Create app.py

app.py

@@ -0,0 +1,86 @@
+import os
+import gradio as gr
+import elevenlabs
+from speech_recognition import Recognizer, AudioFile
+import io
+import google.generativeai as genai
+
+# --- API Keys (Set as environment variables for security!) ---
+genai.configure(api_key=os.environ.get("GOOGLE_API_KEY"))  # Gemini API Key
+elevenlabs_api_key = os.environ.get("ELEVENLABS_API_KEY")
+
+elevenlabs.set_api_key(elevenlabs_api_key)
+
+# --- ElevenLabs Voice ---
+voice = "Bella"  # Choose a voice from ElevenLabs
+
+# --- Language Tutor Parameters ---
+target_language = "Arabic"
+difficulty = 1  # 1 = Easy, 2 = Medium, 3 = Hard
+
+# --- Gemini Model ---
+model = genai.GenerativeModel('gemini-pro')  # Or 'gemini-pro-vision' if you need image input
+
+# --- Functions ---
+
+def generate_question(difficulty):
+    """Generates a Arabic question based on difficulty using Gemini."""
+    prompt = f"Generate a simple {target_language} question for a language learner at difficulty level {difficulty}. Just the question, no extra text."
+    response = model.generate_content(prompt)
+    return response.text.strip()
+
+def evaluate_answer(question, answer):
+    """Evaluates the user's answer using Gemini."""
+    prompt = f"You are a Arabic language tutor. Evaluate the following answer to the question: '{question}'. Answer: '{answer}'. Provide feedback on grammar, vocabulary, and fluency. Keep the feedback concise (under 50 words). Also, give a score from 1-5 (1 being very poor, 5 being excellent)."
+    response = model.generate_content(prompt)
+    return response.text.strip()
+
+def text_to_speech(text, voice):
+    """Converts text to speech using ElevenLabs."""
+    audio = elevenlabs.generate(text=text, voice=voice, model="eleven_monolingual_v1")
+    return audio
+
+def transcribe_audio(audio_file):
+    """Transcribes audio using SpeechRecognition."""
+    r = Recognizer()
+    with AudioFile(audio_file) as source:
+        audio = r.record(source)
+    try:
+        text = r.recognize_google(audio, language=target_language) # You might need to adjust the language code
+        return text
+    except Exception as e:
+        return f"Error transcribing audio: {e}"
+
+def run_tutor(audio_file):
+    """Main function to run the tutor."""
+    question = generate_question(difficulty)
+    question_audio = text_to_speech(question, voice)
+
+    # Transcribe the user's answer
+    user_answer = transcribe_audio(audio_file)
+
+    # Evaluate the answer
+    feedback = evaluate_answer(question, user_answer)
+    feedback_audio = text_to_speech(feedback, voice)
+
+    return question_audio, feedback_audio, question, user_answer, feedback
+
+# --- Gradio Interface ---
+with gr.Blocks() as demo:
+    gr.Markdown("# Adaptive Language Tutor (Spanish)")
+    with gr.Row():
+        question_audio_output = gr.Audio(label="Question")
+        feedback_audio_output = gr.Audio(label="Feedback")
+    with gr.Row():
+        question_text_output = gr.Textbox(label="Question Text")
+        answer_text_output = gr.Textbox(label="Your Answer")
+        feedback_text_output = gr.Textbox(label="Feedback")
+    mic_input = gr.Audio(source="microphone", type="filepath", label="Speak Your Answer")
+
+    mic_input.change(
+        fn=run_tutor,
+        inputs=mic_input,
+        outputs=[question_audio_output, feedback_audio_output, question_text_output, answer_text_output, feedback_text_output]
+    )
+
+demo.launch()