Update app.py

app.py

@@ -1,41 +1,55 @@
 import gradio as gr
-from transformers import MarianMTModel, MarianTokenizer, pipeline
-import pyttsx3  # Text-to-Speech for feedback
+from transformers import pipeline
+import pyttsx3
+import speech_recognition as sr
 
-# Initialize Translation Model
-model_name = 'Helsinki-NLP/opus-mt-en-ur'
-tokenizer = MarianTokenizer.from_pretrained(model_name)
-model = MarianMTModel.from_pretrained(model_name)
-translator = pipeline('translation', model=model, tokenizer=tokenizer)
+# Initialize translation model (English <-> Urdu)
+translation_model = pipeline("translation_en_to_ur", model="Helsinki-NLP/opus-mt-en-ur")
+reverse_translation_model = pipeline("translation_ur_to_en", model="Helsinki-NLP/opus-mt-ur-en")
 
-# Initialize Text-to-Speech Engine
+# Initialize text-to-speech engine
 engine = pyttsx3.init()
 
-# Function to translate text
-def translate_text(input_text, lang_direction):
-    if lang_direction == 'English to Urdu':
-        translation = translator(input_text, max_length=100)[0]['translation_text']
+# Function to translate text and provide feedback
+def translate_and_speak(text, direction):
+    if direction == "English to Urdu":
+        translated_text = translation_model(text)[0]['translation_text']
     else:
-        translation = translator(input_text, max_length=100)[0]['translation_text']
-    return translation
-
-# Function to provide pronunciation feedback (using TTS)
-def provide_pronunciation_feedback(text, lang):
-    engine.say(text)
+        translated_text = reverse_translation_model(text)[0]['translation_text']
+    
+    # Use TTS to speak the translated text
+    engine.say(translated_text)
     engine.runAndWait()
+    
+    return translated_text
+
+# Function for Speech-to-Text
+def speech_to_text(audio_file, direction):
+    # Recognize speech from audio
+    recognizer = sr.Recognizer()
+    audio = sr.AudioFile(audio_file.name)
+    with audio as source:
+        audio_data = recognizer.record(source)
+    
+    try:
+        text = recognizer.recognize_google(audio_data)
+    except sr.UnknownValueError:
+        text = "Sorry, could not understand the audio."
+    except sr.RequestError:
+        text = "Could not request results from Google Speech Recognition service."
+    
+    # Translate text
+    return translate_and_speak(text, direction)
 
-# Gradio Interface
-def translate_and_practice(sentence, lang_direction):
-    translation = translate_text(sentence, lang_direction)
-    provide_pronunciation_feedback(translation, 'ur' if lang_direction == 'English to Urdu' else 'en')
-    return translation
+# Gradio interface
+iface = gr.Interface(
+    fn=translate_and_speak, 
+    inputs=[gr.Textbox(label="Enter Text"), 
+            gr.Radio(choices=["English to Urdu", "Urdu to English"], label="Translation Direction")],
+    outputs="text", 
+    live=True,
+    title="AI-Powered Language Tutor",
+    description="An interactive tutor to help you practice English-Urdu translations with speech feedback!"
+)
 
-# UI with Gradio
-iface = gr.Interface(fn=translate_and_practice,
-                     inputs=[gr.Textbox(label="Enter Sentence"), 
-                             gr.Radio(choices=["English to Urdu", "Urdu to English"], label="Language Direction")],
-                     outputs="text",
-                     live=True,  # To enable real-time updates
-                     title="English-Urdu Language Tutor",
-                     description="Enter a sentence and get the translation with pronunciation feedback.")
 iface.launch()