Update app.py

app.py

@@ -5,17 +5,15 @@ import yt_dlp
 import subprocess
 from huggingface_hub import InferenceClient
 
-# ✅ Initialize Hugging Face client
-client = InferenceClient("openai/whisper-large-v3")
-
-# Summarization and tutorial generation models
+# ✅ Initialize clients
+whisper_client = InferenceClient("openai/whisper-large-v3")
 english_summarizer = InferenceClient("facebook/bart-large-cnn")
 urdu_summarizer = InferenceClient("openai/gpt-oss-120b")
 
 # --- Helper Functions ---
 
 def download_youtube_audio(url):
-    """Download audio from YouTube video"""
+    """Download YouTube audio using yt_dlp"""
     try:
         with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp:
             ydl_opts = {
@@ -34,8 +32,11 @@ def convert_to_wav(audio_path):
     """Ensure audio is in .wav format"""
     wav_path = tempfile.mktemp(suffix=".wav")
     try:
-        subprocess.run(["ffmpeg", "-y", "-i", audio_path, "-ar", "16000", "-ac", "1", wav_path],
-                       check=True, capture_output=True)
+        subprocess.run(
+            ["ffmpeg", "-y", "-i", audio_path, "-ar", "16000", "-ac", "1", wav_path],
+            check=True,
+            capture_output=True
+        )
         return wav_path
     except subprocess.CalledProcessError as e:
         raise RuntimeError(f"❌ ffmpeg conversion failed: {e}")
@@ -51,12 +52,11 @@ def transcribe_audio(audio_file=None, youtube_url=None):
         wav_path = convert_to_wav(audio_path)
 
         with open(wav_path, "rb") as f:
-            text = client.text_generation(
-                prompt="Transcribe this English audio accurately:",
-                inputs=f.read(),
-                max_new_tokens=5000,
-            )
-        return text
+            audio_data = f.read()
+
+        # ✅ Correct way to call Whisper for transcription
+        response = whisper_client.post(json=None, data=audio_data)
+        return response.get("text", "❌ No transcription returned.")
     except Exception as e:
         return f"❌ Error during transcription: {e}"
 
@@ -65,13 +65,13 @@ def summarize_text(transcribed_text, language):
     try:
         if language == "English":
             response = english_summarizer.text_generation(
-                prompt="Summarize this text comprehensively:\n" + transcribed_text,
+                prompt=f"Summarize the following text in detail:\n\n{transcribed_text}",
                 max_new_tokens=1024,
             )
             return response
         else:
             response = urdu_summarizer.text_generation(
-                prompt="مندرجہ ذیل انگریزی متن کا جامع اردو خلاصہ تحریر کریں:\n" + transcribed_text,
+                prompt=f"مندرجہ ذیل انگریزی متن کا جامع اردو خلاصہ لکھیں:\n\n{transcribed_text}",
                 max_new_tokens=2048,
             )
             return response
@@ -83,8 +83,8 @@ def generate_tutorial(transcribed_text, summary, language):
     try:
         model = urdu_summarizer if language == "Urdu" else english_summarizer
         prompt = (
-            f"Write a simple, step-by-step tutorial in {language} for beginners based on this transcript:\n\n"
-            f"Transcript:\n{transcribed_text}\n\nSummary:\n{summary}\n"
+            f"Write a simple, step-by-step tutorial in {language} for absolute beginners "
+            f"based on the following transcript and summary.\n\nTranscript:\n{transcribed_text}\n\nSummary:\n{summary}"
         )
         response = model.text_generation(prompt=prompt, max_new_tokens=2500)
         return response