Update app.py

app.py

@@ -1,24 +1,54 @@
 import os
 import gradio as gr
 import asyncio
+import logging
 from telethon import TelegramClient
-from huggingface_hub import HfApi
+from huggingface_hub import upload_file
 from dotenv import load_dotenv
 
+# === Load secrets from .env ===
 load_dotenv()
 
-api_id = int(os.getenv("API_ID"))
-api_hash = os.getenv("API_HASH")
-hf_token = os.getenv("HF_TOKEN")
-channel = os.getenv("CHANNEL_USERNAME")
-dataset_repo = os.getenv("DATASET_REPO")
-
-client = TelegramClient("my_session", api_id, api_hash)
-hf_api = HfApi()
-hf_api.set_access_token(hf_token)
-
+API_ID = int(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 = "telegram_uploads"
+
+# === Logging setup ===
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(asctime)s — %(levelname)s — %(message)s",
+    handlers=[
+        logging.FileHandler("upload.log"),
+        logging.StreamHandler()
+    ]
+)
+
+# === Setup Telegram Client ===
+client = TelegramClient("my_session", API_ID, API_HASH)
+
+# === Ensure download folder exists ===
 os.makedirs("downloads", exist_ok=True)
 
+# === Upload wrapper ===
+def upload_to_dataset(filepath):
+    try:
+        upload_file(
+            path_or_fileobj=filepath,
+            path_in_repo=f"{DATA_PATH}/{os.path.basename(filepath)}",
+            repo_id=REPO_ID,
+            repo_type="dataset",
+            token=HF_TOKEN
+        )
+        logging.info(f"[↑] Uploaded: {filepath}")
+        return True, f"✅ Uploaded: {os.path.basename(filepath)}"
+    except Exception as e:
+        logging.error(f"[!] Upload failed: {filepath} — {e}")
+        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()
 
@@ -26,7 +56,7 @@ async def process_filenames(name_input, progress=None):
     results = []
     found = set()
 
-    messages = [msg async for msg in client.iter_messages(channel, limit=300)]
+    messages = [msg async for msg in client.iter_messages(CHANNEL, limit=300)]
     total = len(messages)
 
     for i, msg in enumerate(messages):
@@ -43,15 +73,8 @@ async def process_filenames(name_input, progress=None):
 
                     if not os.path.exists(path):
                         await msg.download_media(file=path)
-                        hf_api.upload_file(
-                            path_or_fileobj=path,
-                            path_in_repo=os.path.basename(path),
-                            repo_id=dataset_repo,
-                            repo_type="dataset",
-                            token=hf_token  # Pass token here
-                        )
-
-                        results.append(f"✅ Uploaded: {fname}")
+                        success, msg_text = upload_to_dataset(path)
+                        results.append(msg_text)
                     else:
                         results.append(f"⏩ Already exists: {fname}")
                     break
@@ -63,17 +86,23 @@ async def process_filenames(name_input, progress=None):
 
     return "\n".join(results) if results else "❌ No files matched."
 
+# === Gradio UI handler ===
 def handle_upload(name_input):
     with gr.Progress() as p:
         return asyncio.run(process_filenames(name_input, p))
 
+# === Gradio UI ===
 def launch_ui():
     with gr.Blocks() as app:
-        gr.Markdown("## 📥 Telegram File Fetcher → 📤 HF Dataset Uploader")
-        input_box = gr.Textbox(label="Enter filenames (comma or newline separated)", lines=6)
-        result_box = gr.Textbox(label="Results", lines=10)
-        go = gr.Button("Start Upload")
-        go.click(fn=handle_upload, inputs=input_box, outputs=result_box)
+        gr.Markdown("## 📥 Telegram File Fetcher → 📤 Hugging Face Dataset 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
 
 if __name__ == "__main__":