Update app.py

app.py

@@ -1,20 +1,21 @@
 import os
-import gradio as gr
 import asyncio
 import logging
 from telethon import TelegramClient
 from huggingface_hub import upload_file
 from dotenv import load_dotenv
+from flask import Flask, request, render_template, jsonify
+import threading
 
 # === Load secrets from .env ===
 load_dotenv()
 
-API_ID = int(os.getenv("API_ID"))
+API_ID = os.getenv("API_ID")
 API_HASH = os.getenv("API_HASH")
 HF_TOKEN = os.getenv("HF_TOKEN")
 CHANNEL = os.getenv("CHANNEL_USERNAME")
 REPO_ID = os.getenv("DATASET_REPO")
-DATA_PATH = "AEffects"
+DATA_PATH = "telegram_uploads"
 
 # === Logging setup ===
 logging.basicConfig(
@@ -26,14 +27,23 @@ logging.basicConfig(
     ]
 )
 
-# === Setup Telegram Client ===
-client = TelegramClient("my_session", API_ID, API_HASH)
+# === Setup Telegram Client (only if credentials are available) ===
+client = None
+if API_ID and API_HASH:
+    try:
+        client = TelegramClient("my_session", int(API_ID), API_HASH)
+    except ValueError as e:
+        logging.error(f"Failed to initialize Telegram client: {e}")
+        client = None
 
 # === Ensure download folder exists ===
 os.makedirs("downloads", exist_ok=True)
 
 # === Upload wrapper ===
 def upload_to_dataset(filepath):
+    if not HF_TOKEN or not REPO_ID:
+        return False, f"❌ Missing Hugging Face credentials"
+    
     try:
         upload_file(
             path_or_fileobj=filepath,
@@ -49,62 +59,127 @@ def upload_to_dataset(filepath):
         return False, f"❌ Upload failed: {os.path.basename(filepath)} — {e}"
 
 # === Main file processing logic ===
-async def process_filenames(name_input, progress=None):
-    await client.start()
-
-    filenames = [name.strip().lower() for name in name_input.replace(",", "\n").splitlines() if name.strip()]
-    results = []
-    found = set()
-
-    messages = [msg async for msg in client.iter_messages(CHANNEL, limit=300)]
-    total = len(messages)
-
-    for i, msg in enumerate(messages):
-        if progress:
-            progress(i / total, desc=f"Scanning message {i+1}/{total}...")
-
-        if msg.media and msg.file:
-            fname = msg.file.name or f"file_{msg.id}{msg.file.ext}"
-
-            for search in filenames:
-                if search in fname.lower() and fname not in found:
-                    found.add(fname)
-                    path = f"downloads/{fname}"
-
-                    if not os.path.exists(path):
-                        await msg.download_media(file=path)
-                        success, msg_text = upload_to_dataset(path)
-                        results.append(msg_text)
-                    else:
-                        results.append(f"⏩ Already exists: {fname}")
-                    break
-
-    matched_names = [r.split(":")[-1].strip().lower() for r in results if r.startswith("✅") or r.startswith("⏩")]
-    for name in filenames:
-        if not any(name in matched for matched in matched_names):
-            results.append(f"❌ Not found: {name}")
-
-    return "\n".join(results) if results else "❌ No files matched."
-
-# === Gradio async handler ===
-async def handle_upload(name_input):
-    with gr.Progress() as p:
-        return await process_filenames(name_input, p)
-
-# === Gradio UI ===
-def launch_ui():
-    with gr.Blocks() as app:
-        gr.Markdown("## Hugging Face Uploader")
-        input_box = gr.Textbox(
-            label="Enter filenames (comma or newline separated)",
-            lines=6,
-            placeholder="e.g.\nreport, summary.pdf\nmeeting_notes"
-        )
-        result_box = gr.Textbox(label="Upload Results", lines=12)
-        btn = gr.Button("Start Upload")
-        btn.click(fn=handle_upload, inputs=input_box, outputs=result_box)
-    return app
-
-# === Launch ===
-if __name__ == "__main__":
-    launch_ui().launch(server_name="0.0.0.0", server_port=7860)
+async def process_filenames(name_input):
+    if not client:
+        return "❌ Error: Telegram client not initialized. Please check your API credentials."
+    
+    if not CHANNEL:
+        return "❌ Error: Channel username not configured."
+    
+    try:
+        await client.start()
+
+        filenames = [name.strip().lower() for name in name_input.replace(",", "\n").splitlines() if name.strip()]
+        results = []
+        found = set()
+
+        messages = [msg async for msg in client.iter_messages(CHANNEL, limit=300)]
+        total = len(messages)
+
+        for i, msg in enumerate(messages):
+            if msg.media and msg.file:
+                fname = msg.file.name or f"file_{msg.id}{msg.file.ext}"
+
+                for search in filenames:
+                    if search in fname.lower() and fname not in found:
+                        found.add(fname)
+                        path = f"downloads/{fname}"
+
+                        if not os.path.exists(path):
+                            await msg.download_media(file=path)
+                            success, msg_text = upload_to_dataset(path)
+                            results.append(msg_text)
+                        else:
+                            results.append(f"⏩ Already exists: {fname}")
+                        break
+
+        matched_names = [r.split(":")[-1].strip().lower() for r in results if r.startswith("✅") or r.startswith("⏩")]
+        for name in filenames:
+            if not any(name in matched for matched in matched_names):
+                results.append(f"❌ Not found: {name}")
+
+        return "\n".join(results) if results else "❌ No files matched."
+    
+    except Exception as e:
+        logging.error(f"Error in process_filenames: {e}")
+        return f"❌ Error: {str(e)}"
+
+def run_async_in_thread(coro):
+    """Run async function in a separate thread with its own event loop"""
+    def run_in_thread():
+        loop = asyncio.new_event_loop()
+        asyncio.set_event_loop(loop)
+        try:
+            return loop.run_until_complete(coro)
+        finally:
+            loop.close()
+    
+    import concurrent.futures
+    with concurrent.futures.ThreadPoolExecutor() as executor:
+        future = executor.submit(run_in_thread)
+        return future.result()
+
+app = Flask(__name__)
+
+@app.route('/')
+def index():
+    return render_template('index.html')
+
+@app.route('/upload', methods=['POST'])
+def upload():
+    try:
+        filenames_input = request.form.get('filenames', '').strip()
+        if not filenames_input:
+            return "❌ Error: No filenames provided", 400
+        
+        # Check if credentials are configured
+        if not client:
+            return "❌ Error: Application not configured. Please set up your .env file with API credentials.", 500
+        
+        # Run the async function in a separate thread
+        results = run_async_in_thread(process_filenames(filenames_input))
+        return results
+    
+    except Exception as e:
+        logging.error(f"Error in upload route: {e}")
+        return f"❌ Error: {str(e)}", 500
+
+@app.route('/health')
+def health():
+    status = {
+        "status": "healthy",
+        "message": "Hugging Face Uploader is running",
+        "configured": {
+            "telegram": bool(client),
+            "huggingface": bool(HF_TOKEN and REPO_ID),
+            "channel": bool(CHANNEL)
+        }
+    }
+    return jsonify(status)
+
+@app.route('/config')
+def config():
+    """Show configuration status"""
+    config_status = {
+        "API_ID": "✅ Set" if API_ID else "❌ Missing",
+        "API_HASH": "✅ Set" if API_HASH else "❌ Missing", 
+        "HF_TOKEN": "✅ Set" if HF_TOKEN else "❌ Missing",
+        "CHANNEL_USERNAME": "✅ Set" if CHANNEL else "❌ Missing",
+        "DATASET_REPO": "✅ Set" if REPO_ID else "❌ Missing"
+    }
+    
+    return jsonify(config_status)
+
+if __name__ == '__main__':
+    print("Starting Hugging Face Uploader...")
+    print("Configuration status:")
+    print(f"  API_ID: {'✅ Set' if API_ID else '❌ Missing'}")
+    print(f"  API_HASH: {'✅ Set' if API_HASH else '❌ Missing'}")
+    print(f"  HF_TOKEN: {'✅ Set' if HF_TOKEN else '❌ Missing'}")
+    print(f"  CHANNEL_USERNAME: {'✅ Set' if CHANNEL else '❌ Missing'}")
+    print(f"  DATASET_REPO: {'✅ Set' if REPO_ID else '❌ Missing'}")
+    print("\nTo configure, copy .env.example to .env and fill in your credentials.")
+    print("Visit http://localhost:5000 to use the application.")
+    
+    app.run(host='0.0.0.0', port=5000, debug=True)
+