Update app.py

app.py

@@ -9,6 +9,74 @@ from pathlib import Path
 import re
 import uuid
 import pymupdf
+import telebot
+from threading import Thread
+import requests
+from io import BytesIO
+
+# Add your Telegram bot token as an environment variable in Hugging Face space
+BOT_TOKEN = os.getenv('BOT_TOKEN')
+bot = telebot.TeleBot(BOT_TOKEN)
+
+def download_file_from_telegram(file_id):
+    file_info = bot.get_file(file_id)
+    file_path = file_info.file_path
+    file_url = f'https://api.telegram.org/file/bot{BOT_TOKEN}/{file_path}'
+    response = requests.get(file_url)
+    return BytesIO(response.content)
+
+@bot.message_handler(content_types=['document'])
+def handle_docs(message):
+    try:
+        # Get file from telegram
+        file_id = message.document.file_id
+        file_name = message.document.file_name
+        
+        if not file_name.lower().endswith(('.pdf', '.png', '.jpg', '.jpeg')):
+            bot.reply_to(message, "Please send only PDF or image files.")
+            return
+            
+        # Download file
+        file_data = download_file_from_telegram(file_id)
+        
+        # Save temporarily
+        temp_path = f"/tmp/{file_name}"
+        with open(temp_path, 'wb') as f:
+            f.write(file_data.getvalue())
+            
+        # Process file using your existing function
+        md_content, txt_content, archive_zip_path, new_pdf_path = to_markdown(
+            temp_path, 
+            end_pages=10,  # default value
+            is_ocr=False,  # default value
+            layout_mode="doclayout_yolo",  # default value
+            formula_enable=True,  # default value
+            table_enable=True,  # default value
+            language='auto'  # default value
+        )
+        
+        # Send back results
+        with open(archive_zip_path, 'rb') as zip_file:
+            bot.send_document(message.chat.id, zip_file)
+        
+        # Send markdown content in chunks if it's too long
+        max_length = 4096
+        for i in range(0, len(md_content), max_length):
+            chunk = md_content[i:i + max_length]
+            bot.send_message(message.chat.id, chunk)
+            
+        # Cleanup
+        os.remove(temp_path)
+        
+    except Exception as e:
+        bot.reply_to(message, f"Error processing document: {str(e)}")
+
+@bot.message_handler(commands=['start', 'help'])
+def send_welcome(message):
+    bot.reply_to(message, "Hello! Send me a PDF or image file and I'll process it for you.")
+
+def run_bot():
+    bot.infinity_polling()
 
 # os.system('pip install -U magic-pdf==0.10.5')
 os.system('pip uninstall -y magic-pdf')
@@ -211,6 +279,10 @@ def to_pdf(file_path):
 
 
 if __name__ == "__main__":
+       # Start bot in a separate thread
+    bot_thread = Thread(target=run_bot)
+    bot_thread.start()
+    
     with gr.Blocks() as demo:
         gr.HTML(header)
         with gr.Row():