Update app.py

app.py

@@ -1,4 +1,4 @@
-import gradio as gr
+mport gradio as gr
 from gtts import gTTS
 import time
 import difflib
@@ -14,41 +14,17 @@ def play_text(text):
     os.system(f"start {temp_file.name}")  # Windows
     return "✅ Text is being read out. Please listen and read it yourself."
 
+# Function to transcribe user's audio
 def transcribe_audio(audio, original_text):
     recognizer = sr.Recognizer()
     with sr.AudioFile(audio) as source:
         audio_data = recognizer.record(source)
-    
     try:
         start_time = time.time()
-        
-        # Debug: Check audio data length and properties
-        print(f"Audio Data Length: {len(audio_data.frame_data)}")
-        print(f"Sample Rate: {audio_data.sample_rate}")
-        
-        # Split the audio into chunks (1-minute chunks in this example)
-        audio_length = len(audio_data.frame_data)
-        chunk_size = 60000  # 1 minute (60,000 ms)
-        
-        # Splitting audio data into chunks
-        chunks = [audio_data.frame_data[i:i+chunk_size] for i in range(0, audio_length, chunk_size)]
-        print(f"Number of chunks: {len(chunks)}")
-        
-        transcription = ""
-        for chunk in chunks:
-            audio_chunk = sr.AudioData(chunk, audio_data.sample_rate, audio_data.sample_width)
-            
-            # Debug: Print chunk size
-            print(f"Transcribing Chunk: {len(chunk)}")
-            
-            # Using Google Speech Recognition (supports Hindi)
-            transcription += recognizer.recognize_google(audio_chunk, language="hi-IN") + " "
-        
+        # Using Google Speech Recognition (supports Hindi)
+        transcription = recognizer.recognize_google(audio_data, language="hi-IN")
         end_time = time.time()
         
-        # Debug: Output full transcription
-        print(f"Transcribed Text: {transcription}")
-        
         # Calculate Accuracy
         original_words = original_text.strip().split()
         transcribed_words = transcription.strip().split()
@@ -57,26 +33,19 @@ def transcribe_audio(audio, original_text):
         
         # Calculate speed
         duration = end_time - start_time  # time to process (not speaking time)
-        speed = round(len(transcribed_words) / duration, 2)  # words per second
+        # Better: estimate speaking time from audio length if needed (advanced)
         
-        # Compare words and highlight mistakes
-        wrong_words = []
-        for i, word in enumerate(original_words):
-            if i >= len(transcribed_words) or word != transcribed_words[i]:
-                wrong_words.append(f"🔴 {word}")
+        speed = round(len(transcribed_words) / duration, 2)  # words per second
         
         result = {
             "📝 Transcribed Text": transcription,
             "🎯 Accuracy (%)": accuracy,
-            "⏱️ Speaking Speed (words/sec)": speed,
-            "❌ Incorrect Words": ' '.join(wrong_words) if wrong_words else "None"
+            "⏱️ Speaking Speed (words/sec)": speed
         }
         return result
     except Exception as e:
         return {"error": str(e)}
 
-        return {"error": str(e)}
-
 # Gradio App
 with gr.Blocks() as app:
     gr.Markdown("## 🗣️ Hindi Reading & Pronunciation Practice App")
@@ -98,3 +67,5 @@ with gr.Blocks() as app:
 
 # Launch the app
 app.launch()
+
+