
- filename = f'combined_qr_{int(time.time())}_{i+1}_of_{len(chunks)}.png'
- qr_path = generate_stylish_qr(
- data=chunk, # Pass the chunk dictionary
- filename=filename,
- fill_color="#1a365d", # Deep blue
- back_color="#ffffff"
- )
- if qr_path:
- paths.append(qr_path)
- else:
- logger.warning(f"Failed to generate QR for chunk {i+1}/{len(chunks)}.")
- else:
- # Process individual items (each dictionary in the list)
- if data: # Ensure data is not empty
- for idx, item in enumerate(data):
- chunks = file_processor.chunk_data(item) # chunk_data works on individual dict
- if not chunks:
- logger.warning(f"No chunks generated for item {idx+1}.")
- continue
- for chunk_idx, chunk in enumerate(chunks):
- filename = f'item_{idx+1}_chunk_{chunk_idx+1}_of_{len(chunks)}_{int(time.time())}.png'
- qr_path = generate_stylish_qr(
- data=chunk, # Pass the chunk dictionary
- filename=filename,
- fill_color="#1a365d", # Deep blue
- back_color="#ffffff"
- )
- if qr_path:
- paths.append(qr_path)
- else:
- logger.warning(f"Failed to generate QR for item {idx+1} chunk {chunk_idx+1}/{len(chunks)}.")
- else:
- logger.warning("No items in data list to process individually.")
+ # Generate HTML for labels
+ labels_html = ' '.join(f'{gr.Textbox.sanitize_html(l)}' for l in issue.get('labels', [])) or 'None'
+
+ # Generate HTML for status indicators
+ status_indicators = []
+ if issue_num in manager.stale_issues:
+ status_indicators.append(f"[Stale]")
+ if issue_num in manager.high_priority_candidates:
+ severity = manager._determine_severity(issue.get('labels', []))
+ if severity == "Critical": color, bgcolor = "#ef4444", "#fee2e2"
+ elif severity == "High": color, bgcolor = "#f97316", "#ffedd5"
+ else: color, bgcolor = "#c2410c", "#fffbeb" # Should ideally not happen if logic works
+ status_indicators.append(f"[{severity}]")
+ status_html = " ".join(status_indicators)
+
+ # --- Get Precomputed Data with Helper ---
+ summary_text = manager._get_precomputed_text(issue_num, manager.precomputed_summaries, "summary", "Summary")
+ missing_info_text = manager._get_precomputed_text(issue_num, manager.precomputed_missing_info, "info_needed", "Missing Info")
+ analysis_text = manager._get_precomputed_text(issue_num, manager.precomputed_analysis, "hypothesis", "Analysis")
+ duplicate_text = manager._get_duplicate_info_text(issue_num) # This already includes formatting and links
+
+ # --- Format AI Sections ---
+ ai_sections = []
+ # Only show sections if they have content or an error, not just pending state
+ # Use the presence of icons (❌, ⏳, ℹ️) to determine if it's just a status message
+ if not summary_text.startswith("⏳") and not summary_text.startswith("(ℹ️"):
+ color = "#f0e6ff" if not summary_text.startswith("❌") else "#fee2e2"
+ border_color = "#ddd6fe" if not summary_text.startswith("❌") else "#fecaca"
+ ai_sections.append(f"""
+
+ 🤖 AI Summary: {summary_text}
+
""")
+ elif summary_text.startswith("❌"):
+ ai_sections.append(f"""
+
+ 🤖 {summary_text}
+
""")
+
+
+ if not missing_info_text.startswith("⏳") and not missing_info_text.startswith("(ℹ️"):
+ color = "#fffbeb" if "None needed." not in missing_info_text and not missing_info_text.startswith("❌") else "#f0fdf4" if "None needed." in missing_info_text else "#fee2e2"
+ border_color = "#fef3c7" if "None needed." not in missing_info_text and not missing_info_text.startswith("❌") else "#bbf7d0" if "None needed." in missing_info_text else "#fecaca"
+ ai_sections.append(f"""
+
+ 🤔 AI Missing Info Analysis: {missing_info_text}
+
""")
+ elif missing_info_text.startswith("❌"):
+ ai_sections.append(f"""
+
+ 🤔 {missing_info_text}
+
""")
+
+
+ if not analysis_text.startswith("⏳") and not analysis_text.startswith("(ℹ️"):
+ color = "#e0f2fe" if not analysis_text.startswith("❌") else "#fee2e2"
+ border_color = "#bae6fd" if not analysis_text.startswith("❌") else "#fecaca"
+ ai_sections.append(f"""
+
+ 🔬 AI Preliminary Analysis: {analysis_text}
+
""")
+ elif analysis_text.startswith("❌"):
+ ai_sections.append(f"""
+
+ 🔬 {analysis_text}
+
""")
+
+
+ if duplicate_text: # duplicate_text already contains HTML links if duplicates exist
+ ai_sections.append(f"""
+
+ {duplicate_text}
+
""")
+
+ ai_sections_html = "\n".join(ai_sections)
+
+ # Construct the full preview HTML
+ preview_html = f"""
+
+
+
+
+ State: {issue.get('state', 'N/A')}
+ Assignee: {issue.get('assignee') or 'None'}
+ Severity: {manager._determine_severity(issue.get('labels', []))}
+
+
+ Labels: {labels_html}
+
+
+ {ai_sections_html}
+
+
+
Description:
+ {html_body}
+
+
Error generating preview for issue {issue_num}. Check logs.
"
+
+ async def get_ai_suggestion_wrapper(issue_num: Optional[int], model_key: str, progress=gr.Progress()) -> str:
+ """UI wrapper for getting AI suggestions, handles state and progress."""
+ progress(0, desc="Preparing request...")
+ if issue_num is None or issue_num not in manager.issues:
+ return "⚠️ Error: Please select a valid issue first."
+ if not manager.hf_token:
+ return "🔒 Error: Hugging Face Token is not configured."
+ if model_key not in HF_MODELS or HF_MODELS.get(model_key) is None:
+ return f"⚠️ Error: Invalid or unconfigured model key selected: {model_key}"
+
+ issue = manager.issues[issue_num]
+ issue_hash = manager._get_issue_hash(issue)
+ logger.info(f"Requesting suggestion for issue {issue_num} (hash: {issue_hash}) using model {model_key}.")
+
+ try:
+ progress(0.3, desc=f"Querying {model_key}...")
+ # The cached_suggestion method handles the cache lookup and actual generation
+ suggestion = await manager.cached_suggestion(issue_hash, model_key)
+ progress(1, desc="Suggestion received.")
+
+ if suggestion.lower().startswith("error:"):
+ return f"⚠️ {suggestion}"
+ else:
+ # Format the output clearly
+ return f"**💡 Suggestion based on {model_key}:**\n\n---\n{suggestion}"
+ except Exception as e:
+ logger.exception(f"Error in get_ai_suggestion_wrapper for issue {issue_num}: {e}")
+ return f"❌ An unexpected error occurred while getting the suggestion: {e}"
+
+ async def get_ai_patch_wrapper(issue_num: Optional[int], model_key: str, progress=gr.Progress()) -> str:
+ """UI wrapper for getting AI patches, handles state and progress."""
+ progress(0, desc="Preparing request...")
+ if issue_num is None or issue_num not in manager.issues:
+ return "⚠️ Error: Please select a valid issue first."
+ if not manager.hf_token:
+ return "🔒 Error: Hugging Face Token is not configured."
+ if not manager.repo:
+ return "❌ Error: Repository not loaded. Please scan the repository first."
+ if model_key not in HF_MODELS or HF_MODELS.get(model_key) is None:
+ return f"⚠️ Error: Invalid or unconfigured model key selected: {model_key}"
+
+ logger.info(f"Requesting patch for issue {issue_num} using model {model_key}.")
+ progress(0.1, desc="Gathering code context (using cache if available)...")
+ try:
+ # Context is gathered inside generate_code_patch, which uses the precomputed_context cache
+ progress(0.4, desc=f"Querying {model_key} for patch...")
+ result = await manager.generate_code_patch(issue_num, model_key)
+ progress(1, desc="Patch result received.")
+
+ if "error" in result:
+ logger.error(f"Patch generation failed for issue {issue_num}: {result['error']}")
+ return f"**❌ Error generating patch:**\n\n{result['error']}"
+ else:
+ model_used = result.get('model_used', model_key)
+ explanation = result.get('explanation', '(No explanation provided)')
+ patch_content = result.get('patch')
+ status_msg = result.get('status') # Get specific status if available
+
+ header = f"**🩹 Patch Suggestion from {model_used}:**"
+ if status_msg:
+ header = f"**🩹 Patch Generation Result from {model_used}:** ({status_msg})"
+
+ if patch_content:
+ # Escape backticks in patch content for markdown code block
+ patch_content_sanitized = patch_content.replace('`', '\\`')
+ logger.info(f"Successfully generated patch for issue {issue_num} using {model_used}.")
+ return f"""{header}
+**Explanation:**
+{explanation}
+---
+**Patch:**
+```diff
+{patch_content_sanitized}
+```"""
+ else:
+ logger.warning(f"AI provided explanation but no patch for issue {issue_num}. Explanation: {explanation}")
+ return f"""{header}
+**Explanation:**
+{explanation}
+---
+**(No valid diff block generated)**"""
+
+ except Exception as e:
+ logger.exception(f"Error in get_ai_patch_wrapper for issue {issue_num}: {e}")
+ return f"❌ An unexpected error occurred while generating the patch: {e}"
+
+ async def handle_issue_select(evt: gr.SelectData):
+ """
+ Handles issue selection in the Dataframe: updates preview, loads code context into editor.
+ Reads the selected issue ID from the event data.
+ Returns updates for multiple components.
+ """
+ # Default state for deselection or error
+ default_updates = {
+ # Update the hidden state element first
+ "selected_issue_id_hidden": gr.update(value=""),
+ "issue_preview_html": gr.update(value="Select an issue from the 'Issue Board' tab.
"),
+ # Clear the code editor
+ "code_edit_component": gr.update(value={"placeholder.txt": "# Select an issue to load relevant code context."}, interactive=True, language="text"),
+ "ai_output_display": gr.update(value="*AI suggestions and patches will appear here after selecting an issue and action.*"),
+ "copy_patch_btn": gr.update(visible=False), # Hide copy button on selection change
+ }
- logger.info(f"Generated {len(paths)} QR codes.")
- return paths
+ # Check if a row was actually selected and has a value
+ if evt.index is None or not hasattr(evt, 'value') or not evt.value or evt.value[0] is None:
+ logger.info("Issue deselected or invalid selection event.")
+ # Return default updates to clear the UI
+ return default_updates
- except Exception as e:
- logger.error(f"QR code generation error: {e}")
- return []
-
-# Keep the Gradio UI definition and main function as they are,
-# as the changes are internal to the processing classes and the
-# process_inputs function already handles calling them and getting
-# the combined list of results.
-
-def create_modern_interface():
- """Create a modern and visually appealing Gradio interface"""
-
- # Modern CSS styling
- css = """
- /* Modern color scheme */
- :root {
- --primary-color: #1a365d;
- --secondary-color: #2d3748;
- --accent-color: #4299e1;
- --background-color: #f7fafc;
- --success-color: #48bb78;
- --error-color: #f56565;
- --warning-color: #ed8936;
- }
- /* Container styling */
- .container {
- max-width: 1200px;
- margin: auto;
- padding: 2rem;
- background-color: var(--background-color);
- border-radius: 1rem;
- box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
- }
- /* Component styling */
- .input-container {
- background-color: white;
- padding: 1.5rem;
- border-radius: 0.5rem;
- border: 1px solid #e2e8f0;
- margin-bottom: 1rem;
- }
- /* Button styling */
- .primary-button {
- background-color: var(--primary-color);
- color: white;
- padding: 0.75rem 1.5rem;
- border-radius: 0.375rem;
- border: none;
- cursor: pointer;
- transition: all 0.2s;
- }
- .primary-button:hover {
- background-color: var(--accent-color);
- transform: translateY(-1px);
- }
- /* Status messages */
- .status {
- padding: 1rem;
- border-radius: 0.375rem;
- margin: 1rem 0;
- }
- .status.success { background-color: #f0fff4; color: var(--success-color); }
- .status.error { background-color: #fff5f5; color: var(--error-color); }
- .status.warning { background-color: #fffaf0; color: var(--warning-color); }
- /* Gallery styling */
- .gallery {
- display: grid;
- grid-template-columns: repeat(auto-fit, minmax(200px, 1fr));
- gap: 1rem;
- padding: 1rem;
- background-color: white;
- border-radius: 0.5rem;
- border: 1px solid #e2e8f0;
- }
- .gallery img {
- width: 100%;
- height: auto;
- border-radius: 0.375rem;
- transition: transform 0.2s;
- }
- .gallery img:hover {
- transform: scale(1.05);
- }
- /* QR Code Viewport Styling */
- .viewport-container {
- display: grid;
- gap: 0.5rem;
- padding: 1rem;
- background-color: white;
- border-radius: 0.5rem;
- border: 1px solid #e2e8f0;
- margin-top: 1rem;
- }
- .viewport-item {
- display: flex;
- flex-direction: column;
- align-items: center;
- }
- .viewport-item img {
- width: 100%;
- height: auto;
- border-radius: 0.375rem;
- transition: transform 0.2s;
- max-width: 150px; /* Adjust as needed */
- max-height: 150px; /* Adjust as needed */
- }
- """
- # Create interface with modern design
- with gr.Blocks(css=css, title="Advanced Data Processor & QR Generator") as interface:
- interface.head += """
-
- """
- qr_code_paths = gr.State([])
+
+ # --- Analytics Helper ---
+ def update_cluster_analytics(mgr: IssueManager):
+ """Generates data for the cluster analytics dataframe."""
+ if not mgr.issue_clusters or not mgr.issue_list_for_clustering:
+ return [["N/A", 0, "No clusters found"]]
+
+ cluster_data = []
+ # Expanded stop words list (lowercase)
+ stop_words = set("a an the is are was were be been being has have had do does did will would shall should can could may might must it its this that these those there their then than and or but if because as at by for with without about against between into through during before after above below to from up down in out on off over under again further then once here there when where why how all any both each few more most other some such no nor not only own same so than too very s t can will just don should now d ll m o re ve yo ain aren couldn didn doesn isn it's mustn ouren shan shouldn wasn weren won wouldn".split())
+ # Add common code/issue terms (lowercase)
+ stop_words.update({'issue', 'bug', 'error', 'fix', 'feat', 'chore', 'refactor', 'docs', 'test', 'file', 'line', 'code', 'when', 'user', 'report', 'problem', 'need', 'want', 'get', 'use', 'try', 'make', 'add', 'remove', 'change', 'update', 'create', 'build', 'run', 'start', 'stop', 'click', 'button', 'page', 'app', 'data', 'system', 'service', 'component', 'module', 'function', 'class', 'method', 'variable', 'value', 'string', 'number', 'list', 'dict', 'object', 'array', 'true', 'false', 'null', 'none', 'type', 'size', 'color', 'style', 'width', 'height', 'margin', 'padding', 'border', 'display', 'position', 'float', 'clear', 'overflow', 'index', 'key', 'value', 'id', 'name', 'text', 'html', 'css', 'js', 'py', 'java', 'c', 'cpp', 'h', 'hpp', 'go', 'rs', 'php', 'rb', 'json', 'yaml', 'xml', 'sql', 'log', 'config', 'setup', 'install', 'version', 'release', 'branch', 'commit', 'pull', 'request', 'merge', 'diff', 'patch', 'context', 'suggestion', 'analysis', 'missing', 'information', 'duplicate', 'potential', 'cluster', 'severity', 'label', 'assignee', 'state', 'title', 'body', 'url', 'created', 'updated', 'time', 'day', 'week', 'month', 'year', 'ago', 'now', 'new', 'old', 'different', 'same', 'similar', 'group', 'count', 'distribution', 'analytics', 'overview', 'studio', 'board', 'resolution', 'collaborative', 'editor', 'context', 'aware', 'assistance', 'tools', 'output', 'patch', 'steps', 'approach', 'plan', 'identify', 'propose', 'recommendation', 'workflow', 'standard', 'important', 'critical', 'high', 'medium', 'low', 'unknown', 'none', 'needed', 'analysis', 'hypothesis', 'preliminary', 'relevant', 'area', 'investigation', 'implementation', 'testing', 'next'})
+
+
+ def get_top_keywords(indices):
+ text = ""
+ for idx in indices:
+ # Ensure index is valid for the current issue list snapshot
+ if 0 <= idx < len(mgr.issue_list_for_clustering):
+ issue = mgr.issue_list_for_clustering[idx]
+ text += issue.get('title', '') + " " + issue.get('body', '')[:500] + " " # Use more body text
+ text = text.lower()
+ # Remove punctuation
+ text = text.translate(str.maketrans('', '', string.punctuation))
+ # Split into words and filter
+ words = [w for w in text.split() if len(w) > 2 and w not in stop_words] # Min word length 3
+ if not words: return "N/A"
+ counts = Counter(words)
+ # Get top 5 most common words
+ top_5 = [word for word, count in counts.most_common(5)]
+ return ", ".join(top_5)
+
+ # Sort clusters by size (descending)
+ sorted_clusters = sorted(mgr.issue_clusters.items(), key=lambda item: len(item[1]), reverse=True)
+
+ # Limit to top N clusters for display
+ top_n_clusters = 10 # Display top 10 clusters
+ cluster_data = []
+ for cluster_id, indices in sorted_clusters[:top_n_clusters]:
+ keywords = get_top_keywords(indices)
+ cluster_data.append([cluster_id, len(indices), keywords])
+
+ if not cluster_data:
+ return [["N/A", 0, "No clusters found"]]
+
+ return cluster_data
+
+ # --- Gradio Blocks ---
+ with gr.Blocks(theme=theme, title="AI Issue Resolver Pro", css="""
+ #collab-list .collab-item { margin-bottom: 4px; font-size: 0.9em; }
+ .gradio-container { max-width: 1600px !important; }
+ #issue_preview_div { max-height: 80vh; overflow-y: auto; } /* Ensure preview is scrollable */
+ #ai_output_md { max-height: 60vh; overflow-y: auto; } /* Ensure AI output is scrollable */
+ #code_editor_component { min-height: 500px; height: 70vh; } /* Give editor more space */
+ #ai_tools_accordion > .label { background-color: #f0f9ff; } /* Light blue background for tools header */
+ #ai_tools_accordion.closed > .label { background-color: #f0f9ff; } /* Keep background when closed */
+ #ai_tools_accordion > .label:hover { background-color: #e0f2fe; } /* Hover effect */
+ #ai_tools_accordion > .label > .icon { color: #0ea5e9; } /* Tool icon color */
+ .panel { border: 1px solid #e5e7eb; border-radius: 8px; padding: 15px; background-color: #f9fafb; } /* Styled panel */
+ #config-panel { margin-bottom: 20px; }
+ #status_output_txt textarea { font-family: monospace; font-size: 0.9em; background-color: #eef; } /* Style status bar */
+ .gradio-dataframe-container { overflow-x: auto; } /* Ensure dataframe is scrollable horizontally */
+ #issue_list_df table { width: 100%; } /* Make dataframe table use full width */
+ #issue_list_df th, #issue_list_df td { white-space: nowrap; overflow: hidden; text-overflow: ellipsis; } /* Prevent text wrapping in table headers/cells */
+ #issue_list_df td:nth-child(1) { width: 80px; min-width: 80px; } /* Fix ID column width */
+ #issue_list_df td:nth-child(2) { width: 60%; max-width: 400px; } /* Give Title column more space */
+ #issue_list_df td:nth-child(3) { width: 15%; min-width: 100px; } /* Severity */
+ #issue_list_df td:nth-child(4) { width: 15%; min-width: 80px; } /* Cluster */
+ .gradio-dataframe-container table td { cursor: pointer; } /* Indicate rows are clickable */
+
+
+ """) as demo_app:
gr.Markdown("""
- # 🌐 Advanced Data Processing & QR Code Generator
- Transform your data into beautifully designed, sequenced QR codes with our cutting-edge processor.
+
+
🚀 AI Issue Resolver Pro
+
Collaborative Issue Resolution Powered by AI
+
+
👥 Active Collaborators
+
+ Connecting...
+
+
Select an issue from the 'Issue Board' tab.
", elem_id="issue_preview_div")
+
+ with gr.Accordion("🛠️ AI Assistance Tools", open=True, elem_id="ai_tools_accordion"):
+ suggest_btn = gr.Button("🧠 Suggest Resolution Steps", icon="💡", elem_id="suggest_btn", interactive=bool(HF_MODELS))
+ patch_btn = gr.Button("📝 Generate Code Patch", icon="🩹", elem_id="patch_btn", interactive=bool(HF_MODELS))
+
+ gr.Markdown("### AI Output")
+ # Use a Markdown component with scrollability
+ ai_output_display = gr.Markdown(value="*AI suggestions and patches will appear here...*", elem_id="ai_output_md")
+ with gr.Row():
+ copy_patch_btn = gr.Button("📋 Copy Patch", elem_id="copy_patch_btn", visible=False) # Initially hidden
+ clear_ai_output_btn = gr.Button("🧹 Clear Output", elem_id="clear_ai_output_btn")
+
+
+ with gr.Column(scale=2, min_width=600):
+ gr.Markdown("### Collaborative Code Editor (Context-Aware)")
+ gr.Markdown("⚠️ Warning: Real-time collaborative editing is experimental and may lose data with simultaneous edits. Use with caution and save work frequently!
")
+ # Initialize with placeholder content
+ code_edit_component = code_editor(
+ label="Code Context / Editor",
+ language="python", # Default language, can be changed based on file extension in JS
+ interactive=True,
+ elem_id="code_editor_component",
+ value={"placeholder.txt": "# Select an issue to load relevant code context."}
+ )
+
+ with gr.Tab("📈 Analytics", id="analytics", elem_id="tab-analytics"):
+ gr.Markdown("### Repository Analytics")
+ with gr.Row(equal_height=False):
+ with gr.Column(scale=1):
+ gr.Markdown("#### Issue Severity Distribution")
+ # Initialize with empty plot, will be updated after crawl
+ analytics_severity_plot = gr.Plot(label="Severity Distribution (Analytics)", elem_id="analytics_severity_plot", value=manager._generate_stats_plot({}))
+ with gr.Column(scale=1):
+ gr.Markdown("#### Issue Cluster Analysis (Top Clusters)")
+ cluster_info_df = gr.Dataframe(
+ headers=["Cluster ID", "Issue Count", "Top Keywords (Example)"],
+ datatype=["number", "number", "str"],
+ value=[["Scan a repository to see cluster data.", 0, ""]], # Initial placeholder data
+ label="Issue Clusters", elem_id="cluster_info_df",
+ interactive=False, # Make this dataframe non-interactive
+ row_count=(5, "dynamic"),
+ col_count=(3, "fixed"),
+ wrap=True
+ )
+ gr.Markdown("*(Analytics update after scanning the repository. More detailed analytics could be added.)*")
+
+ # --- Event Handlers ---
+ # The crawl button updates the issue list, stats plot, status, and analytics plot
+ crawl_btn.click(
+ fn=manager.crawl_issues,
+ inputs=[repo_url, github_token, hf_token],
+ outputs=[issue_list, stats_plot, status_output, analytics_severity_plot],
+ api_name="crawl_issues",
+ show_progress="full"
+ ).then(
+ # After crawl_issues completes, update the cluster analytics dataframe
+ fn=lambda: update_cluster_analytics(manager),
+ inputs=[],
+ outputs=[cluster_info_df]
)
- output_text = gr.Textbox(
- label="Processing Status",
- interactive=False
+
+ # The issue list selection updates the preview, code editor, AI output area, and hidden state
+ issue_list.select(
+ fn=handle_issue_select,
+ # Pass the event data, which contains the selected row's value (including ID)
+ inputs=[gr.SelectData()],
+ outputs=[selected_issue_id_hidden, issue_preview_html, code_edit_component, ai_output_display, copy_patch_btn],
+ show_progress="minimal",
+ # Trigger the JS function to update the tracked issue ID and editor listeners
+ # This is handled by the MutationObserver and reading the hidden input value in JS
)
- with gr.Tab("🖼️ QR Code Viewport") as viewport_tab:
- viewport_output = gr.HTML(label="QR Code Sequence Viewport")
- enabled_qr_codes = gr.State([]) # To store the enabled/disabled state
-
- # Load example data
- def load_example():
- example = {
- "type": "product_catalog",
- "items": [
- {
- "id": "123",
- "name": "Premium Widget",
- "description": "High-quality widget with advanced features",
- "price": 299.99,
- "category": "electronics",
- "tags": ["premium", "featured", "new"]
- },
- {
- "id": "456",
- "name": "Basic Widget",
- "description": "Reliable widget for everyday use",
- "price": 149.99,
- "category": "electronics",
- "tags": ["basic", "popular"]
- }
- ],
- "metadata": {
- "timestamp": datetime.now().isoformat(),
- "version": "2.0",
- "source": "example"
- }
- }
- return json.dumps(example, indent=2)
+ # AI Suggestion button
+ suggest_btn.click(
+ fn=get_ai_suggestion_wrapper,
+ inputs=[selected_issue_id_hidden, model_select], # Read selected issue ID from hidden state
+ outputs=[ai_output_display],
+ api_name="suggest_resolution",
+ show_progress="full"
+ ).then(
+ # After getting suggestion, hide the copy patch button
+ fn=lambda: gr.update(visible=False),
+ inputs=[],
+ outputs=[copy_patch_btn]
+ )
- def clear_input():
- return "", None, "" # Clear url, files, text
+ # AI Patch button
+ patch_btn.click(
+ fn=get_ai_patch_wrapper,
+ inputs=[selected_issue_id_hidden, model_select], # Read selected issue ID from hidden state
+ outputs=[ai_output_display],
+ api_name="generate_patch",
+ show_progress="full"
+ ).then(
+ # After getting patch, check if it contains a diff block and show the copy button if so
+ # Use a lambda to check the output text
+ fn=lambda output_text: gr.update(visible="```diff" in output_text),
+ inputs=[ai_output_display],
+ outputs=[copy_patch_btn]
+ )
- def update_viewport(paths, enabled_states):
- if not paths:
- return "No QR codes generated yet.
"
+ # Clear AI Output button
+ clear_ai_output_btn.click(
+ fn=lambda: ["*AI suggestions and patches will appear here...*", gr.update(visible=False)],
+ inputs=[],
+ outputs=[ai_output_display, copy_patch_btn]
+ )
- num_qr_codes = len(paths)
- cols = math.ceil(math.sqrt(num_qr_codes)) # Calculate columns for a roughly square grid
- cols = max(1, min(cols, 6)) # Limit max columns for small screens
- rows = math.ceil(num_qr_codes / cols)
+ # --- JavaScript for WebSocket Communication and UI Interaction ---
+ def web_socket_js(ws_port, gradio_port):
+ # Generate a unique client ID on page load
+ # Note: This JS is generated *once* when the Gradio app is created.
+ # The actual client ID is generated and logged when the JS runs in the browser.
+ # The Python-side logging here is just for context during app startup.
+ temp_client_id_placeholder = f"client_{hashlib.sha1(os.urandom(16)).hexdigest()[:8]}"
+ logger.info(f"Generating JS with placeholder Client ID: {temp_client_id_placeholder}")
+
+ return f"""
+
+ """
+ # The _js parameter injects the JavaScript code into the Gradio page
+ demo_app.load(_js=web_socket_js(WS_PORT, GRADIO_PORT), fn=None, inputs=None, outputs=None)
+
+ return demo_app
+
+# ========== WebSocket Server Logic ==========
+async def handle_ws_connection(websocket: WebSocketServerProtocol, path: str, manager: IssueManager):
+ """Handles incoming WebSocket connections and messages for collaboration."""
+ # Generate a client ID and attach it to the websocket object
+ client_id = f"client_{hashlib.sha1(os.urandom(16)).hexdigest()[:8]}"
+ setattr(websocket, 'client_id', client_id)
+ remote_addr = websocket.remote_address
+ logger.info(f"WebSocket client connected: {remote_addr} assigned ID {client_id}")
+
+ # Add the new client to the list of active clients
+ manager.ws_clients.append(websocket)
+ logger.info(f"Client list size: {len(manager.ws_clients)}")
+ try:
+ # Wait for the first message (expected to be 'join') or other messages
+ async for message in websocket:
try:
- # Process JSON input
- if text and text.strip():
- try:
- json_data = json.loads(text)
- # Wrap direct JSON input in a dictionary for consistency with file/URL output structure
- results.append({
- 'source': 'json_input',
- 'extracted_data': json_data,
- 'timestamp': datetime.now().isoformat(),
- 'processing_notes': ['Parsed from direct JSON input.']
- })
- processing_status_messages.append("✅ Successfully parsed direct JSON input.")
- except json.JSONDecodeError as e:
- processing_status_messages.append(f"❌ Invalid JSON format in text input: {str(e)}")
- except Exception as e:
- processing_status_messages.append(f"❌ Error processing direct JSON input: {str(e)}")
-
-
- # Process URLs
- if urls and urls.strip():
- url_list = re.split(r'[,\n]', urls)
- url_list = [url.strip() for url in url_list if url.strip()]
- for url in url_list:
- validation = url_processor.validate_url(url)
- if validation['is_valid']:
- processing_status_messages.append(f"🌐 Fetching URL: {url}...")
- content_result = url_processor.fetch_content(url)
- if content_result:
- results.append(content_result)
- processing_status_messages.append(f"✅ Fetched and processed URL: {url}")
- else:
- processing_status_messages.append(f"❌ Failed to fetch/process URL: {url}")
- if validation['details'].get('final_url'):
- processing_status_messages[-1] += f" (Redirected to {validation['details']['final_url']})"
- else:
- processing_status_messages.append(f"⚠️ Skipping invalid URL: {url} ({validation['message']})")
-
- # Process files
- if files:
- for file in files:
- processing_status_messages.append(f"📁 Processing file: {file.name}...")
- file_results = file_processor.process_file(file)
- if file_results:
- results.extend(file_results)
- processing_status_messages.append(f"✅ Processed file: {file.name}")
- else:
- processing_status_messages.append(f"❌ Failed to process file: {file.name}")
+ # Ensure message is bytes or string before decoding/parsing
+ if isinstance(message, bytes):
+ message = message.decode('utf-8')
+ if not isinstance(message, str):
+ logger.warning(f"Received non-string/bytes message from {client_id}: {message!r}")
+ continue
+
+ data = json.loads(message)
+ msg_type = data.get("type")
+ # Use the client_id assigned by the server, not one sent by the client, for security
+ sender_id = client_id
+
+ logger.debug(f"Received WS message type '{msg_type}' from {sender_id} ({remote_addr})")
+
+ if msg_type == "join":
+ # Store collaborator info when they explicitly join
+ # Use the name provided by the client, default if not provided
+ client_name = data.get("name", f"User_{sender_id[:4]}")
+ # Ensure name is a string and not too long
+ client_name = str(client_name)[:50] if client_name else f"User_{sender_id[:4]}"
+
+ if sender_id in manager.collaborators:
+ # Update existing entry if client reconnects or sends join again
+ manager.collaborators[sender_id].update({"name": client_name, "status": "Connected"})
+ logger.info(f"Client {sender_id} ({client_name}) updated status to Connected.")
+ else:
+ # Add new entry
+ manager.collaborators[sender_id] = {"name": client_name, "status": "Connected"}
+ logger.info(f"Client {sender_id} ({client_name}) joined collaboration. Current collaborators: {list(manager.collaborators.keys())}")
+
+ # Broadcast updated status list to all clients
+ await manager.broadcast_collaboration_status_once()
+
+ elif msg_type == "code_update":
+ issue_num = data.get("issue_num")
+ delta_str = data.get("delta")
+ # Ensure data is valid and sender ID matches
+ if issue_num is not None and delta_str is not None and sender_id == client_id:
+ # FIX: Corrected call - handle_ws_connection is already async, just await the async manager method
+ await manager.handle_code_editor_update(int(issue_num), delta_str, sender_id)
+ else:
+ logger.warning(f"Invalid or unauthorized 'code_update' message from {sender_id}: Missing issue_num/delta or sender mismatch. Data: {str(data)[:200]}")
+
+ elif msg_type == "status_update":
+ status = data.get("status", "Idle")
+ # Only update status for the client ID that sent the message
+ if sender_id == client_id:
+ # Ensure status is a string and not too long
+ status = str(status)[:100] if status else "Idle"
+
+ if sender_id in manager.collaborators:
+ manager.collaborators[sender_id]["status"] = status
+ # Broadcast updated status list
+ await manager.broadcast_collaboration_status_once()
+ else:
+ # This might happen if 'join' wasn't received first, or state is out of sync
+ logger.warning(f"Received status update from client {sender_id} not in collaborator list. Adding/Updating with default name.")
+ manager.collaborators[sender_id] = {"name": f"User_{sender_id[:4]} (Re-added)", "status": status}
+ await manager.broadcast_collaboration_status_once()
+ else:
+ logger.warning(f"Unauthorized status update from {sender_id} attempting to update status for another client. Ignoring.")
- # Generate QR codes
- qr_paths = []
- final_json_output = None
- if results:
- # Use the collected results (list of dicts) for QR code generation
- qr_paths = generate_qr_codes(results, combine)
- final_json_output = results # Show the structured data in the JSON output box
+ else:
+ logger.warning(f"Unknown WebSocket message type '{msg_type}' received from {sender_id} ({remote_addr}). Message: {str(message)[:200]}")
- if qr_paths:
- processing_status_messages.append(f"✅ Successfully generated {len(qr_paths)} QR codes.")
- else:
- processing_status_messages.append("❌ Failed to generate QR codes.")
+ except json.JSONDecodeError:
+ logger.error(f"Received invalid JSON over WebSocket from {client_id} ({remote_addr}): {str(message)[:200]}...")
+ except Exception as e:
+ logger.exception(f"Error processing WebSocket message from {client_id} ({remote_addr}): {e}")
- else:
- processing_status_messages.append("⚠️ No valid content collected from inputs.")
+ # Catch standard socket exceptions for disconnects
+ except (ConnectionClosed, ConnectionClosedOK, ConnectionAbortedError, ConnectionResetError, WebSocketException) as e:
+ logger.info(f"WebSocket client {client_id} ({remote_addr}) disconnected: Type={type(e).__name__}, Code={getattr(e, 'code', 'N/A')}, Reason='{getattr(e, 'reason', 'N/A')}'")
+ except Exception as e:
+ logger.exception(f"Unexpected error in WebSocket handler for {client_id} ({remote_addr}): {e}")
+ finally:
+ # Ensure cleanup happens regardless of how the loop exits
+ logger.info(f"Cleaning up connection for client {client_id} ({remote_addr})")
+ # Pass the websocket object itself for removal
+ # Schedule this in the main loop if not already in it
+ if manager.main_loop.is_running():
+ manager.main_loop.call_soon_threadsafe(manager.remove_ws_client, websocket)
+ else:
+ logger.warning("Main loop not running, cannot schedule final client removal.")
- except Exception as e:
- logger.error(f"Overall processing error in process_inputs: {e}")
- processing_status_messages.append(f"❌ An unexpected error occurred during processing: {str(e)}")
+async def start_websocket_server(manager: IssueManager, port: int):
+ """Starts the WebSocket server."""
+ # The handler needs the manager instance
+ handler_with_manager = lambda ws, path: handle_ws_connection(ws, path, manager)
+ server = None
+ # Use a Future to signal when the server should stop
+ stop_event = asyncio.Future()
- return (
- final_json_output,
- [str(path) for path in qr_paths], # Gradio Gallery expects list of paths (strings)
- "\n".join(processing_status_messages) # Join status messages
- )
+ # Add a method to signal the server to stop from outside the async function
+ # This is needed for graceful shutdown from the main thread
+ manager.stop_ws_server = lambda: stop_event.set_result(True) if not stop_event.done() else None
- def on_qr_generation(qr_paths_list):
- # When QR codes are generated, update the state with the list of paths
- # and initialize the enabled_qr_codes state with all indices enabled
- num_qrs = len(qr_paths_list)
- initial_enabled_states = list(range(num_qrs))
- return qr_paths_list, initial_enabled_states # Return paths list and initial enabled state
-
-
- # Link events
- example_btn.click(load_example, inputs=[], outputs=text_input)
- clear_btn.click(clear_input, inputs=[], outputs=[url_input, file_input, text_input]) # Clear all inputs
-
- process_btn.click(
- process_inputs,
- inputs=[url_input, file_input, text_input, combine_data],
- outputs=[output_json, output_gallery, output_text]
- ).then( # Chain a .then() to update the QR paths state and trigger viewport update
- on_qr_generation,
- inputs=[output_gallery], # Get the list of paths from the gallery output
- outputs=[qr_code_paths, enabled_qr_codes] # Update the state variables
+ try:
+ # Start the websockets server
+ server = await websockets.serve(
+ handler_with_manager,
+ "0.0.0.0", # Listen on all interfaces
+ port,
+ ping_interval=20, # Send ping every 20 seconds
+ ping_timeout=20 # Close connection if no pong received within 20 seconds
)
+ logger.info(f"WebSocket server started successfully on ws://0.0.0.0:{port}")
+
+ # Wait until the stop_event is set (signaled from main thread or shutdown)
+ await stop_event
+
+ except OSError as e:
+ logger.error(f"Failed to start WebSocket server on port {port}: {e}. Is the port already in use?")
+ # Signal the main loop to stop gracefully if possible, or re-raise
+ # If this happens during startup, main thread will catch it and trigger shutdown.
+ # No need to explicitly call shutdown_handler here, the main thread's except/finally will handle it.
+ raise SystemExit(f"WebSocket Port {port} unavailable. Application cannot start.") from e
+ except asyncio.CancelledError:
+ logger.info("WebSocket server task cancelled.")
+ except Exception as e:
+ logger.exception(f"An unexpected error occurred starting or running the WebSocket server: {e}")
+ # Ensure the stop event is set so the await completes
+ if not stop_event.done(): stop_event.set_result(True)
+ raise # Re-raise to potentially stop the main loop
+
+ finally:
+ if server:
+ logger.info("Attempting to stop WebSocket server...")
+ server.close() # Signal server to close
+ await server.wait_closed() # Wait for server to finish closing connections
+ logger.info("WebSocket server stopped.")
+ # Ensure the stop event is completed even if there was an error before await
+ if not stop_event.done():
+ stop_event.set_result(True)
+
+
+def run_webhook_server(manager: IssueManager, port: int, main_loop: asyncio.AbstractEventLoop):
+ """Starts the HTTP webhook server in a separate thread."""
+ # Pass the manager instance and the main asyncio loop reference to the handler class
+ WebhookHandler.manager_instance = manager
+ WebhookHandler.main_loop = main_loop
+ httpd = None
+
+ # Add a method to signal the webhook server thread to stop
+ # This is needed for graceful shutdown from the main thread
+ # Use a simple flag and check it periodically, or rely on httpd.shutdown()
+ # httpd.shutdown() is the standard way for BaseHTTPRequestHandler servers
+ manager.stop_webhook_server = lambda: httpd.shutdown() if httpd else None
- # The viewport tab's select event will trigger update_viewport to render the grid
- viewport_tab.select(update_viewport, inputs=[qr_code_paths, enabled_qr_codes], outputs=[viewport_output])
- # Add helpful documentation
- gr.Markdown("""
- ### 🚀 Features
-
- - **Enhanced URL Scraping**: Extracts HTML text, title, meta description, links, and attempts parsing JSON/XML from URLs based on content type.
- - **Advanced File Processing**: Reads various text-based files (.txt, .md, .log etc.), HTML, XML, CSV, and attempts text extraction from common documents (.pdf, .docx, .rtf, .odt - *requires extra dependencies*).
- - **Smart JSON Handling**: Parses valid JSON from direct input, files (.json or content), or URLs.
- - **Archive Support**: Extracts and processes supported files from .zip, .tar, .gz archives.
- - **Robust Encoding Detection**: Uses `chardet` for reliable character encoding identification.
- - **Structured Output**: Provides a consistent JSON output format containing raw content (if applicable), extracted data, and processing notes for each processed item.
- - **Sequential QR Codes**: Maintains data integrity across multiple codes by chunking the combined/individual processed data.
- - **QR Code Viewport**: Visualize generated QR codes in a sequenced square grid with options to enable/disable individual codes for selective scanning/sharing.
- - **Modern Design**: Clean, responsive interface with visual feedback.
-
- ### 💡 Tips
-
- 1. **URLs**: Enter multiple URLs separated by commas or newlines. The processor will attempt to fetch and structure the content based on its type.
- 2. **Files**: Upload any type of file. The processor will attempt to handle supported text-based files, archives (.zip, .tar, .gz), and specific document/structured formats.
- 3. **JSON**: Use the "Direct JSON Input" tab for pasting JSON data. The system also tries to detect JSON content in file uploads and URLs. Use the "Load Example" button to see a sample JSON structure.
- 4. **Dependencies**: Processing PDF, DOCX, RTF, and ODT files requires installing optional Python libraries. Check the console logs for warnings if a library is missing.
- 5. **QR Codes**: Choose whether to "Combine all data into sequence" or generate separate sequences for each input item.
- 6. **Processing**: Monitor the "Processing Status" box for real-time updates and notes about errors or processing steps.
- 7. **Output**: The "Processed Data" JSON box shows the structured data extracted from your inputs. The "Generated QR Codes" gallery shows the QR code images.
-
- ### 🎨 Output Details
-
- - The "Processed Data" JSON will be a list of dictionaries. Each dictionary represents one processed input (URL or file).
- - Each item will have keys like `source`, `filename` (for files), `url` (for URLs), `mime_type`, `raw_content` (if readable), `extracted_data`, and `processing_notes`.
- - `extracted_data` will contain the parsed/extracted content, structured according to the input type (e.g., dictionary for JSON, text for documents, list of rows for CSV, dictionary with title/text/links for HTML).
- - `processing_notes` will list any issues encountered during extraction.
- - Generated QR codes are saved in the `output/qr_codes` directory.
-
- ### ⚙️ QR Code Viewport Instructions
-
- 1. Navigate to the **QR Code Viewport** tab after generating QR codes.
- 2. The generated QR codes will be displayed in a grid based on their total count.
- 3. Use the checkboxes below each QR code to enable or disable it for visual selection. Enabled codes have a green border and full opacity.
- 4. This viewport is currently for visualization and selection *within the UI*; it doesn't change the generated files themselves. You would manually select which physical QR codes to scan based on this view.
+ try:
+ server_address = ("0.0.0.0", port)
+ # Create the HTTP server instance
+ httpd = HTTPServer(server_address, WebhookHandler)
+ logger.info(f"Webhook HTTP server starting on http://0.0.0.0:{port}")
+ # Start serving requests (this call blocks the thread)
+ httpd.serve_forever()
+ except OSError as e:
+ logger.error(f"Failed to start Webhook server on port {port}: {e}. Is the port already in use?")
+ # If the server fails to start, signal the main loop to stop
+ if main_loop.is_running():
+ # Use call_soon_threadsafe as this is in a different thread
+ main_loop.call_soon_threadsafe(main_loop.stop)
+ except Exception as e:
+ logger.exception(f"Unexpected error in Webhook server thread: {e}")
+ # If an unexpected error occurs, signal the main loop to stop
+ if main_loop.is_running():
+ main_loop.call_soon_threadsafe(main_loop.stop)
+ finally:
+ if httpd:
+ logger.info("Shutting down Webhook HTTP server...")
+ # This method stops the serve_forever() loop
+ # It needs to be called from a different thread than the one running serve_forever()
+ # The manager.stop_webhook_server lambda is designed for this.
+ # If this finally block is reached due to an exception *within* serve_forever(),
+ # httpd.shutdown() might not be needed or might fail, but calling it is safer.
+ try:
+ httpd.shutdown() # Signal the server to stop accepting new connections and finish current ones
+ except Exception as e:
+ logger.warning(f"Error during httpd.shutdown(): {e}")
+ try:
+ httpd.server_close() # Close the server socket
+ except Exception as e:
+ logger.warning(f"Error during httpd.server_close(): {e}")
- """)
- return interface
+ logger.info("Webhook server thread finished.")
-def main():
- """Initialize and launch the application"""
+
+# ========== Main Execution ==========
+if __name__ == "__main__":
+ print("--- Acknowledging potential TensorFlow/CUDA warnings ---")
+ print("If you see warnings like 'Could not load dynamic library...' or related to CUDA/GPU,")
+ print("they are often harmless if you are not using a local GPU-accelerated model.")
+ print("Hugging Face Inference API calls run remotely and do not require local GPU setup.")
+ print("--- Starting Application ---")
+
+ # Get or create the event loop for the main thread
+ # This loop will run the async tasks (WS server, idle tasks, broadcast)
try:
- # Configure system settings
- mimetypes.init()
-
- # Create and launch interface
- interface = create_modern_interface()
-
- # Launch with configuration
- interface.launch(
- share=False,
- debug=False, # Set to True for more verbose Gradio logging
- show_error=True,
- show_api=False
+ loop = asyncio.get_event_loop()
+ logger.info(f"Using existing asyncio loop in main thread: {id(loop)}")
+ except RuntimeError:
+ loop = asyncio.new_event_loop()
+ asyncio.set_event_loop(loop)
+ logger.info(f"Created and set new asyncio loop in main thread: {id(loop)}")
+
+ manager = IssueManager()
+
+ # Start the webhook server in a separate thread
+ webhook_thread = threading.Thread(
+ target=run_webhook_server,
+ args=(manager, WEBHOOK_PORT, loop), # Pass the main loop to the webhook thread
+ name="WebhookServerThread",
+ daemon=True # Allow the main program to exit even if this thread is running
+ )
+ webhook_thread.start()
+
+ # Create the Gradio UI
+ app = create_ui(manager)
+
+ # Start the WebSocket server as an asyncio task in the main loop
+ ws_task = loop.create_task(start_websocket_server(manager, WS_PORT))
+
+ # Start the manager's background tasks (idle processing, broadcast)
+ # These tasks are created in the manager's loop (which is the main loop here)
+ manager.start_idle_processing()
+ manager.start_broadcast_loop()
+
+
+ def shutdown_handler(signum, frame):
+ """Graceful shutdown logic."""
+ logger.info(f"Signal {signum} received. Initiating shutdown...")
+ # Signal the WebSocket server to stop
+ if hasattr(manager, 'stop_ws_server') and manager.stop_ws_server:
+ manager.stop_ws_server()
+ # Signal the webhook server to stop
+ if hasattr(manager, 'stop_webhook_server') and manager.stop_webhook_server:
+ manager.stop_webhook_server()
+ # Stop manager's internal async tasks
+ manager.stop_idle_processing()
+ manager.stop_broadcast_loop()
+
+ # Stop the main asyncio loop
+ if loop.is_running():
+ logger.info("Stopping main asyncio loop...")
+ # Use call_soon_threadsafe because signal handlers run in a different context
+ loop.call_soon_threadsafe(loop.stop)
+ else:
+ logger.warning("Main asyncio loop not running during shutdown handler.")
+
+
+ # Add signal handlers for graceful shutdown (e.g., Ctrl+C)
+ try:
+ # For Unix-like systems
+ loop.add_signal_handler(signal.SIGINT, shutdown_handler, signal.SIGINT, None)
+ loop.add_signal_handler(signal.SIGTERM, shutdown_handler, signal.SIGTERM, None)
+ logger.info("Added signal handlers for SIGINT and SIGTERM.")
+ except NotImplementedError:
+ # Signal handlers are not available on Windows
+ logger.warning("Signal handlers not available on this platform (likely Windows). Ctrl+C may not be graceful.")
+ # On Windows, KeyboardInterrupt is usually caught directly in the main thread
+
+
+ # Launch Gradio app
+ # Use prevent_thread_lock=True to run Gradio server in a separate thread,
+ # freeing the main thread to run the asyncio loop.
+ try:
+ logger.info(f"Launching Gradio app on port {GRADIO_PORT}...")
+ app.launch(
+ server_name="0.0.0.0",
+ server_port=GRADIO_PORT,
+ share=False, # Set to True to share publicly (requires auth token usually)
+ debug=True, # Keep debug=True for development logs
+ inbrowser=True,
+ prevent_thread_lock=True # Run Gradio server in a separate thread
)
+ logger.info("Gradio app launched in separate thread.")
+
+ # Run the main asyncio loop indefinitely to keep async tasks running
+ logger.info("Running main asyncio loop...")
+ loop.run_forever() # This blocks the main thread until loop.stop() is called
+
+ except KeyboardInterrupt:
+ logger.info("KeyboardInterrupt received in main thread.")
+ # If signal handlers are not implemented (e.g., Windows), KeyboardInterrupt
+ # is caught here. Trigger shutdown manually.
+ # Check if signal handlers were likely not registered by checking if signal.SIGINT exists
+ if not hasattr(signal, 'SIGINT'):
+ shutdown_handler(None, None) # Call shutdown logic
+
+ except SystemExit as e:
+ logger.error(f"SystemExit received: {e}")
+ # SystemExit might be raised by port binding errors etc.
+ # The source of SystemExit might have already triggered loop.stop() or shutdown.
+ # Ensure cleanup happens if not already started.
+ # Check if the loop is still running; if so, stop it.
+ if loop.is_running():
+ logger.info("SystemExit caught before loop stopped, triggering shutdown.")
+ shutdown_handler(None, None)
+ else:
+ logger.info("SystemExit caught, but loop already stopped. Proceeding with final cleanup.")
+
+
except Exception as e:
- logger.error(f"Application startup error: {e}")
- # Optionally print a user-friendly message before exiting
- print(f"\nFatal Error: {e}\nCheck the logs for details.")
- raise # Re-raise the exception to ensure the process exits if launch fails
+ logger.exception(f"An unexpected error occurred in the main thread: {e}")
+ # Trigger graceful shutdown on unexpected error
+ shutdown_handler(None, None)
+
+ finally:
+ logger.info("Main thread exiting finally block.")
+ # Wait for background tasks/threads to complete shutdown...
+ logger.info("Running loop briefly to complete pending tasks (e.g., cancellations)...")
+ try:
+ # Gather remaining tasks (like ws_task cancellation, idle_task cancellation)
+ # Exclude the current task if inside a task (though we are in the main thread here)
+ # asyncio.all_tasks() returns tasks for the *current* thread's loop
+ pending_tasks = [task for task in asyncio.all_tasks(loop=loop) if not task.done()]
+ if pending_tasks:
+ # Use a timeout for waiting for tasks to finish
+ try:
+ # Wait for pending tasks, ignoring exceptions during shutdown
+ loop.run_until_complete(asyncio.wait(pending_tasks, timeout=5, return_when=asyncio.ALL_COMPLETED))
+ logger.info(f"Completed pending tasks.")
+ except asyncio.TimeoutError:
+ logger.warning(f"Timed out waiting for {len(pending_tasks)} pending tasks to complete.")
+ except Exception as e:
+ logger.error(f"Error during final loop run_until_complete: {e}")
+ else:
+ logger.info("No pending tasks to complete.")
+ except Exception as e:
+ logger.error(f"Error checking pending tasks: {e}")
-if __name__ == "__main__":
- main()
\ No newline at end of file
+
+ # Wait for the webhook thread to join (if it's not daemon, or if shutdown() needs time)
+ # Since it's daemon, it will be killed, but explicit shutdown is better.
+ # Add a small timeout for join in case shutdown() hangs.
+ # Check if the thread is still alive before joining
+ if webhook_thread.is_alive():
+ logger.info("Waiting for webhook thread to join...")
+ webhook_thread.join(timeout=2) # Wait up to 2 seconds
+
+ # Close the loop
+ if not loop.is_closed():
+ logger.info("Closing asyncio loop.")
+ loop.close()
+
+ logger.info("Application shutdown complete.")
\ No newline at end of file