Update app.py

app.py

@@ -1,121 +1,209 @@
 import streamlit as st
-import requests
-import os
-os.environ["SDL_AUDIODRIVER"] = "alsa"
-os.environ["PULSE_SERVER"] = "127.0.0.1"  # Ensures it looks for a PulseAudio server
-os.environ["PA_ALSA_PLUGHW"] = "1"  # Ensures ALSA hardware compatibility
-import sounddevice as sd
-import base64
-import tempfile
-import io
+import torch
 import numpy as np
-import librosa
-from scipy.io.wavfile import write
-from PIL import Image
-
-# Hugging Face API Keys
-HF_API_KEY = os.getenv("HF_API_KEY")
-HEADERS = {"Authorization": f"Bearer {HF_API_KEY}"}
-
-# Hugging Face API Endpoints
-HF_SPEECH_TO_TEXT_API = "https://api-inference.huggingface.co/models/openai/whisper-small"
-HF_CHATBOT_API = "https://api-inference.huggingface.co/models/mistralai/Mistral-7B-Instruct"
-HF_TEXT_TO_SPEECH_API = "https://api-inference.huggingface.co/models/facebook/mms-tts-eng"
-
-# Avatars and their personalities
-avatars = {
-    "rat": {"desc": "Clever and resourceful. Loves puzzles and word games.", "voice": "Hello! Let's solve some fun puzzles together!"},
-    "ox": {"desc": "Steady and dependable. Encourages patience in learning.", "voice": "Consistency is key! Let's learn step by step."},
-    "tiger": {"desc": "Bold and adventurous. Loves dynamic conversations.", "voice": "Let's jump into an exciting conversation!"},
-    "rabbit": {"desc": "Gentle and kind. A patient and encouraging teacher.", "voice": "Take your time. Learning should be fun!"},
-    "dragon": {"desc": "Wise and powerful. Inspires confidence in learning.", "voice": "Believe in yourself! You can do this!"},
-    "snake": {"desc": "Calm and analytical. Gives insightful explanations.", "voice": "Let's explore the beauty of language together."},
-    "horse": {"desc": "Energetic and friendly. Encourages active learning.", "voice": "Let's have fun while we learn!"},
-    "goat": {"desc": "Gentle and artistic. Uses creativity to teach.", "voice": "Imagine and express yourself freely!"},
-    "monkey": {"desc": "Playful and clever. Loves challenges and fun exercises.", "voice": "I love games! Let's play with words!"},
-    "rooster": {"desc": "Confident and expressive. Loves discussions.", "voice": "Speak up! Let's share ideas together."},
-    "dog": {"desc": "Loyal and friendly. Encourages persistence.", "voice": "I'll always be here to help you practice!"},
-    "pig": {"desc": "Warm and cheerful. Creates a positive learning atmosphere.", "voice": "Learning is fun! Let's enjoy this together!"},
+import re
+from transformers import (
+    AutoModelForCausalLM, 
+    AutoTokenizer, 
+    AutoModelForSpeechSeq2Seq,
+    AutoProcessor
+)
+import speech_recognition as sr
+import streamlit.components.v1 as components
+
+# Avatar Personas (same as previous implementation)
+AVATAR_PERSONAS = {
+    "Rat": {
+        "image": "rat.png",
+        "name": "Puzzle Master Rat",
+        "description": """
+        🧩 Clever and curious, I turn language learning into an exciting puzzle! 
+        I love breaking down complex ideas into fun, logical challenges. 
+        My teaching style is all about critical thinking and playful problem-solving.
+        """,
+        "teaching_style": "Analytical and game-based learning",
+        "sample_phrase": "Let's solve this English mystery together!",
+        "voice_sample": "I love solving word puzzles and learning new things!"
+    },
+    "Ox": {
+        "image": "ox.png",
+        "name": "Professor Ox",
+        "description": """
+        📚 Patient and methodical, I guide you through English step by step. 
+        I break down grammar and vocabulary into clear, manageable segments. 
+        Learning is a journey, and I'll be your steady companion.
+        """,
+        "teaching_style": "Structured and systematic approach",
+        "sample_phrase": "Let's learn English systematically!",
+        "voice_sample": "English learning is about understanding each concept carefully."
+    },
+    "Tiger": {
+        "image": "tiger.png",
+        "name": "Adventure Coach Tiger",
+        "description": """
+        🏆 Energetic and brave, I make English learning an exciting adventure! 
+        I turn every lesson into a challenge, motivating you to push your limits 
+        and discover the joy of language learning.
+        """,
+        "teaching_style": "High-energy and challenge-driven",
+        "sample_phrase": "Are you ready to level up your English?",
+        "voice_sample": "Let's make learning English an exciting journey!"
+    }
 }
 
-# Function to process speech-to-text
-def speech_to_text(audio_bytes):
-    files = {"file": ("audio.wav", audio_bytes, "audio/wav")}
-    response = requests.post(HF_SPEECH_TO_TEXT_API, headers=HEADERS, files=files)
-    text = response.json().get("text", "Sorry, could not transcribe audio.")
-    return text
-
-# Function to get chatbot response
-def chatbot_response(user_input):
-    payload = {"inputs": f"You are a friendly AI coach for English learners. Help them practice speaking naturally with supportive feedback.\nUser: {user_input}\nAI:"}
-    response = requests.post(HF_CHATBOT_API, headers=HEADERS, json=payload)
-    ai_reply = response.json().get("generated_text", "I'm here to help! Keep practicing.")
-    return ai_reply
-
-# Function for text-to-speech
-def text_to_speech(text):
-    payload = {"inputs": text}
-    response = requests.post(HF_TEXT_TO_SPEECH_API, headers=HEADERS, json=payload)
-    return response.content
-
-# Function to record audio using sounddevice
-def record_audio(duration=5, samplerate=16000):
-    st.write("🎙️ Recording... Speak now!")
-    audio_data = sd.rec(int(duration * samplerate), samplerate=samplerate, channels=1, dtype="float32")
-    sd.wait()
-    st.write("✅ Recording finished!")
-
-    # Convert to WAV format
-    audio_data = (audio_data * 32767).astype(np.int16)  # Convert to 16-bit PCM
-
-    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmpfile:
-        write(tmpfile.name, samplerate, audio_data)
-        tmpfile.seek(0)
-        return tmpfile.read()
-
-# Streamlit UI
-st.title("🎙️ AI Speaking Pal")
-st.write("Press the button and chat with your pal!")
-
-# Start/Stop Conversation Button
-if "conversation_active" not in st.session_state:
-    st.session_state.conversation_active = False
-
-def toggle_conversation():
-    st.session_state.conversation_active = not st.session_state.conversation_active
-
-st.button("🎤 Start/Stop Conversation", on_click=toggle_conversation)
-
-# Conversation Text Box
-show_text = st.checkbox("Show conversation text")
-if show_text:
-    conversation_box = st.text_area("Conversation:", height=200)
-
-# Handle microphone input
-if st.session_state.conversation_active:
-    audio_bytes = record_audio(duration=5)  # Record for 5 seconds
-    user_text = speech_to_text(audio_bytes)
-    st.write(f"**You:** {user_text}")
-
-    ai_reply = chatbot_response(user_text)
-    st.write(f"**AI:** {ai_reply}")
-
-    speech_audio = text_to_speech(ai_reply)
-    st.audio(io.BytesIO(speech_audio), format="audio/wav")
-
-# Swiper Carousel Component
-avatar_list = list(avatars.keys())
-selected_index = st.slider("Pick your pal:", 0, len(avatar_list) - 1, 0)
-selected_avatar = avatar_list[selected_index]
-avatar_info = avatars[selected_avatar]
-
-# Display Avatar Image
-avatar_img = Image.open(f"images/{selected_avatar}.png")
-st.image(avatar_img, caption=f"{selected_avatar.capitalize()} - {avatar_info['desc']}", use_column_width=True)
-
-# Play avatar's sample voice
-if st.button("Hear Sample"): 
-    audio_bytes = text_to_speech(avatar_info["voice"])
-    st.audio(io.BytesIO(audio_bytes), format="audio/wav")
-
-st.write("Have fun learning English with your AI pal!")
+class AdvancedLanguageLearningAssistant:
+    def __init__(self, avatar_name):
+        # Existing implementation from previous version
+        self.avatar = AVATAR_PERSONAS[avatar_name]
+        
+        # Simplified model loading for Hugging Face Spaces
+        try:
+            self.llm_model = AutoModelForCausalLM.from_pretrained("microsoft/DialoGPT-small")
+            self.llm_tokenizer = AutoTokenizer.from_pretrained("microsoft/DialoGPT-small")
+        except Exception as e:
+            st.error(f"Model loading error: {e}")
+            self.llm_model = None
+            self.llm_tokenizer = None
+
+    def generate_response(self, user_input):
+        if not self.llm_model:
+            return "Sorry, I'm having trouble generating a response."
+
+        # Incorporate avatar's unique personality
+        prompt = f"""
+        Avatar: {self.avatar['name']}
+        Teaching Style: {self.avatar['teaching_style']}
+        User Input: {user_input}
+        
+        Generate a response that:
+        1. Reflects the avatar's unique personality
+        2. Provides encouraging language learning support
+        3. Maintains an engaging tone
+        """
+
+        try:
+            inputs = self.llm_tokenizer.encode(prompt, return_tensors='pt')
+            outputs = self.llm_model.generate(
+                inputs, 
+                max_length=150, 
+                temperature=0.7
+            )
+            
+            response = self.llm_tokenizer.decode(outputs[0], skip_special_tokens=True)
+            return response
+        except Exception as e:
+            return f"Sorry, I'm having trouble generating a response. Error: {e}"
+
+def avatar_selection_page():
+    """
+    First stage: Avatar Selection
+    """
+    st.title("Choose Your Learning Companion")
+    
+    # Custom CSS for avatar selection
+    st.markdown("""
+    <style>
+    .avatar-grid {
+        display: grid;
+        grid-template-columns: repeat(3, 1fr);
+        gap: 20px;
+        padding: 20px;
+    }
+    .avatar-card {
+        border: 2px solid #f0f0f0;
+        border-radius: 10px;
+        padding: 15px;
+        text-align: center;
+        transition: all 0.3s ease;
+    }
+    .avatar-card:hover {
+        transform: scale(1.05);
+        border-color: #4CAF50;
+    }
+    .avatar-image {
+        width: 200px;
+        height: 200px;
+        object-fit: cover;
+        border-radius: 50%;
+    }
+    </style>
+    """, unsafe_allow_html=True)
+
+    # Avatar Selection Grid
+    st.markdown('<div class="avatar-grid">', unsafe_allow_html=True)
+    
+    for avatar_key, avatar_info in AVATAR_PERSONAS.items():
+        st.markdown(f'''
+        <div class="avatar-card">
+            <img src="images/{avatar_info['image']}" class="avatar-image" alt="{avatar_info['name']}">
+            <h3>{avatar_info['name']}</h3>
+            <p>{avatar_info['description']}</p>
+            <button onclick="selectAvatar('{avatar_key}')">Select {avatar_info['name']}</button>
+        </div>
+        ''', unsafe_allow_html=True)
+    
+    st.markdown('</div>', unsafe_allow_html=True)
+
+    # JavaScript to handle avatar selection
+    st.markdown("""
+    <script>
+    function selectAvatar(avatarKey) {
+        window.parent.postMessage({type: 'avatarSelected', avatarKey: avatarKey}, '*');
+    }
+    </script>
+    """, unsafe_allow_html=True)
+
+def conversation_page(selected_avatar):
+    """
+    Second stage: Conversation Interface
+    """
+    st.title(f"Learning English with {AVATAR_PERSONAS[selected_avatar]['name']}")
+    
+    # Initialize Assistant
+    assistant = AdvancedLanguageLearningAssistant(selected_avatar)
+    
+    # Conversation Interface
+    user_input = st.text_input("Type your message in English:")
+    
+    if st.button("Send"):
+        if user_input:
+            response = assistant.generate_response(user_input)
+            st.write(f"🤖 {response}")
+
+def main():
+    # Initialize session state for tracking app stage
+    if 'stage' not in st.session_state:
+        st.session_state.stage = 'avatar_selection'
+    if 'selected_avatar' not in st.session_state:
+        st.session_state.selected_avatar = None
+
+    # Handle avatar selection
+    components.html("""
+    <script>
+    window.addEventListener('message', function(event) {
+        if (event.data.type === 'avatarSelected') {
+            window.parent.postMessage({
+                type: 'streamlit:setComponentValue', 
+                key: 'selected_avatar', 
+                value: event.data.avatarKey
+            }, '*');
+        }
+    });
+    </script>
+    """, height=0)
+
+    # Retrieve selected avatar
+    selected_avatar = st.experimental_get_query_params().get('selected_avatar', [None])[0]
+    
+    if selected_avatar:
+        st.session_state.selected_avatar = selected_avatar
+        st.session_state.stage = 'conversation'
+
+    # Render appropriate page based on stage
+    if st.session_state.stage == 'avatar_selection':
+        avatar_selection_page()
+    elif st.session_state.stage == 'conversation':
+        conversation_page(st.session_state.selected_avatar)
+
+if __name__ == "__main__":
+    main()