Create app.py

app.py

@@ -0,0 +1,78 @@
+import os
+import tempfile
+import torch
+import speech_recognition as sr
+from transformers import pipeline
+from gtts import gTTS
+from flask import Flask, request, jsonify
+import gradio as gr
+
+app = Flask(__name__)
+
+# Load translation model
+translator = pipeline("translation", model="Helsinki-NLP/opus-mt-ur-en")
+
+# Speech recognition function
+def recognize_speech(audio_file):
+    recognizer = sr.Recognizer()
+    with sr.AudioFile(audio_file) as source:
+        audio_data = recognizer.record(source)
+        try:
+            text = recognizer.recognize_google(audio_data, language="ur-PK")  # Detect Urdu/Pashto
+            return text
+        except sr.UnknownValueError:
+            return "Could not understand audio"
+        except sr.RequestError:
+            return "Could not request results"
+
+# Text-to-speech conversion
+def text_to_speech(text, lang):
+    tts = gTTS(text=text, lang=lang)
+    temp_audio = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3")
+    temp_audio.close()
+    tts.save(temp_audio.name)
+    return temp_audio.name
+
+@app.route("/process", methods=["POST"])
+def process_audio():
+    file = request.files["audio"]
+    filename = "input.wav"
+    file.save(filename)
+
+    text = recognize_speech(filename)
+
+    if "پښتو" in text or "Pashto" in text:
+        response = translator(text, src="ps", tgt="ur")[0]["translation_text"]
+        response_audio = text_to_speech(response, "ur")
+    else:
+        response = translator(text, src="ur", tgt="ps")[0]["translation_text"]
+        response_audio = text_to_speech(response, "ps")
+
+    return jsonify({"response": response, "audio": response_audio})
+
+# Gradio UI
+def chat_interface(audio_input):
+    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as temp_file:
+        temp_file.write(audio_input)
+        temp_file_path = temp_file.name
+    
+    text = recognize_speech(temp_file_path)
+    
+    if "پښتو" in text or "Pashto" in text:
+        response = translator(text, src="ps", tgt="ur")[0]["translation_text"]
+        response_audio = text_to_speech(response, "ur")
+    else:
+        response = translator(text, src="ur", tgt="ps")[0]["translation_text"]
+        response_audio = text_to_speech(response, "ps")
+
+    return response, response_audio
+
+gr.Interface(
+    fn=chat_interface,
+    inputs=gr.Audio(source="microphone", type="filepath"),
+    outputs=[gr.Textbox(label="Translation"), gr.Audio(label="AI Voice Response")],
+    live=True
+).launch()
+
+if __name__ == "__main__":
+    app.run(host="0.0.0.0", port=7860)