Update app.py

app.py

@@ -4,6 +4,11 @@ from pydub import AudioSegment
 import tempfile
 from langdetect import detect
 import os
+from telegram import Update
+from telegram.ext import Application, CommandHandler, MessageHandler, filters, ContextTypes
+
+# Telegram bot token (to be set via Hugging Face Space secrets)
+TELEGRAM_BOT_TOKEN = os.getenv(8030235633:AAHKvxM9Nzp0DkxfdotMux3572tC_5CGEUA)
 
 # Process audio and transcribe
 def process_audio(audio_input):
@@ -14,22 +19,25 @@ def process_audio(audio_input):
         if isinstance(audio_input, tuple):  # Recorded audio (sample_rate, numpy_array)
             sample_rate, audio_data = audio_input
             AudioSegment(audio_data, sample_rate=sample_rate, frame_rate=sample_rate, channels=1).export(temp_file.name, format="wav")
-        else:  # Uploaded audio file
-            # Load the uploaded audio file and convert it to WAV
+        else:  # Uploaded audio file (file path or Telegram file)
             audio = AudioSegment.from_file(audio_input)
             audio = audio.set_channels(1)  # Convert to mono for consistency
             audio.export(temp_file.name, format="wav")
         audio_file_path = temp_file.name
 
-    # Transcribe the WAV file
+    # Debug: Check if the WAV file is valid
+    if os.path.getsize(audio_file_path) == 0:
+        raise ValueError("The converted WAV file is empty. The input audio may be corrupted.")
+
+    # Transcribe the WAV file using pocketsphinx (offline)
     with sr.AudioFile(audio_file_path) as source:
         audio = recognizer.record(source)
         try:
-            transcription = recognizer.recognize_google(audio)
+            transcription = recognizer.recognize_sphinx(audio)  # Use pocketsphinx for offline transcription
         except sr.UnknownValueError:
             transcription = "Could not understand the audio."
-        except sr.RequestError:
-            transcription = "Transcription service unavailable."
+        except sr.RequestError as e:
+            transcription = f"Transcription failed: {str(e)}"
 
     # Detect language
     try:
@@ -59,7 +67,33 @@ def audio_transcriptor(audio_file, audio_record):
 
     return language, transcription, text_file
 
-# Custom HTML for styled transcription display
+# Telegram bot handlers
+async def start(update: Update, context: ContextTypes.DEFAULT_TYPE):
+    await update.message.reply_text("Hello! Send me an audio file, and I'll transcribe it for you.")
+
+async def handle_audio(update: Update, context: ContextTypes.DEFAULT_TYPE):
+    # Download the audio file from Telegram
+    audio_file = await update.message.audio.get_file()
+    audio_path = f"/tmp/{audio_file.file_id}.ogg"  # Telegram audio files are typically in OGG format
+    await audio_file.download_to_drive(audio_path)
+
+    # Process the audio using the existing transcriptor function
+    language, transcription, text_file_path = process_audio(audio_path)
+
+    # Send the transcription back to the user
+    await update.message.reply_text(f"Detected Language: {language}\nTranscription: {transcription}")
+
+    # Send the transcription file
+    with open(text_file_path, 'rb') as f:
+        await update.message.reply_document(document=f, filename="transcription.txt")
+
+    # Clean up temporary files
+    if os.path.exists(audio_path):
+        os.remove(audio_path)
+    if os.path.exists(text_file_path):
+        os.remove(text_file_path)
+
+# Custom HTML for styled transcription display (for Gradio interface)
 transcription_html = """
 <div class="transcription-container" id="transcriptionContainer">
     <h2>Transcription Results</h2>
@@ -103,7 +137,7 @@ transcription_html = """
 # Gradio interface
 with gr.Blocks() as demo:
     gr.Markdown("# Audio Transcriptor")
-    gr.Markdown("Upload an audio file or record audio to transcribe the speech and detect the language.")
+    gr.Markdown("Upload an audio file or record audio to transcribe the speech and detect the language. You can also interact with the bot via Telegram!")
 
     with gr.Row():
         audio_file = gr.Audio(sources=["upload"], type="filepath", label="Upload Audio")
@@ -132,5 +166,27 @@ with gr.Blocks() as demo:
         outputs=[audio_file, audio_record]
     )
 
-# Launch Gradio app
-demo.launch()
+# Start the Telegram bot in a separate thread
+def run_telegram_bot():
+    if not TELEGRAM_BOT_TOKEN:
+        print("Telegram bot token not found. Please set TELEGRAM_BOT_TOKEN in the Space secrets.")
+        return
+
+    application = Application.builder().token(TELEGRAM_BOT_TOKEN).build()
+
+    # Add handlers
+    application.add_handler(CommandHandler("start", start))
+    application.add_handler(MessageHandler(filters.AUDIO, handle_audio))
+
+    # Start the bot
+    print("Starting Telegram bot...")
+    application.run_polling(allowed_updates=Update.ALL_TYPES)
+
+# Launch Gradio app and Telegram bot
+if __name__ == "__main__":
+    import threading
+    # Start the Telegram bot in a separate thread
+    bot_thread = threading.Thread(target=run_telegram_bot)
+    bot_thread.start()
+    # Launch Gradio app
+    demo.launch()