app.py

#!/usr/bin/env python3
"""
BabelSlide v2.0 - Professional Document Translator
Streamlit application for translating PDF, DOCX, and PPTX documents using AI
"""

import streamlit as st
import tempfile
from pathlib import Path
import sys
import os

from translators.chatgpt_translator import ChatGPTTranslator
from translators.deepseek_translator import DeepSeekTranslator
from processors.pdf_processor import PDFProcessor
from processors.docx_processor import DOCXProcessor
from processors.pptx_processor import PPTXProcessor
from utils.constants import LANGUAGES, API_PROVIDERS
from utils.validator import FileValidator
from utils.logger import setup_logger, ProcessLogger
from core.exceptions import (
    BabelSlideException, 
    ValidationError, 
    UnsupportedFileError, 
    FileSizeError,
    APIKeyError,
    TranslationError,
    ProcessorError
)

class BabelSlideStreamlitApp:
    """Streamlit interface for BabelSlide application"""
    
    def __init__(self):
        self.logger = setup_logger("BabelSlideUI")
        self.process_logger = ProcessLogger(self.logger)
        
        # Initialize session state
        if 'processing' not in st.session_state:
            st.session_state.processing = False
        if 'translation_result' not in st.session_state:
            st.session_state.translation_result = None
        if 'review_result' not in st.session_state:
            st.session_state.review_result = None
    
    def setup_page_config(self):
        """Configure Streamlit page"""
        st.set_page_config(
            page_title="BabelSlide - Document Translator",
            page_icon="🌍",
            layout="wide",
            initial_sidebar_state="expanded"
        )
        
        # Custom CSS
        st.markdown("""
        <style>
        .main-header {
            text-align: center;
            padding: 2rem 0;
            background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
            color: white;
            border-radius: 12px;
            margin-bottom: 2rem;
        }
        
        .success-box {
            background: #d1fae5;
            border: 1px solid #10b981;
            border-radius: 8px;
            padding: 1rem;
            margin: 1rem 0;
        }
        
        .error-box {
            background: #fef2f2;
            border: 1px solid #ef4444;
            border-radius: 8px;
            padding: 1rem;
            margin: 1rem 0;
        }
        
        .info-box {
            background: #eff6ff;
            border: 1px solid #3b82f6;
            border-radius: 8px;
            padding: 1rem;
            margin: 1rem 0;
        }
        
        .stButton > button {
            background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
            color: white;
            border: none;
            border-radius: 8px;
            padding: 0.5rem 2rem;
            font-weight: 600;
        }
        </style>
        """, unsafe_allow_html=True)
    
    def render_header(self):
        """Render application header"""
        st.markdown("""
        <div class="main-header">
            <h1>🌍 BabelSlide v2.0</h1>
            <p>Professional Document Translation using AI • PDF • DOCX • PPTX</p>
        </div>
        """, unsafe_allow_html=True)
    
    def render_sidebar(self):
        """Render configuration sidebar"""
        st.sidebar.markdown("## ⚙️ Configuration")
        
        # API Provider
        api_provider = st.sidebar.selectbox(
            "AI Provider",
            options=list(API_PROVIDERS.keys()),
            index=0,
            help="Choose your preferred translation AI"
        )
        
        # API Key
        api_key = st.sidebar.text_input(
            "API Key",
            type="password",
            placeholder="Enter your API key (sk-... for OpenAI)",
            help="Your API key is never stored permanently"
        )
        
        st.sidebar.markdown("---")
        
        # Languages
        col1, col2 = st.sidebar.columns(2)
        
        with col1:
            source_lang = st.selectbox(
                "Source Language",
                options=list(LANGUAGES.keys()),
                index=list(LANGUAGES.keys()).index("English"),
                help="Language of the original document"
            )
        
        with col2:
            target_lang = st.selectbox(
                "Target Language",
                options=list(LANGUAGES.keys()),
                index=list(LANGUAGES.keys()).index("Polish"),
                help="Language to translate to"
            )
        
        st.sidebar.markdown("---")
        st.sidebar.markdown("### 📝 Supported Formats")
        st.sidebar.info("• PDF documents\n• DOCX (Word) files\n• PPTX (PowerPoint) presentations")
        st.sidebar.warning("Maximum file size: 50 MB")
        
        return api_provider, api_key, source_lang, target_lang
    
    def render_file_upload(self):
        """Render file upload section"""
        st.markdown("## 📄 Document Upload")
        
        uploaded_file = st.file_uploader(
            "Choose a document to translate",
            type=['pdf', 'docx', 'pptx'],
            help="Upload PDF, DOCX, or PPTX files (max 50 MB)",
            accept_multiple_files=False
        )
        
        if uploaded_file:
            col1, col2, col3 = st.columns([2, 1, 1])
            with col1:
                st.info(f"📁 **File:** {uploaded_file.name}")
            with col2:
                file_size = len(uploaded_file.getvalue()) / (1024 * 1024)
                st.info(f"📏 **Size:** {file_size:.1f} MB")
            with col3:
                file_type = uploaded_file.name.split('.')[-1].upper()
                st.info(f"📋 **Type:** {file_type}")
        
        return uploaded_file
    
    def validate_inputs(self, file, api_provider, api_key, source_lang, target_lang):
        """Validate all inputs before processing"""
        errors = []
        
        if not file:
            errors.append("Please upload a document")
        
        if not api_key or not api_key.strip():
            errors.append("Please provide an API key")
        
        if source_lang == target_lang:
            errors.append("Source and target languages must be different")
        
        # Validate file if provided
        if file:
            try:
                # Create temporary file for validation
                with tempfile.NamedTemporaryFile(delete=False, suffix=f".{file.name.split('.')[-1]}") as tmp_file:
                    tmp_file.write(file.getvalue())
                    tmp_file_path = Path(tmp_file.name)
                
                FileValidator.validate_file(tmp_file_path)
                tmp_file_path.unlink()  # Clean up
                
            except (ValidationError, UnsupportedFileError, FileSizeError) as e:
                errors.append(f"File validation error: {str(e)}")
        
        # Validate API key format
        try:
            if api_key:
                FileValidator.validate_api_key(api_key.strip(), api_provider)
        except ValidationError as e:
            errors.append(f"API key error: {str(e)}")
        
        return errors
    
    def process_document(self, file, api_provider, api_key, source_lang, target_lang):
        """Process document translation"""
        try:
            # Create temporary file
            with tempfile.NamedTemporaryFile(delete=False, suffix=f".{file.name.split('.')[-1]}") as tmp_file:
                tmp_file.write(file.getvalue())
                tmp_file_path = Path(tmp_file.name)
            
            # Create translator
            if api_provider == "ChatGPT":
                translator = ChatGPTTranslator(api_key.strip())
            elif api_provider == "DeepSeek":
                translator = DeepSeekTranslator(api_key.strip())
            else:
                raise ValueError(f"Unsupported provider: {api_provider}")
            
            # Create processor based on file extension
            extension = tmp_file_path.suffix.lower()
            if extension == '.pdf':
                processor = PDFProcessor(translator)
            elif extension == '.docx':
                processor = DOCXProcessor(translator)
            elif extension == '.pptx':
                processor = PPTXProcessor(translator)
            else:
                raise ValueError(f"Unsupported file format: {extension}")
            
            # Progress tracking
            progress_bar = st.progress(0)
            status_text = st.empty()
            
            def progress_callback(progress_val, desc):
                progress_bar.progress(progress_val)
                status_text.text(desc)
            
            # Process document
            status_text.text("Starting translation...")
            output_path, summary_text = processor.process_document(
                tmp_file_path,
                source_lang,
                target_lang,
                progress_callback
            )
            
            # Generate review
            status_text.text("Generating review...")
            review_text = self.generate_review(summary_text, source_lang, translator)
            
            # Clean up temp file
            tmp_file_path.unlink()
            
            progress_bar.progress(1.0)
            status_text.text("✅ Translation completed!")
            
            return output_path, review_text, summary_text
            
        except Exception as e:
            self.logger.error(f"Translation error: {str(e)}")
            raise
    
    def generate_review(self, translated_text: str, source_lang: str, translator) -> str:
        """Generate translation review"""
        try:
            system_prompt = f"""Generate a comprehensive translation review in {source_lang} covering:
            1. Translation quality assessment
            2. Coherence and consistency
            3. Technical terminology accuracy  
            4. Overall readability
            5. Recommendations for improvement
            
            Keep the review concise but informative."""
            
            # Use translator's API to generate review
            review = translator._make_translation_request(
                f"Review this translated text:\n\n{translated_text[:2000]}...",
                "English",
                source_lang
            )
            
            return translator._clean_translation_output(review)
            
        except Exception as e:
            return f"Review generation failed: {str(e)}"
    
    def render_results(self):
        """Render translation results"""
        if st.session_state.translation_result:
            st.markdown("## 📥 Results")
            
            col1, col2 = st.columns(2)
            
            with col1:
                st.markdown("### 📄 Translated Document")
                if st.session_state.translation_result:
                    with open(st.session_state.translation_result, 'rb') as file:
                        st.download_button(
                            label="⬇️ Download Translated Document",
                            data=file.read(),
                            file_name=Path(st.session_state.translation_result).name,
                            mime="application/octet-stream"
                        )
            
            with col2:
                st.markdown("### 📋 Translation Review")
                if st.session_state.review_result:
                    st.download_button(
                        label="⬇️ Download Review",
                        data=st.session_state.review_result,
                        file_name="translation_review.txt",
                        mime="text/plain"
                    )
            
            # Summary
            if hasattr(st.session_state, 'summary_text') and st.session_state.summary_text:
                st.markdown("### 📝 Translation Summary")
                with st.expander("View Summary", expanded=False):
                    st.text_area(
                        "Summary",
                        value=st.session_state.summary_text[:1000] + "..." if len(st.session_state.summary_text) > 1000 else st.session_state.summary_text,
                        height=200,
                        disabled=True,
                        label_visibility="collapsed"
                    )
    
    def run(self):
        """Main application loop"""
        self.setup_page_config()
        self.render_header()
        
        # Sidebar configuration
        api_provider, api_key, source_lang, target_lang = self.render_sidebar()
        
        # Main content
        uploaded_file = self.render_file_upload()
        
        # Translation button
        st.markdown("---")
        col1, col2, col3 = st.columns([1, 2, 1])
        with col2:
            translate_button = st.button(
                "🚀 Translate Document",
                disabled=st.session_state.processing,
                use_container_width=True
            )
        
        # Process translation
        if translate_button:
            # Validate inputs
            errors = self.validate_inputs(uploaded_file, api_provider, api_key, source_lang, target_lang)
            
            if errors:
                st.error("❌ **Please fix the following errors:**")
                for error in errors:
                    st.error(f"• {error}")
            else:
                st.session_state.processing = True
                
                try:
                    with st.spinner("Translating document..."):
                        output_path, review_text, summary_text = self.process_document(
                            uploaded_file, api_provider, api_key, source_lang, target_lang
                        )
                    
                    # Store results
                    st.session_state.translation_result = output_path
                    st.session_state.review_result = review_text
                    st.session_state.summary_text = summary_text
                    
                    st.success(f"✅ **Translation completed successfully!**\n\n"
                              f"📄 **File:** {uploaded_file.name}\n"
                              f"🔄 **Translation:** {source_lang} → {target_lang}\n"
                              f"🤖 **Provider:** {api_provider}")
                    
                    # Auto-refresh to show results
                    st.rerun()
                    
                except Exception as e:
                    st.error(f"❌ **Translation failed:** {str(e)}")
                
                finally:
                    st.session_state.processing = False
        
        # Show results if available
        self.render_results()
        
        # Footer
        st.markdown("---")
        st.markdown(
            "<div style='text-align: center; color: #666;'>"
            "<strong>BabelSlide v2.0</strong> • Professional document translation • Built for global communication"
            "</div>",
            unsafe_allow_html=True
        )

# Main entry point
if __name__ == "__main__":
    app = BabelSlideStreamlitApp()
    app.run()