app.py

"""
🎌 Anime Translator with Lip-Sync
=================================

A Streamlit application that translates text between English and Hindi,
converts it to speech, and generates a lip-synced anime avatar animation.
"""

import streamlit as st
from pathlib import Path
import tempfile
import time
import shutil
import os
import subprocess
from shutil import which
from typing import Tuple, Optional
import base64

# Import utility modules
from utils.translator import translate_text, detect_language
from utils.tts_engine import synthesize_speech, get_audio_duration
from utils.lipsync import generate_lipsync_gif
from utils.speech_to_text import transcribe_audio, get_language_code
from utils.avatar_manager import list_avatars, get_avatar_preview, ensure_sample_avatar

# =============================================================================
# FFmpeg Configuration
# =============================================================================

def configure_ffmpeg():
    """Configure FFmpeg path for pydub on Windows."""
    possible_paths = [
        r"C:\ffmpeg\bin",
        r"C:\Program Files\ffmpeg\bin",
        r"C:\Program Files (x86)\ffmpeg\bin",
        os.path.expanduser("~\\ffmpeg\\bin"),
        r"C:\Users\Nishant Pratap\ffmpeg\bin",
    ]
    
    if which("ffmpeg") is not None:
        return True
    
    for path in possible_paths:
        ffmpeg_exe = os.path.join(path, "ffmpeg.exe")
        if os.path.exists(ffmpeg_exe):
            os.environ["PATH"] = path + os.pathsep + os.environ.get("PATH", "")
            try:
                from pydub import AudioSegment
                AudioSegment.converter = os.path.join(path, "ffmpeg.exe")
                AudioSegment.ffprobe = os.path.join(path, "ffprobe.exe")
            except:
                pass
            return True
    
    return False


def check_ffmpeg_detailed():
    """Check FFmpeg installation and return detailed status."""
    status = {
        "ffmpeg_in_path": False,
        "ffmpeg_works": False,
        "ffprobe_works": False,
        "pydub_works": False,
        "error_message": None
    }
    
    ffmpeg_path = which("ffmpeg")
    status["ffmpeg_in_path"] = ffmpeg_path is not None
    
    try:
        result = subprocess.run(
            ["ffmpeg", "-version"], 
            capture_output=True, 
            text=True,
            timeout=5
        )
        status["ffmpeg_works"] = result.returncode == 0
    except Exception as e:
        status["error_message"] = str(e)
    
    try:
        result = subprocess.run(
            ["ffprobe", "-version"], 
            capture_output=True, 
            text=True,
            timeout=5
        )
        status["ffprobe_works"] = result.returncode == 0
    except Exception:
        pass
    
    try:
        from pydub import AudioSegment
        silence = AudioSegment.silent(duration=100)
        status["pydub_works"] = True
    except Exception as e:
        status["pydub_works"] = False
        if not status["error_message"]:
            status["error_message"] = str(e)
    
    return status


ffmpeg_found = configure_ffmpeg()

# =============================================================================
# Configuration
# =============================================================================

AVATARS_DIR = Path("./avatars")
TEMP_DIR = Path(tempfile.gettempdir()) / "anime_translator"

AVATARS_DIR.mkdir(parents=True, exist_ok=True)
TEMP_DIR.mkdir(parents=True, exist_ok=True)

# Page configuration
st.set_page_config(
    page_title="🎌 Anime Translator",
    page_icon="🎌",
    layout="wide",
    initial_sidebar_state="expanded"
)

# Initialize session state for animation control
if 'animation_playing' not in st.session_state:
    st.session_state.animation_playing = True
if 'current_gif_path' not in st.session_state:
    st.session_state.current_gif_path = None

# =============================================================================
# Custom CSS Styling - UPDATED WITH ANIMATION FIX
# =============================================================================

st.markdown("""
<style>
    .main {
        background: linear-gradient(135deg, #1a1a2e 0%, #16213e 50%, #0f3460 100%);
    }
    
    .main-header {
        background: linear-gradient(90deg, #e94560, #ff6b6b);
        -webkit-background-clip: text;
        -webkit-text-fill-color: transparent;
        font-size: 3rem;
        font-weight: bold;
        text-align: center;
        padding: 1rem;
        margin-bottom: 2rem;
    }
    
    .stButton > button {
        background: linear-gradient(90deg, #e94560, #ff6b6b);
        color: white;
        border: none;
        border-radius: 25px;
        padding: 0.75rem 2rem;
        font-weight: bold;
        transition: all 0.3s ease;
        width: 100%;
    }
    
    .stButton > button:hover {
        transform: translateY(-2px);
        box-shadow: 0 5px 20px rgba(233, 69, 96, 0.4);
    }
    
    .result-box {
        background: linear-gradient(135deg, rgba(233, 69, 96, 0.1), rgba(255, 107, 107, 0.1));
        border-radius: 15px;
        padding: 1.5rem;
        border: 1px solid rgba(233, 69, 96, 0.3);
        margin: 1rem 0;
    }
    
    .info-box {
        background: rgba(100, 200, 255, 0.1);
        border-left: 4px solid #64c8ff;
        padding: 1rem;
        border-radius: 0 10px 10px 0;
        margin: 1rem 0;
    }
    
    .success-box {
        background: rgba(100, 255, 150, 0.1);
        border-left: 4px solid #64ff96;
        padding: 1rem;
        border-radius: 0 10px 10px 0;
    }
    
    #MainMenu {visibility: hidden;}
    footer {visibility: hidden;}
    
    .stTabs [data-baseweb="tab-list"] {
        gap: 8px;
    }
    
    .stTabs [data-baseweb="tab"] {
        background: rgba(255, 255, 255, 0.05);
        border-radius: 10px;
        padding: 10px 20px;
    }
    
    .stTabs [aria-selected="true"] {
        background: linear-gradient(90deg, #e94560, #ff6b6b);
    }
    
    /* ============================================= */
    /* ANIMATION CONTAINER - FIXED SIZE */
    /* ============================================= */
    
    .animation-container {
        width: 100%;
        max-width: 400px;
        height: 400px;
        margin: 0 auto;
        display: flex;
        align-items: center;
        justify-content: center;
        background: rgba(0, 0, 0, 0.2);
        border-radius: 15px;
        overflow: hidden;
        position: relative;
    }
    
    .animation-container img {
        max-width: 100%;
        max-height: 100%;
        object-fit: contain;
    }
    
    .animation-container.paused img {
        animation-play-state: paused !important;
    }
    
    /* Static image when paused */
    .animation-static {
        width: 100%;
        max-width: 400px;
        height: 400px;
        margin: 0 auto;
        display: flex;
        align-items: center;
        justify-content: center;
        background: rgba(0, 0, 0, 0.2);
        border-radius: 15px;
        overflow: hidden;
    }
    
    .animation-static img {
        max-width: 100%;
        max-height: 100%;
        object-fit: contain;
    }
    
    /* Animation controls */
    .animation-controls {
        display: flex;
        justify-content: center;
        gap: 10px;
        margin-top: 10px;
    }
    
    .control-btn {
        background: linear-gradient(90deg, #e94560, #ff6b6b);
        color: white;
        border: none;
        border-radius: 20px;
        padding: 8px 20px;
        cursor: pointer;
        font-weight: bold;
        transition: all 0.3s ease;
    }
    
    .control-btn:hover {
        transform: scale(1.05);
        box-shadow: 0 3px 15px rgba(233, 69, 96, 0.4);
    }
    
    .control-btn.stop {
        background: linear-gradient(90deg, #666, #888);
    }
    
    /* Fixed height result column */
    .result-column {
        min-height: 600px;
    }
</style>
""", unsafe_allow_html=True)

# =============================================================================
# Helper Functions
# =============================================================================

def cleanup_temp_files(older_than_sec: int = 3600) -> None:
    """Clean up old temporary files."""
    now = time.time()
    try:
        for path in TEMP_DIR.iterdir():
            try:
                if now - path.stat().st_mtime > older_than_sec:
                    if path.is_file():
                        path.unlink()
                    elif path.is_dir():
                        shutil.rmtree(path)
            except Exception:
                pass
    except Exception:
        pass


def get_gif_first_frame(gif_path: str) -> Optional[str]:
    """Extract the first frame of a GIF as a static image."""
    try:
        from PIL import Image
        import io
        
        with Image.open(gif_path) as img:
            # Get first frame
            img.seek(0)
            first_frame = img.copy()
            
            # Save to bytes
            buffer = io.BytesIO()
            first_frame.save(buffer, format='PNG')
            buffer.seek(0)
            
            # Convert to base64
            img_base64 = base64.b64encode(buffer.getvalue()).decode()
            return img_base64
    except Exception as e:
        print(f"Error extracting first frame: {e}")
        return None


def display_animation_with_controls(gif_path: str, key_prefix: str = ""):
    """Display animation with play/pause/stop controls."""
    
    if not gif_path or not os.path.exists(gif_path):
        st.info("ℹ️ No animation available")
        return
    
    # Read GIF file
    with open(gif_path, "rb") as f:
        gif_data = f.read()
    gif_base64 = base64.b64encode(gif_data).decode()
    
    # Get first frame for static display
    first_frame_base64 = get_gif_first_frame(gif_path)
    
    # Animation state key
    state_key = f"{key_prefix}_playing"
    if state_key not in st.session_state:
        st.session_state[state_key] = True
    
    # Control buttons
    col1, col2, col3 = st.columns([1, 1, 1])
    
    with col1:
        if st.button("▶️ Play", key=f"{key_prefix}_play", use_container_width=True):
            st.session_state[state_key] = True
            st.rerun()
    
    with col2:
        if st.button("⏸️ Pause", key=f"{key_prefix}_pause", use_container_width=True):
            st.session_state[state_key] = False
            st.rerun()
    
    with col3:
        if st.button("⏹️ Stop", key=f"{key_prefix}_stop", use_container_width=True):
            st.session_state[state_key] = False
            st.rerun()
    
    # Display animation or static frame
    if st.session_state[state_key]:
        # Playing - show animated GIF
        st.markdown(
            f'''
            <div class="animation-container">
                <img src="data:image/gif;base64,{gif_base64}" alt="Lip-sync animation">
            </div>
            ''',
            unsafe_allow_html=True
        )
    else:
        # Paused/Stopped - show first frame
        if first_frame_base64:
            st.markdown(
                f'''
                <div class="animation-static">
                    <img src="data:image/png;base64,{first_frame_base64}" alt="Animation paused">
                </div>
                <p style="text-align: center; color: #888; margin-top: 10px;">⏸️ Animation Paused</p>
                ''',
                unsafe_allow_html=True
            )
        else:
            st.info("Animation paused")
    
    # Download button
    st.download_button(
        label="📥 Download Animation",
        data=gif_data,
        file_name="lipsync_animation.gif",
        mime="image/gif",
        key=f"{key_prefix}_download",
        use_container_width=True
    )


def process_translation_pipeline(
    text: str,
    source_lang: str,
    target_lang: str,
    avatar_name: str
) -> Tuple[str, Optional[str], Optional[str]]:
    """Main processing pipeline: translate, synthesize speech, generate animation."""
    
    # Step 1: Translate text
    try:
        translated_text = translate_text(text, source_lang, target_lang)
    except Exception as e:
        raise Exception(f"Translation failed: {str(e)}")
    
    # Step 2: Synthesize speech
    try:
        audio_path = synthesize_speech(translated_text, target_lang, TEMP_DIR)
    except Exception as e:
        raise Exception(f"Speech synthesis failed: {str(e)}")
    
    # Step 3: Generate lip-sync animation
    gif_path = None
    try:
        gif_path = generate_lipsync_gif(
            avatar_name=avatar_name,
            audio_path=audio_path,
            avatars_dir=AVATARS_DIR,
            output_dir=TEMP_DIR,
            fps=12
        )
    except Exception as e:
        print(f"Animation generation warning: {str(e)}")
        gif_path = None
    
    return translated_text, audio_path, gif_path


# =============================================================================
# Sidebar
# =============================================================================

with st.sidebar:
    st.markdown("## ⚙️ Settings")
    
    # Avatar selection
    st.markdown("### 🎭 Avatar Selection")
    avatars = list_avatars(AVATARS_DIR)
    
    if avatars:
        selected_avatar = st.selectbox(
            "Choose your avatar",
            options=avatars,
            index=0,
            help="Select an anime avatar for lip-sync animation"
        )
        
        preview = get_avatar_preview(selected_avatar, AVATARS_DIR)
        if preview:
            st.image(preview, caption=f"Preview: {selected_avatar}", use_container_width=True)
    else:
        st.warning("No avatars found. Creating sample avatar...")
        ensure_sample_avatar(AVATARS_DIR)
        selected_avatar = "sample"
        st.rerun()
    
    st.markdown("---")
    
    # Language settings
    st.markdown("### 🌐 Language Settings")
    
    source_language = st.selectbox(
        "Source Language",
        options=["auto", "en", "hi"],
        format_func=lambda x: {"auto": "🔄 Auto-detect", "en": "🇬🇧 English", "hi": "🇮🇳 Hindi"}[x],
        index=0
    )
    
    target_language = st.selectbox(
        "Target Language",
        options=["en", "hi"],
        format_func=lambda x: {"en": "🇬🇧 English", "hi": "🇮🇳 Hindi"}[x],
        index=1
    )
    
    st.markdown("---")
    
    # Animation settings
    st.markdown("### 🎬 Animation Settings")
    
    animation_size = st.slider(
        "Animation Size",
        min_value=200,
        max_value=500,
        value=350,
        step=50,
        help="Adjust the display size of the animation"
    )
    
    auto_play = st.checkbox("Auto-play animation", value=True)
    
    st.markdown("---")
    
    # System status
    st.markdown("### 🔧 System Status")
    
    ffmpeg_status = check_ffmpeg_detailed()
    
    if ffmpeg_status["ffmpeg_works"]:
        st.success("✅ FFmpeg: Working")
    else:
        st.error("❌ FFmpeg: Not working")
        
    if ffmpeg_status["pydub_works"]:
        st.success("✅ Pydub: Working")
    else:
        st.warning("⚠️ Pydub: Limited (fallback mode)")
    
    if ffmpeg_status["error_message"]:
        with st.expander("🔍 Error Details"):
            st.code(ffmpeg_status["error_message"])
            st.markdown("""
            **To fix FFmpeg:**
            ```bash
            conda install -c conda-forge ffmpeg
            ```
            Or download from: https://www.gyan.dev/ffmpeg/builds/
            """)
    
    st.markdown("---")
    
    # Info section
    st.markdown("### ℹ️ About")
    st.markdown("""
    Translate text between English and Hindi with lip-synced avatar animation.
    
    **Features:**
    - 🎤 Voice input
    - 🔄 Auto detection
    - 🗣️ Text-to-speech
    - 🎬 Lip-sync animation
    """)
    
    if st.button("🧹 Clear Temp Files"):
        cleanup_temp_files(older_than_sec=0)
        st.success("Cleared!")


# =============================================================================
# Main Content
# =============================================================================

st.markdown('<h1 class="main-header">🎌 Anime Translator</h1>', unsafe_allow_html=True)
st.markdown(
    '<p style="text-align: center; color: #888; font-size: 1.2rem;">'
    'Translate • Speak • Animate</p>',
    unsafe_allow_html=True
)

# Tabs
tab1, tab2 = st.tabs(["📝 Text Input", "🎤 Voice Input"])

# =============================================================================
# Tab 1: Text Input
# =============================================================================

with tab1:
    col1, col2 = st.columns([1, 1])
    
    with col1:
        st.markdown("### 📝 Enter Your Text")
        
        text_input = st.text_area(
            "Type or paste your text here",
            height=150,
            placeholder="Enter text in English or Hindi...\nउदाहरण: नमस्ते, आप कैसे हैं?\nExample: Hello, how are you?",
            key="text_input"
        )
        
        if text_input:
            detected = detect_language(text_input)
            st.markdown(
                f'<div class="info-box">'
                f'📊 Characters: {len(text_input)} | '
                f'🔍 Detected: {"🇮🇳 Hindi" if detected == "hi" else "🇬🇧 English"}'
                f'</div>',
                unsafe_allow_html=True
            )
        
        translate_btn = st.button(
            "🚀 Translate & Animate",
            key="translate_text_btn",
            use_container_width=True
        )
    
    with col2:
        st.markdown("### 🎬 Result")
        
        # Create a container with fixed height
        result_container = st.container()
        
        with result_container:
            if translate_btn and text_input:
                with st.spinner("🔄 Processing..."):
                    progress = st.progress(0)
                    status_text = st.empty()
                    
                    try:
                        status_text.text("📝 Translating...")
                        progress.progress(33)
                        
                        translated, audio_path, gif_path = process_translation_pipeline(
                            text_input,
                            source_language,
                            target_language,
                            selected_avatar
                        )
                        
                        status_text.text("🗣️ Generating speech...")
                        progress.progress(66)
                        
                        status_text.text("🎬 Creating animation...")
                        progress.progress(100)
                        
                        progress.empty()
                        status_text.empty()
                        
                        # Store results in session state
                        st.session_state['text_result'] = {
                            'translated': translated,
                            'audio_path': audio_path,
                            'gif_path': gif_path
                        }
                        st.session_state['text_animation_playing'] = auto_play
                        
                    except Exception as e:
                        progress.empty()
                        status_text.empty()
                        st.error(f"❌ Error: {str(e)}")
            
            elif translate_btn:
                st.warning("⚠️ Please enter some text to translate.")
            
            # Display stored results
            if 'text_result' in st.session_state:
                result = st.session_state['text_result']
                
                # Display translated text
                st.markdown(
                    f'<div class="result-box">'
                    f'<h4>📜 Translated Text:</h4>'
                    f'<p style="font-size: 1.2rem;">{result["translated"]}</p>'
                    f'</div>',
                    unsafe_allow_html=True
                )
                
                # Audio player
                if result['audio_path'] and os.path.exists(result['audio_path']):
                    st.markdown("#### 🔊 Audio")
                    st.audio(result['audio_path'], format="audio/mp3")
                
                # Animation display with controls
                if result['gif_path'] and os.path.exists(result['gif_path']):
                    st.markdown("#### 🎭 Lip-Sync Animation")
                    display_animation_with_controls(result['gif_path'], key_prefix="text")
                else:
                    st.info("ℹ️ Animation not available (FFmpeg may be missing)")


# =============================================================================
# Tab 2: Voice Input
# =============================================================================

with tab2:
    col1, col2 = st.columns([1, 1])
    
    with col1:
        st.markdown("### 🎤 Voice Recording")
        
        st.markdown("""
        <div class="info-box">
        <strong>Instructions:</strong><br>
        1. Upload an audio file (WAV, MP3, etc.)<br>
        2. Or use the audio recorder below<br>
        3. Click "Transcribe & Translate"
        </div>
        """, unsafe_allow_html=True)
        
        uploaded_audio = st.file_uploader(
            "Upload an audio file",
            type=["wav", "mp3", "ogg", "flac", "m4a"],
            help="Supported formats: WAV, MP3, OGG, FLAC, M4A"
        )
        
        recorded_audio = None
        try:
            from audio_recorder_streamlit import audio_recorder
            st.markdown("**Or record directly:**")
            recorded_audio = audio_recorder(
                text="🎙️ Click to record",
                recording_color="#e94560",
                neutral_color="#6c757d",
                icon_name="microphone",
                icon_size="2x"
            )
        except ImportError:
            st.info("💡 For recording: `pip install audio-recorder-streamlit`")
        
        voice_lang = st.selectbox(
            "Recording Language",
            options=["en", "hi"],
            format_func=lambda x: {"en": "🇬🇧 English", "hi": "🇮🇳 Hindi"}[x]
        )
        
        voice_btn = st.button(
            "🎯 Transcribe & Translate",
            key="voice_btn",
            use_container_width=True
        )
    
    with col2:
        st.markdown("### 🎬 Result")
        
        audio_to_process = None
        
        if uploaded_audio is not None:
            temp_audio_path = TEMP_DIR / f"uploaded_{int(time.time()*1000)}.wav"
            with open(temp_audio_path, "wb") as f:
                f.write(uploaded_audio.getbuffer())
            audio_to_process = str(temp_audio_path)
            st.audio(uploaded_audio)
        
        elif recorded_audio is not None:
            temp_audio_path = TEMP_DIR / f"recorded_{int(time.time()*1000)}.wav"
            with open(temp_audio_path, "wb") as f:
                f.write(recorded_audio)
            audio_to_process = str(temp_audio_path)
            st.audio(recorded_audio, format="audio/wav")
        
        if voice_btn:
            if audio_to_process:
                with st.spinner("🔄 Processing voice..."):
                    try:
                        st.text("🎤 Transcribing...")
                        lang_code = get_language_code(voice_lang)
                        transcribed_text, success = transcribe_audio(audio_to_process, lang_code)
                        
                        if success:
                            translated, audio_path, gif_path = process_translation_pipeline(
                                transcribed_text,
                                voice_lang,
                                target_language,
                                selected_avatar
                            )
                            
                            # Store results in session state
                            st.session_state['voice_result'] = {
                                'transcribed': transcribed_text,
                                'translated': translated,
                                'audio_path': audio_path,
                                'gif_path': gif_path
                            }
                            st.session_state['voice_animation_playing'] = auto_play
                            
                        else:
                            st.error(f"❌ {transcribed_text}")
                    except Exception as e:
                        st.error(f"❌ Error: {str(e)}")
            else:
                st.warning("⚠️ Please upload or record audio first.")
        
        # Display stored results
        if 'voice_result' in st.session_state:
            result = st.session_state['voice_result']
            
            st.markdown(
                f'<div class="success-box">'
                f'<strong>📝 Transcribed:</strong> {result["transcribed"]}'
                f'</div>',
                unsafe_allow_html=True
            )
            
            st.markdown(
                f'<div class="result-box">'
                f'<h4>📜 Translated:</h4>'
                f'<p style="font-size: 1.2rem;">{result["translated"]}</p>'
                f'</div>',
                unsafe_allow_html=True
            )
            
            if result['audio_path'] and os.path.exists(result['audio_path']):
                st.markdown("#### 🔊 Audio")
                st.audio(result['audio_path'], format="audio/mp3")
            
            if result['gif_path'] and os.path.exists(result['gif_path']):
                st.markdown("#### 🎭 Animation")
                display_animation_with_controls(result['gif_path'], key_prefix="voice")


# =============================================================================
# Footer
# =============================================================================

st.markdown("---")
st.markdown(
    """
    <div style="text-align: center; color: #666; padding: 1rem;">
        <p>Made By Praveen</p>
    </div>
    """,
    unsafe_allow_html=True
)