Add bot application files and Dockerfile

.hf_space

@@ -0,0 +1,17 @@
+# .hf_space
+title: Telegram Summary Bot
+emoji: 🤖
+colorFrom: blue
+colorTo: green
+sdk: docker # Specify we are using a Dockerfile
+# sdk_version: # Not needed for docker
+app_file: main.py # Reference the main file (though CMD in Dockerfile takes precedence)
+# app_port: 7860 # Port is defined in Dockerfile EXPOSE and CMD
+# Add secrets needed by the Space (these will be environment variables)
+secrets:
+  - TELEGRAM_TOKEN
+  # Add ONLY the keys you actually have secrets for:
+  - OPENROUTER_API_KEY
+  - URLTOTEXT_API_KEY
+  - SUPADATA_API_KEY
+  - APIFY_API_TOKEN

Dockerfile

@@ -0,0 +1,31 @@
+# Use an official Python runtime as a parent image
+FROM python:3.10-slim
+
+# Set the working directory in the container
+WORKDIR /app
+
+# Copy the requirements file into the container at /app
+COPY requirements.txt .
+
+# Install any needed system dependencies (if any - bs4, requests usually don't need much)
+# RUN apt-get update && apt-get install -y --no-install-recommends some-package && rm -rf /var/lib/apt/lists/*
+# For this bot, we likely don't need extra apt packages currently.
+
+# Install Python dependencies
+# Using --no-cache-dir reduces image size slightly
+RUN pip install --no-cache-dir -r requirements.txt
+
+# Copy the rest of the application code into the container at /app
+COPY . .
+
+# Make port 7860 available to the world outside this container
+# Hugging Face Spaces typically expect apps to run on port 7860
+EXPOSE 7860
+
+# Define environment variable to ensure Python output is sent straight to logs
+ENV PYTHONUNBUFFERED=1
+
+# Command to run the application using Gunicorn
+# It will run the Flask 'app' object found in the 'main' module (main.py)
+# Listen on all interfaces (0.0.0.0) on the port specified by HF (usually 7860)
+CMD ["gunicorn", "--bind", "0.0.0.0:7860", "main:app"]

main.py

@@ -0,0 +1,858 @@
+# main.py
+import os
+import re
+import logging
+import asyncio
+import json
+from flask import Flask, request, Response # For web server
+
+from telegram import Update, InlineKeyboardButton, InlineKeyboardMarkup
+from telegram.ext import (
+    Application,
+    CommandHandler,
+    MessageHandler,
+    filters,
+    ContextTypes,
+    CallbackQueryHandler
+)
+from telegram.constants import ParseMode # Import ParseMode explicitly
+
+# Import specific libraries (Ensure these are covered in requirements.txt)
+from youtube_transcript_api import YouTubeTranscriptApi
+import requests
+from bs4 import BeautifulSoup
+# Only import ApifyClient if you might use it (i.e., have the secret)
+if os.environ.get('APIFY_API_TOKEN'):
+    from apify_client import ApifyClient
+else:
+    ApifyClient = None # Define it as None if not used, to avoid errors later
+
+# Apply nest_asyncio early, can help prevent event loop conflicts in web frameworks
+import nest_asyncio
+nest_asyncio.apply()
+
+# --- Environment Variable Loading (using Replit Secrets) ---
+TELEGRAM_TOKEN = os.environ.get('TELEGRAM_TOKEN')
+OPENROUTER_API_KEY = os.environ.get('OPENROUTER_API_KEY')
+URLTOTEXT_API_KEY = os.environ.get('URLTOTEXT_API_KEY') # Will be None if not set in Secrets
+SUPADATA_API_KEY = os.environ.get('SUPADATA_API_KEY')   # Will be None if not set in Secrets
+APIFY_API_TOKEN = os.environ.get('APIFY_API_TOKEN')     # Will be None if not set in Secrets
+
+# --- Logging Setup ---
+logging.basicConfig(
+    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
+    level=logging.INFO # Set to INFO to see processing steps, DEBUG for more detail
+)
+# Reduce noise from libraries
+logging.getLogger("httpx").setLevel(logging.WARNING)
+if ApifyClient: # Only set level if imported
+    logging.getLogger("apify_client").setLevel(logging.WARNING)
+logging.getLogger("telegram.ext").setLevel(logging.INFO)
+logging.getLogger('telegram.bot').setLevel(logging.INFO)
+logging.getLogger("urllib3").setLevel(logging.INFO) # From requests
+logger = logging.getLogger(__name__)
+
+# --- Input Validation ---
+if not TELEGRAM_TOKEN:
+    logger.critical("FATAL: TELEGRAM_TOKEN environment variable not set in Replit Secrets!")
+    # In a deployed environment, exiting might not be helpful. Log critical error.
+    # exit("Telegram Token Missing")
+else:
+    logger.info("Telegram Token loaded.")
+
+if not OPENROUTER_API_KEY:
+    # Log error but allow running, summaries will just fail gracefully later
+    logger.error("OpenRouter API Key not set in Replit Secrets! Summarization will fail.")
+else:
+     logger.info("OpenRouter API Key loaded.")
+
+# Log presence of optional keys
+if URLTOTEXT_API_KEY: logger.info("URLToText API Key found.")
+else: logger.warning("URLToText API Key not found. Website fallback 2 unavailable.")
+if SUPADATA_API_KEY: logger.info("Supadata API Key found.")
+else: logger.warning("Supadata API Key not found. YT Transcript fallback 1 unavailable.")
+if APIFY_API_TOKEN: logger.info("Apify API Token found.")
+else: logger.warning("Apify API Token not found. YT Transcript fallback 2 unavailable.")
+
+
+# --- Bot Logic Functions (Simplified Version - No Crawl4AI) ---
+
+# Helper Functions
+def is_youtube_url(url):
+    """Checks if the URL is a valid YouTube video or shorts URL."""
+    youtube_regex = r'(https?://)?(www\.)?(youtube\.com/(watch\?v=|shorts/)|youtu\.be/)([\w-]{11})'
+    return bool(re.search(youtube_regex, url))
+
+def extract_youtube_id(url):
+    """Extracts the YouTube video ID from a URL."""
+    youtube_id_regex = r'(?:youtube\.com/(?:watch\?v=|shorts/)|youtu\.be/)([\w-]{11})'
+    match = re.search(youtube_id_regex, url)
+    if match:
+        return match.group(1)
+    logger.warning(f"Could not extract YouTube ID from URL: {url}")
+    return None
+
+# Supadata Transcript Fetching
+async def get_transcript_via_supadata(video_id: str, api_key: str):
+    """Fetches YouTube transcript via Supadata API."""
+    if not video_id: logger.error("[Supadata] get_transcript_via_supadata called with no video_id"); return None
+    if not api_key: logger.error("[Supadata] API key is missing."); return None # Already checked before calling
+    logger.info(f"[Supadata] Attempting fetch for video ID: {video_id}")
+    api_endpoint = f"https://api.supadata.net/v1/youtube/transcript"
+    params = {"videoId": video_id, "format": "text"}
+    headers = {"X-API-Key": api_key}
+    try:
+        # Use asyncio.to_thread to run blocking requests.get in a separate thread
+        response = await asyncio.to_thread(requests.get, api_endpoint, headers=headers, params=params, timeout=30)
+        logger.debug(f"[Supadata] Received status code {response.status_code} for {video_id}")
+        if response.status_code == 200:
+            try:
+                data = response.json()
+                content = data if isinstance(data, str) else data.get("transcript") or data.get("text") or data.get("data")
+                if content and isinstance(content, str):
+                    logger.info(f"[Supadata] Successfully fetched transcript for {video_id}. Length: {len(content)}")
+                    return content.strip()
+                else:
+                    logger.warning(f"[Supadata] API success but content empty/invalid for {video_id}. Response: {data}")
+                    return None
+            except json.JSONDecodeError: # Handle cases where API might return plain text on success
+                if response.text:
+                    logger.info(f"[Supadata] Successfully fetched transcript (plain text) for {video_id}. Length: {len(response.text)}")
+                    return response.text.strip()
+                else:
+                    logger.error(f"[Supadata] Failed to decode JSON response (and no text body) for {video_id}. Response: {response.text[:200]}...")
+                    return None
+            except Exception as e:
+                logger.error(f"[Supadata] Error processing successful response for {video_id}: {e}", exc_info=True)
+                return None
+        elif response.status_code in [401, 403]:
+            logger.error(f"[Supadata] Authentication error ({response.status_code}). Check API key.")
+            return None # Don't retry if key is bad
+        elif response.status_code == 404:
+            logger.warning(f"[Supadata] Transcript not found ({response.status_code}) for {video_id}.")
+            return None
+        else:
+             logger.error(f"[Supadata] Unexpected status code {response.status_code} for {video_id}. Response: {response.text[:200]}...")
+             return None
+    except requests.exceptions.Timeout:
+        logger.error(f"[Supadata] Timeout error connecting to API for {video_id}")
+        return None
+    except requests.exceptions.RequestException as e:
+        logger.error(f"[Supadata] Request error connecting to API for {video_id}: {e}")
+        return None
+    except Exception as e:
+        logger.error(f"[Supadata] Unexpected error during API call for {video_id}: {e}", exc_info=True)
+        return None
+
+# Apify Transcript Fetching
+async def get_transcript_via_apify(video_url: str, api_token: str):
+    """Fetches YouTube transcript via Apify API."""
+    if not video_url: logger.error("[Apify] get_transcript_via_apify called with no video_url"); return None
+    if not api_token: logger.error("[Apify] API token is missing."); return None # Already checked
+    if not ApifyClient: logger.error("[Apify] ApifyClient not available/imported."); return None
+
+    logger.info(f"[Apify] Attempting fetch for URL: {video_url}")
+    actor_id = "karamelo~youtube-transcripts"
+    api_endpoint = f"https://api.apify.com/v2/acts/{actor_id}/run-sync-get-dataset-items"
+    params = {"token": api_token}
+    payload = json.dumps({
+        "urls": [video_url],
+        "outputFormat": "singleStringText",
+        "maxRetries": 5,
+        "channelHandleBoolean": False,
+        "channelNameBoolean": False,
+        "datePublishedBoolean": False,
+        "relativeDateTextBoolean": False,
+    })
+    headers = {"Content-Type": "application/json"}
+    try:
+        logger.debug(f"[Apify] Sending request to run actor {actor_id} synchronously for {video_url}")
+        response = await asyncio.to_thread(requests.post, api_endpoint, headers=headers, params=params, data=payload, timeout=90) # Longer timeout for actor run
+        logger.debug(f"[Apify] Received status code {response.status_code} for {video_url}")
+        if response.status_code == 200:
+            try:
+                results = response.json()
+                if isinstance(results, list) and len(results) > 0:
+                    item = results[0]
+                    content = item.get("text") or item.get("transcript") or item.get("captions_concatenated")
+                    if not content and item.get("captions") and isinstance(item["captions"], list): # Handle 'captions' format if primary keys fail
+                         logger.info("[Apify] Processing 'captions' format.")
+                         content = " ".join(cap.get("text", "") for cap in item["captions"] if cap.get("text"))
+                    if content and isinstance(content, str):
+                        logger.info(f"[Apify] Successfully fetched transcript for {video_url}. Length: {len(content)}")
+                        return content.strip()
+                    else:
+                        logger.warning(f"[Apify] Actor run successful but transcript content not found/empty in result for {video_url}. Result item: {item}")
+                        return None
+                else:
+                    logger.warning(f"[Apify] Actor run successful but dataset was empty for {video_url}. Response: {results}")
+                    return None
+            except json.JSONDecodeError:
+                logger.error(f"[Apify] Failed to decode JSON response for {video_url}. Status: {response.status_code}. Response text: {response.text[:200]}...")
+                return None
+            except Exception as e:
+                logger.error(f"[Apify] Error processing successful response for {video_url}: {e}", exc_info=True)
+                return None
+        elif response.status_code == 400:
+             logger.error(f"[Apify] Bad Request (400) for {video_url}. Check input payload. Response: {response.text[:200]}...")
+             return None
+        elif response.status_code == 401:
+            logger.error("[Apify] Authentication error (401). Check API token.")
+            return None # Don't retry if token is bad
+        else:
+             logger.error(f"[Apify] Unexpected status code {response.status_code} for {video_url}. Response: {response.text[:200]}...")
+             return None
+    except requests.exceptions.Timeout:
+        logger.error(f"[Apify] Timeout error running actor for {video_url}")
+        return None
+    except requests.exceptions.RequestException as e:
+        logger.error(f"[Apify] Request error running actor for {video_url}: {e}")
+        return None
+    except Exception as e:
+        logger.error(f"[Apify] Unexpected error during Apify call for {video_url}: {e}", exc_info=True)
+        return None
+
+# Combined YouTube Transcript Function (with Fallbacks)
+async def get_youtube_transcript(video_id: str, video_url: str, supadata_key: str | None, apify_token: str | None):
+    """Fetches YouTube transcript using library, then Supadata, then Apify."""
+    if not video_id: logger.error("get_youtube_transcript called with no video_id"); return None
+    logger.info(f"Fetching transcript for video ID: {video_id} (URL: {video_url})")
+    transcript_text = None
+
+    # 1. Primary Method: youtube-transcript-api
+    logger.info("[Primary YT] Attempting youtube-transcript-api...")
+    try:
+        # Run synchronous library call in a thread
+        transcript_list = await asyncio.to_thread(
+            YouTubeTranscriptApi.get_transcript,
+            video_id,
+            languages=['en', 'en-GB', 'en-US'] # Prioritize English variations
+        )
+        if transcript_list:
+            transcript_text = " ".join([item['text'] for item in transcript_list if 'text' in item])
+            if transcript_text:
+                logger.info(f"[Primary YT] Successfully fetched transcript via library for {video_id} (length: {len(transcript_text)})")
+                return transcript_text # Return immediately on success
+            else:
+                 logger.warning(f"[Primary YT] Joined transcript text is empty for {video_id}")
+                 transcript_text = None # Ensure it's None if empty after join
+        else:
+            logger.warning(f"[Primary YT] Transcript list empty for {video_id}")
+            transcript_text = None
+    except Exception as e:
+        logger.warning(f"[Primary YT] Error getting transcript via library for {video_id}: {e}")
+        if "No transcript found" in str(e): logger.warning(f"[Primary YT] No transcript found for {video_id}. May be unavailable/private.")
+        elif "disabled" in str(e): logger.warning(f"[Primary YT] Transcripts disabled for {video_id}.")
+        transcript_text = None # Ensure it's None on error
+
+    # 2. Fallback 1: Supadata API
+    if transcript_text is None:
+        logger.info("[Fallback YT 1] Primary method failed. Trying Supadata API...")
+        if supadata_key:
+            transcript_text = await get_transcript_via_supadata(video_id, supadata_key)
+            if transcript_text:
+                logger.info(f"[Fallback YT 1] Successfully fetched transcript via Supadata for {video_id}")
+                return transcript_text # Return on success
+            else:
+                logger.warning(f"[Fallback YT 1] Supadata API failed or returned no content for {video_id}.")
+        else:
+            logger.warning("[Fallback YT 1] Supadata API key not available. Skipping.")
+
+    # 3. Fallback 2: Apify API
+    if transcript_text is None:
+        logger.info("[Fallback YT 2] Primary & Supadata failed. Trying Apify API...")
+        if apify_token:
+            transcript_text = await get_transcript_via_apify(video_url, apify_token)
+            if transcript_text:
+                logger.info(f"[Fallback YT 2] Successfully fetched transcript via Apify for {video_url}")
+                return transcript_text # Return on success
+            else:
+                logger.warning(f"[Fallback YT 2] Apify API failed or returned no content for {video_url}.")
+        else:
+            logger.warning("[Fallback YT 2] Apify API token not available. Skipping.")
+
+    # If all methods failed
+    if transcript_text is None:
+        logger.error(f"All methods failed to fetch transcript for video ID: {video_id}")
+        return None
+
+    # Should not be reached if logic above is correct, but as a safeguard
+    return transcript_text
+
+# Website Content via Requests/BS4 (Primary Method for Simplified Bot)
+async def get_website_content_via_requests(url):
+    """Attempts to scrape website content using requests/BeautifulSoup (Primary Method)."""
+    if not url: logger.error("[Web Scraper - Requests/BS4] called with no URL"); return None
+    logger.info(f"[Web Scraper - Requests/BS4] Fetching website content for: {url}")
+    try:
+        headers = {
+             'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/108.0.0.0 Safari/537.36', # Updated UA
+             'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8',
+             'Accept-Language': 'en-US,en;q=0.9',
+             'Connection': 'keep-alive',
+             'DNT': '1', # Do Not Track
+             'Upgrade-Insecure-Requests': '1'
+        }
+        logger.debug(f"[Web Scraper - Requests/BS4] Sending request to {url}")
+        # Run blocking I/O in a separate thread
+        response = await asyncio.to_thread(requests.get, url, headers=headers, timeout=25, allow_redirects=True)
+        response.raise_for_status() # Raise HTTPError for bad responses (4xx or 5xx)
+        logger.debug(f"[Web Scraper - Requests/BS4] Received response {response.status_code} from {url}")
+
+        content_type = response.headers.get('content-type', '').lower()
+        if 'html' not in content_type:
+            logger.warning(f"[Web Scraper - Requests/BS4] Non-HTML content type received from {url}: {content_type}")
+            return None # Don't try to parse non-html
+
+        # Use html.parser, it's built-in
+        soup = BeautifulSoup(response.text, 'html.parser')
+
+        # Remove common unwanted tags more aggressively
+        for element in soup(["script", "style", "header", "footer", "nav", "aside", "form", "button", "input", "textarea", "select", "option", "label", "iframe", "img", "svg", "link", "meta", "noscript", "figure", "figcaption", "video", "audio"]):
+            element.extract()
+
+        # Try finding common main content containers
+        main_content = soup.find('main') or \
+                       soup.find('article') or \
+                       soup.find(id='content') or \
+                       soup.find(class_='content') or \
+                       soup.find(id='main-content') or \
+                       soup.find(class_='main-content') or \
+                       soup.find(role='main')
+
+        # Fallback to body if no specific container found
+        target_element = main_content if main_content else soup.body
+
+        if not target_element:
+            logger.warning(f"[Web Scraper - Requests/BS4] Could not find body or main content container for parsing {url}")
+            return None # Nothing to parse
+
+        # Get text, joining lines smartly
+        lines = [line.strip() for line in target_element.get_text(separator='\n', strip=True).splitlines() if line.strip()]
+        text = "\n".join(lines) # Join with newlines to preserve some structure
+
+        # Basic length check
+        if not text or len(text) < 50: # Arbitrary short length check
+             logger.warning(f"[Web Scraper - Requests/BS4] Extracted text is very short or empty after cleaning for {url} (Length: {len(text)})")
+             # Consider returning None if too short, depends on use case
+             # return None
+
+        logger.info(f"[Web Scraper - Requests/BS4] Successfully scraped content for {url} (final length: {len(text)})")
+        return text
+
+    except requests.exceptions.Timeout:
+        logger.error(f"[Web Scraper - Requests/BS4] Timeout error scraping website: {url}")
+        return None
+    except requests.exceptions.TooManyRedirects:
+        logger.error(f"[Web Scraper - Requests/BS4] Too many redirects error scraping website: {url}")
+        return None
+    except requests.exceptions.RequestException as e:
+        # This catches ConnectTimeout, HTTPError, ConnectionError etc.
+        logger.error(f"[Web Scraper - Requests/BS4] Request error scraping website {url}: {e}")
+        return None
+    except Exception as e:
+        # Catch-all for unexpected errors during parsing etc.
+        logger.error(f"[Web Scraper - Requests/BS4] Error scraping or parsing website {url}: {e}", exc_info=True)
+        return None
+
+# Website Content via URLToText API (Fallback Method)
+async def get_website_content_via_urltotext_api(url: str, api_key: str):
+    """Fetches website content using the URLToText API (Fallback)."""
+    if not url: logger.error("[Web Scraper - URLToText API] called with no URL"); return None
+    if not api_key: logger.error("[Web Scraper - URLToText API] API key is missing."); return None # Already checked
+    logger.info(f"[Web Scraper - URLToText API] Attempting to fetch content for: {url}")
+    api_endpoint = "https://urltotext.com/api/v1/urltotext/"
+    payload = json.dumps({
+        "url": url,
+        "output_format": "text",
+        "extract_main_content": True,
+        "render_javascript": True, # Often needed for modern sites
+        "residential_proxy": False, # Start with standard
+    })
+    headers = {
+        "Authorization": f"Token {api_key}",
+        "Content-Type": "application/json"
+    }
+    try:
+        logger.debug(f"[Web Scraper - URLToText API] Sending request for {url}")
+        response = await asyncio.to_thread(requests.post, api_endpoint, headers=headers, data=payload, timeout=45)
+        logger.debug(f"[Web Scraper - URLToText API] Received status code {response.status_code} for {url}")
+        if response.status_code == 200:
+            try:
+                data = response.json()
+                content = data.get("data", {}).get("content")
+                credits = data.get("credits_used", "N/A")
+                warning = data.get("data", {}).get("warning")
+                if warning: logger.warning(f"[Web Scraper - URLToText API] Warning for {url}: {warning}")
+                if content:
+                    logger.info(f"[Web Scraper - URLToText API] Successfully fetched content via API for {url}. Length: {len(content)}. Credits: {credits}")
+                    return content.strip()
+                else:
+                    logger.warning(f"[Web Scraper - URLToText API] API returned success but content was empty for {url}. Response: {data}")
+                    return None
+            except json.JSONDecodeError:
+                logger.error(f"[Web Scraper - URLToText API] Failed to decode JSON response for {url}. Response: {response.text[:500]}...")
+                return None
+            except Exception as e:
+                logger.error(f"[Web Scraper - URLToText API] Error processing successful API response for {url}: {e}", exc_info=True)
+                return None
+        elif response.status_code in [400, 402, 422, 500]: # Known client/server errors
+             logger.error(f"[Web Scraper - URLToText API] Error {response.status_code} from API for {url}. Response: {response.text[:200]}...")
+             return None
+        else: # Other unexpected codes
+             logger.error(f"[Web Scraper - URLToText API] Unexpected status {response.status_code} from API for {url}. Response: {response.text[:200]}...")
+             return None
+    except requests.exceptions.Timeout:
+        logger.error(f"[Web Scraper - URLToText API] Timeout error connecting to API for {url}")
+        return None
+    except requests.exceptions.RequestException as e:
+        logger.error(f"[Web Scraper - URLToText API] Request error connecting to API for {url}: {e}")
+        return None
+    except Exception as e:
+        logger.error(f"[Web Scraper - URLToText API] Unexpected error during API call for {url}: {e}", exc_info=True)
+        return None
+
+# DeepSeek Summary Function (via OpenRouter)
+async def generate_summary(text: str, summary_type: str, api_key: str) -> str:
+    """Generates summary using DeepSeek via OpenRouter API."""
+    logger.info(f"Generating {summary_type} summary using DeepSeek/OpenRouter. Input text length: {len(text)}")
+    if not api_key:
+         logger.error("OpenRouter API key was not provided to generate_summary.")
+         return "Error: AI model configuration key (OpenRouter) is missing."
+
+    openrouter_api_endpoint = "https://openrouter.ai/api/v1/chat/completions"
+    # Check OpenRouter docs for the latest recommended free/low-cost models
+    model_name = "deepseek/deepseek-chat:free"
+
+    if summary_type == "paragraph":
+        prompt = "You are an AI model designed to provide concise summaries using British English spellings. Your output MUST be: • Clear and simple language suitable for someone unfamiliar with the topic. • Uses British English spellings throughout. • Straightforward and understandable vocabulary; avoid complex terms. • Presented as ONE SINGLE PARAGRAPH. • No more than 85 words maximum; but does not have to be exactly 85. • Considers the entire text content equally. • Uses semicolons (;) instead of em dashes (– or —). Here is the text to summarise:"
+    else: # points summary
+        prompt = """You are an AI model designed to provide concise summaries using British English spellings. Your output MUST strictly follow this Markdown format:
+• For each distinct topic or section identified in the text, create a heading.
+• Each heading MUST be enclosed in double asterisks for bolding (e.g., **Section Title**).
+• Immediately following each heading, list the key points as a bulleted list.
+• Each bullet point MUST start with a hyphen and a space (`- `) on a new line.
+• The text within each bullet point should NOT contain any bold formatting.
+• Use clear, simple, and straightforward language suitable for someone unfamiliar with the topic.
+• Use British English spellings throughout.
+• Avoid overly complex or advanced vocabulary.
+• Keep bullet points concise.
+• Ensure the entire summary takes no more than two minutes to read.
+• Consider the entire text's content, not just the beginning or a few topics.
+• Use semicolons (;) instead of em dashes (– or —).
+
+Here is the text to summarise:"""
+
+    MAX_INPUT_LENGTH = 500000 # Truncate long inputs to avoid high costs/errors
+    if len(text) > MAX_INPUT_LENGTH:
+        logger.warning(f"Input text length ({len(text)}) exceeds limit ({MAX_INPUT_LENGTH}). Truncating.")
+        text = text[:MAX_INPUT_LENGTH] + "... (Content truncated)"
+    full_prompt = f"{prompt}\n\n{text}"
+
+    headers = {
+        "Authorization": f"Bearer {api_key}",
+        "Content-Type": "application/json",
+        # Recommended headers for OpenRouter identification
+        "HTTP-Referer": "https://replit.com/", # Identify source
+        "X-Title": "Telegram Summary Bot (Replit)", # Identify app
+    }
+    payload = json.dumps({
+        "model": model_name,
+        "messages": [
+            {"role": "user", "content": full_prompt}
+        ],
+        # Optional: Add max_tokens if needed, check model defaults
+        # "max_tokens": 1024,
+    })
+
+    try:
+        logger.debug(f"Sending request to OpenRouter ({model_name})...")
+        # Run blocking request in thread
+        response = await asyncio.to_thread(requests.post, openrouter_api_endpoint, headers=headers, data=payload, timeout=60)
+        logger.debug(f"Received status code {response.status_code} from OpenRouter.")
+
+        if response.status_code == 200:
+            try:
+                data = response.json()
+                if data.get("choices") and isinstance(data["choices"], list) and len(data["choices"]) > 0:
+                    message = data["choices"][0].get("message")
+                    if message and isinstance(message, dict):
+                        summary = message.get("content")
+                        if summary:
+                            logger.info(f"Successfully generated summary via OpenRouter. Output length: {len(summary)}")
+                            return summary.strip()
+                        else:
+                            logger.warning(f"OpenRouter response successful, but content was empty. Response: {data}")
+                            return "Sorry, the AI model returned an empty summary."
+                    else:
+                         logger.error(f"Unexpected message structure in OpenRouter response: {message}. Full response: {data}")
+                         return "Sorry, could not parse the AI model's response (unexpected format)."
+                else:
+                    # Handle cases like moderation flags, empty choices list
+                    if data.get("error"): logger.error(f"OpenRouter API Error: {data['error']}")
+                    else: logger.error(f"Unexpected choices structure in OpenRouter response: {data.get('choices')}. Full response: {data}")
+                    return "Sorry, could not parse the AI model's response (choices missing/invalid or API error)."
+
+            except json.JSONDecodeError:
+                logger.error(f"Failed to decode JSON response from OpenRouter. Status: {response.status_code}. Response text: {response.text[:500]}...")
+                return "Sorry, failed to understand the response from the AI model."
+            except Exception as e:
+                logger.error(f"Error processing successful OpenRouter response: {e}", exc_info=True)
+                return "Sorry, an error occurred while processing the AI model's response."
+
+        elif response.status_code == 401:
+            logger.error("OpenRouter API key is invalid (401 Unauthorized). Check Replit Secrets.")
+            return "Error: The AI model configuration key (OpenRouter) is invalid."
+        elif response.status_code == 402:
+            logger.error("OpenRouter Payment Required (402). Check credits/limits on OpenRouter.")
+            return "Sorry, there might be an issue with the AI model service limits or payment. Please try again later or check OpenRouter account."
+        elif response.status_code == 429:
+            logger.warning("OpenRouter Rate Limit Exceeded (429).")
+            return "Sorry, the AI model is currently busy due to high demand. Please try again in a moment."
+        elif response.status_code == 500:
+             logger.error(f"OpenRouter Internal Server Error (500). Response: {response.text[:500]}...")
+             return "Sorry, the AI model service encountered an internal error. Please try again later."
+        else:
+             # Handle other potential errors (e.g., 400 Bad Request, 404 Not Found for model)
+             logger.error(f"Unexpected status code {response.status_code} from OpenRouter. Response: {response.text[:500]}...")
+             try: # Try to parse error message from response body
+                 error_data = response.json()
+                 error_msg = error_data.get("error", {}).get("message", response.text[:100])
+                 return f"Sorry, the AI model service returned an error ({response.status_code}): {error_msg}"
+             except: # Fallback if parsing fails
+                return f"Sorry, the AI model service returned an unexpected status ({response.status_code})."
+
+    except requests.exceptions.Timeout:
+        logger.error("Timeout error connecting to OpenRouter API.")
+        return "Sorry, the request to the AI model timed out. Please try again."
+    except requests.exceptions.RequestException as e:
+        logger.error(f"Request error connecting to OpenRouter API: {e}")
+        return "Sorry, there was an error connecting to the AI model service."
+    except Exception as e:
+        logger.error(f"Unexpected error in generate_summary (OpenRouter): {e}", exc_info=True)
+        return "Sorry, an unexpected error occurred while generating the summary."
+
+
+# --- Telegram Bot Handlers (Command, Message, CallbackQuery) ---
+
+async def start(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
+    """Sends a welcome message when the /start command is issued."""
+    user = update.effective_user
+    logger.info(f"User {user.id} ({user.username or 'NoUsername'}) used /start.")
+    mention = user.mention_html() if user.username else user.first_name
+    await update.message.reply_html(
+        f"👋 Hello {mention}! I can summarize YouTube links or website URLs.\n\n"
+        "Just send me a link anytime!",
+        # Optional: disable_web_page_preview=True
+    )
+
+async def help_command(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
+    """Sends a help message when the /help command is issued."""
+    logger.info(f"User {update.effective_user.id} used /help.")
+    await update.message.reply_text(
+        "🔍 **How to use this bot:**\n\n"
+        "1. Send me any YouTube video link or website URL.\n"
+        "2. I'll ask how you want it summarized (paragraph or points).\n"
+        "3. Click the button for your choice.\n"
+        "4. Wait for the summary!\n\n"
+        "I use multiple methods if the first fails (especially for YT transcripts & website content).\n\n"
+        "**Commands:**\n"
+        "/start - Display welcome message\n"
+        "/help - Show this help message",
+        parse_mode=ParseMode.MARKDOWN
+    )
+
+async def handle_potential_url(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
+    """Handles text messages, checks for URLs, and asks for summary type."""
+    if not update.message or not update.message.text: return # Ignore empty messages
+    url = update.message.text.strip()
+    user = update.effective_user
+    logger.info(f"User {user.id} ({user.username or 'NoUsername'}) sent potential URL: {url}")
+
+    # Basic URL validation
+    if not (url.startswith('http://') or url.startswith('https://')) or '.' not in url[8:]:
+        logger.debug(f"Ignoring non-URL message from user {user.id}: {url}")
+        # Optional: Reply if you want to guide the user
+        # await update.message.reply_text("Please send a valid URL starting with http:// or https://")
+        return
+
+    # Store URL in user_data (simple state management)
+    context.user_data['url_to_summarize'] = url
+    logger.debug(f"Stored URL '{url}' for user {user.id} in user_data")
+
+    # Ask for summary type with Inline Keyboard
+    keyboard = [
+        [
+            InlineKeyboardButton("Paragraph Summary", callback_data="paragraph"),
+            InlineKeyboardButton("Points Summary", callback_data="points")
+        ]
+    ]
+    reply_markup = InlineKeyboardMarkup(keyboard)
+    await update.message.reply_text(
+        f"Okay, I see this link:\n{url}\n\nHow would you like it summarized?",
+        reply_markup=reply_markup,
+        disable_web_page_preview=True
+    )
+
+async def handle_summary_type_callback(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
+    """Handles button presses for summary type selection."""
+    query = update.callback_query
+    if not query: return
+    await query.answer() # Acknowledge button press immediately
+
+    summary_type = query.data
+    user = update.effective_user or query.from_user # Get user info
+    url = context.user_data.get('url_to_summarize', None) # Retrieve stored URL
+
+    logger.info(f"User {user.id} chose '{summary_type}' summary. Checking for URL '{url}' in context.")
+
+    # Check if URL is still in context (it might expire or be lost)
+    if not url:
+         logger.warning(f"User {user.id} pressed button, but NO URL found in user_data context.")
+         try:
+             # Edit the message where the button was, informing the user
+             await query.edit_message_text(text="Sorry, I couldn't find the URL associated with this request. Please send the link again.")
+         except Exception as edit_err:
+             # If editing fails (e.g., message too old), log it but don't crash
+             logger.warning(f"Failed to edit message for missing URL context: {edit_err}")
+             # Maybe send a new message as a fallback? Depends on desired behavior.
+             # await context.bot.send_message(chat_id=user.id, text="Sorry, context lost. Please send link again.")
+         return # Stop processing if URL is missing
+
+    # Clear the URL from context now that we're processing it
+    context.user_data.pop('url_to_summarize', None)
+    logger.debug(f"Retrieved and cleared URL {url} from user_data for user {user.id}")
+
+    # --- Get API Keys (Read fresh from environment - cheap operation) ---
+    # This ensures if secrets are updated in Replit UI, the next request uses them
+    current_openrouter_key = os.environ.get('OPENROUTER_API_KEY')
+    current_urltotext_key = os.environ.get('URLTOTEXT_API_KEY')
+    current_supadata_key = os.environ.get('SUPADATA_API_KEY')
+    current_apify_token = os.environ.get('APIFY_API_TOKEN')
+
+    # Check *essential* key for summarization
+    if not current_openrouter_key:
+         logger.error("OpenRouter API key missing in handler. Cannot generate summary.")
+         # Inform user and clean up the button message
+         await context.bot.send_message(chat_id=user.id, text="Error: AI model configuration key (OpenRouter) is missing. Cannot generate summary.")
+         try: await query.delete_message() # Delete the message with buttons
+         except Exception: pass
+         return
+
+    # --- Start Processing ---
+    processing_message = f"Got it! Generating '{summary_type}' summary for:\n{url}\n\nThis might take a moment..."
+    message_to_delete_later = None # In case editing fails
+    try:
+        # Edit the message to show processing status
+        await query.edit_message_text(processing_message)
+    except Exception as e:
+        # If editing fails (e.g., message too old), send a new status message
+        logger.warning(f"Could not edit original message: {e}, sending new status message.")
+        try:
+            message_to_delete_later = await context.bot.send_message(chat_id=user.id, text=processing_message)
+        except Exception as send_err:
+            # If even sending fails, log and give up on this request
+            logger.error(f"Failed to send status message after edit failure: {send_err}")
+            return
+
+    content = None
+    user_feedback_message = None # Stores error messages for the user
+    success = False
+    is_youtube = is_youtube_url(url)
+
+    try:
+        # Show "typing..." status in Telegram chat
+        await context.bot.send_chat_action(chat_id=user.id, action='typing')
+
+        # --- Content Fetching Logic ---
+        if is_youtube:
+            video_id = extract_youtube_id(url)
+            if video_id:
+                # Fetch YT transcript using the function with fallbacks
+                content = await get_youtube_transcript(
+                    video_id,
+                    url, # Pass full URL for Apify
+                    current_supadata_key,
+                    current_apify_token
+                )
+                # Set feedback message only if content fetching failed
+                user_feedback_message = None if content else "Sorry, I couldn't get the transcript for that YouTube video using any available method (unavailable/private/no captions?)."
+            else:
+                user_feedback_message = "Sorry, I couldn't understand that YouTube URL format."
+        else: # Website Logic (Requests/BS4 -> URLToText API)
+            logger.info(f"Attempting website scrape (Requests/BS4) for {url}")
+            content = await get_website_content_via_requests(url)
+
+            if content:
+                logger.info("Primary website scraping (Requests/BS4) successful.")
+                user_feedback_message = None
+            else:
+                logger.warning(f"Primary web scraping failed for {url}. Attempting fallback API (URLToText).")
+                if current_urltotext_key:
+                    await context.bot.send_chat_action(chat_id=user.id, action='typing') # Show activity for fallback
+                    content = await get_website_content_via_urltotext_api(url, current_urltotext_key)
+                    if content:
+                        user_feedback_message = None
+                        logger.info("Fallback URLToText API scraping successful.")
+                    else:
+                        user_feedback_message = "Sorry, I couldn't fetch content from that website using either method (blocked/inaccessible/empty?)."
+                        logger.error(f"Both primary (Requests/BS4) and fallback API failed for website {url}.")
+                else:
+                    # Primary failed, and fallback key is missing
+                    user_feedback_message = "Sorry, I couldn't fetch content from that website (blocked/inaccessible/empty?). The fallback method is not configured."
+                    logger.warning(f"Primary web scraping failed for {url}, and fallback API key (URLToText) is missing.")
+        # --- End Content Fetching ---
+
+        # --- Generate Summary if Content was Fetched ---
+        if content:
+            logger.info("Content fetched successfully, proceeding to generate summary.")
+            await context.bot.send_chat_action(chat_id=user.id, action='typing')
+            # Pass the OpenRouter key to the summary function
+            summary = await generate_summary(content, summary_type, current_openrouter_key)
+
+            # Check if summary generation returned an error message
+            if summary.startswith("Error:") or summary.startswith("Sorry,"):
+                 user_feedback_message = summary # Use the error from the summary function
+                 success = False
+                 logger.warning(f"Summary generation failed or returned error: {summary}")
+            else:
+                 # Send the successful summary
+                 await context.bot.send_message(
+                     chat_id=user.id,
+                     text=summary,
+                     parse_mode=ParseMode.MARKDOWN,
+                     disable_web_page_preview=True
+                 )
+                 success = True
+                 user_feedback_message = None # Clear any previous failure message from fetching stage
+        elif not user_feedback_message:
+             # If content is None but no specific error message was set during fetching
+             user_feedback_message = "Sorry, couldn't retrieve any content to summarize from the provided link."
+             logger.warning(f"Content fetching resulted in None for {url}, but no specific user feedback message was set.")
+
+        # --- Send Feedback if any step failed ---
+        if user_feedback_message and not success:
+            await context.bot.send_message(chat_id=user.id, text=user_feedback_message)
+
+    except Exception as e:
+        # Catch unexpected errors during the whole process
+        logger.error(f"Unexpected error during processing callback for {url}: {e}", exc_info=True)
+        try:
+            # Send a generic error message to the user
+            await context.bot.send_message(chat_id=user.id, text="Oops! Something went really wrong while processing your request. Please try again later.")
+        except Exception as final_err:
+             # If even sending the error message fails... log it.
+             logger.error(f"Failed to send final error message to user {user.id}: {final_err}")
+    finally:
+        # --- Cleanup ---
+        # Delete the "Processing..." status message or the original message with buttons
+        try:
+             if message_to_delete_later: # If we sent a separate status message
+                 await context.bot.delete_message(chat_id=user.id, message_id=message_to_delete_later.message_id)
+             elif query: # Otherwise, delete the original message with the buttons
+                 # We might have already edited it, but deleting ensures cleanup
+                 await query.delete_message()
+        except Exception as del_e:
+             # Log if deletion fails, but don't let it stop anything
+             logger.warning(f"Could not delete status/button message: {del_e}")
+
+
+async def error_handler(update: object, context: ContextTypes.DEFAULT_TYPE) -> None:
+    """Log Errors caused by Updates."""
+    logger.error("Exception while handling an update:", exc_info=context.error)
+    # Optionally, notify the developer or user about the error
+    # developer_chat_id = 12345678 # Replace with your chat ID if desired
+    # if update and hasattr(update, 'effective_chat') and update.effective_chat:
+    #     chat_id = update.effective_chat.id
+    #     error_message = f"Error processing update {update.update_id} for chat {chat_id}. Error: {context.error}"
+    # else:
+    #     error_message = f"Error in handler: {context.error}"
+    # try:
+    #     await context.bot.send_message(chat_id=developer_chat_id, text=error_message)
+    # except Exception as e:
+    #     logger.error(f"Failed to send error notification to developer: {e}")
+
+
+# --- Initialize Telegram Bot Application ---
+# Do this setup once when the script starts
+logger.info("Initializing Telegram Application...")
+if not TELEGRAM_TOKEN: # Check again, as initialization needs it
+     logger.critical("Cannot initialize PTB Application without TELEGRAM_TOKEN.")
+     # Consider how to handle this in deployment - maybe the Flask app still runs but logs errors?
+     ptb_app = None
+else:
+    ptb_app = Application.builder().token(TELEGRAM_TOKEN).build()
+
+    # Register handlers with the PTB application instance
+    ptb_app.add_handler(CommandHandler("start", start))
+    ptb_app.add_handler(CommandHandler("help", help_command))
+    ptb_app.add_handler(MessageHandler(filters.TEXT & ~filters.COMMAND, handle_potential_url))
+    ptb_app.add_handler(CallbackQueryHandler(handle_summary_type_callback))
+    # Add the error handler
+    ptb_app.add_error_handler(error_handler)
+    logger.info("Telegram handlers registered.")
+
+
+# --- Flask App Setup ---
+app = Flask(__name__) # Create Flask web server instance
+logger.info("Flask app created.")
+
+
+# --- Webhook Endpoint ---
+@app.route('/webhook', methods=['POST'])
+async def webhook() -> Response:
+    """Webhook endpoint to receive updates from Telegram"""
+    logger.info("Webhook received request...")
+    if not ptb_app: # Check if PTB initialization failed
+        logger.error("Telegram application not initialized. Cannot process update.")
+        return Response('Bot not configured', status=500)
+
+    if request.is_json:
+        try:
+            update_data = request.get_json()
+            # Use PTB's built-in deserialization
+            update = Update.de_json(update_data, ptb_app.bot)
+            logger.debug(f"Processing update ID: {update.update_id}")
+
+            # Process the update using PTB's internal dispatcher
+            # This ensures all handlers, context, etc., work as expected
+            # Running it in a background task allows Flask to respond quickly
+            # If Replit handles concurrency well, this might not be strictly needed,
+            # but it's safer practice for potentially long-running handlers.
+            asyncio.create_task(ptb_app.process_update(update))
+
+            # Respond quickly to Telegram that we received the update
+            return Response('ok', status=200)
+        except json.JSONDecodeError:
+            logger.error("Failed to decode JSON from Telegram webhook.")
+            return Response('Bad Request - Invalid JSON', status=400)
+        except Exception as e:
+            logger.error(f"Error processing update in webhook handler: {e}", exc_info=True)
+            return Response('Internal Server Error', status=500)
+    else:
+        logger.warning("Received non-JSON request to webhook endpoint.")
+        return Response('Bad Request - Expected JSON', status=400)
+
+@app.route('/')
+def index():
+    """A simple health check endpoint for the web server"""
+    logger.debug("Health check endpoint '/' accessed.")
+    # You could add more checks here if needed
+    bot_status = "PTB App Initialized" if ptb_app else "PTB App FAILED Initialization"
+    return f"Hello! Telegram Bot Webhook Listener ({bot_status}) is running."
+
+
+# --- Main Execution Block ---
+# This part runs the Flask web server when you execute `python main.py`
+if __name__ == '__main__':
+    if not ptb_app:
+        logger.critical("Aborting Flask server start because Telegram Application failed to initialize (Missing Token?).")
+    else:
+        logger.info("Starting Flask web server for Telegram bot webhook...")
+        # Replit expects apps to listen on 0.0.0.0 and often injects the PORT environment variable
+        # Defaulting to 8080 if PORT isn't set by Replit.
+        port = int(os.environ.get('PORT', 8080))
+        # Use app.run for development/simple deployment.
+        # For production, tools like Gunicorn or Uvicorn are often used,
+        # but Replit's runner usually handles this adequately.
+        app.run(host='0.0.0.0', port=port, debug=False) # Set debug=False for deployment

requirements.txt

@@ -0,0 +1,9 @@
+python-telegram-bot[job-queue]>=21.0
+requests
+beautifulsoup4
+youtube-transcript-api
+apify-client # Include if you have the token
+Flask>=2.0
+nest_asyncio
+# Add gunicorn for a production-ready web server
+gunicorn