app.py

import streamlit as st
import spacy
from spacy import displacy
import pandas as pd
from collections import Counter
import plotly.express as px
from utils import analyze_text
from utils import svg_to_png
import base64

# Set page to wide mode for better visualization
st.set_page_config(layout="wide")

# Load English language model
@st.cache_resource
def load_model():
    return spacy.load('en_core_web_md')

nlp = load_model()

# Streamlit UI
st.markdown("<h1 style='text-align: center; color: white;'>English Sentences Analyzer</h1>", unsafe_allow_html=True)

# Text Input and Help side by side
col1, col2 = st.columns([3, 1])
with col1:
    text_input = st.text_area(
        "Enter English text:",
        "The ambitious startup in Silicon Valley developed an innovative AI system last year. " +
        "Google and Microsoft showed interest in acquiring the technology for $50 million.",
        height=200
    )
    analyze_button = st.button("Analyze Text")

with col2:
    with st.expander("Quick Guide", expanded=True):
        st.markdown("""
        1. Enter your text in the input box
        2. Click "Analyze Text" to see:
            - Sentence structure visualization
            - Detailed token analysis
            - Additional analysis in expandable sections
        3. Use mouse wheel or buttons to zoom the visualization
        4. Click and drag to pan around
        """)

if analyze_button:
    if text_input:
        tokens, entities, noun_chunks, stats, doc = analyze_text(nlp, text_input)
        
        # 1. Dependency Parse with improved visualization
        st.header("Sentence Structure Analysis")
        
        # Generate sentence visualizations
        sentences = list(doc.sents)
        sentence_htmls = []
        for sent in sentences:
            sent_html = displacy.render(sent, style="dep", options={
                "distance": 120,
                "arrow_stroke": 2,
                "arrow_width": 8,
                "font": "Arial",
                "bg": "#ffffff",
            })
            # Ensure proper SVG structure
            if not sent_html.startswith('<?xml'):
                sent_html = '<?xml version="1.0" encoding="UTF-8"?>' + sent_html
            sentence_htmls.append(sent_html)
        
        doc_html = "<br><br>".join(sentence_htmls)
        
        # Convert SVG to PNG with error handling
        png_bytes = svg_to_png(doc_html)
        if png_bytes is None:
            st.error("Failed to generate visualization")
        else:
            png_b64 = base64.b64encode(png_bytes).decode()
        
            # CSS for image container
            st.markdown("""
            <style>
            .image-container {
                position: relative;
                overflow: hidden;
                background: #b4b4b4;
                border: 1px solid #ddd;
                border-radius: 5px;
                margin: 10px 0;
            }
            .zoomable-image {
                transform-origin: 0 0;
                transition: transform 0.1s;
            }
            .download-btn {
                position: absolute;
                right: 10px;
                top: 10px;
                background: rgba(255, 255, 255, 0.8);
                border: 1px solid #ddd;
                border-radius: 4px;
                padding: 5px 10px;
                cursor: pointer;
            }
            .download-btn:hover {
                background: white;
            }
            </style>
            """, unsafe_allow_html=True)
            
            # JavaScript for zoom and pan functionality
            js_code = f"""
            <div class="image-container" id="imageContainer">
                <img src="data:image/png;base64,{png_b64}" 
                     class="zoomable-image" 
                     id="zoomableImage"
                     style="max-width: 100%;">
                <a class="download-btn" 
                   href="data:image/png;base64,{png_b64}" 
                   download="sentence_structure.png">
                   📥 Download
                </a>
            </div>
            <script>
                const container = document.getElementById('imageContainer');
                const img = document.getElementById('zoomableImage');
                let scale = 1;
                let isPanning = false;
                let startX, startY, translateX = 0, translateY = 0;
                
                // Zoom functionality
                container.addEventListener('wheel', (e) => {{
                    e.preventDefault();
                    const rect = container.getBoundingClientRect();
                    const mouseX = e.clientX - rect.left;
                    const mouseY = e.clientY - rect.top;
                    
                    const delta = e.deltaY * -0.01;
                    const newScale = Math.max(1, Math.min(scale + delta, 4));
                    const scaleChange = newScale / scale;
                    
                    translateX = mouseX - (mouseX - translateX) * scaleChange;
                    translateY = mouseY - (mouseY - translateY) * scaleChange;
                    
                    scale = newScale;
                    updateTransform();
                }});
                
                // Pan functionality
                container.addEventListener('mousedown', (e) => {{
                    isPanning = true;
                    startX = e.clientX - translateX;
                    startY = e.clientY - translateY;
                    container.style.cursor = 'grabbing';
                }});
                
                container.addEventListener('mousemove', (e) => {{
                    if (!isPanning) return;
                    translateX = e.clientX - startX;
                    translateY = e.clientY - startY;
                    updateTransform();
                }});
                
                container.addEventListener('mouseup', () => {{
                    isPanning = false;
                    container.style.cursor = 'grab';
                }});
                
                container.addEventListener('mouseleave', () => {{
                    isPanning = false;
                    container.style.cursor = 'grab';
                }});
                
                function updateTransform() {{
                    img.style.transform = `translate(${{translateX}}px, ${{translateY}}px) scale(${{scale}})`;
                }}
                
                // Initialize
                container.style.cursor = 'grab';
                container.style.height = '500px';
            </script>
            """
            
            st.markdown(js_code, unsafe_allow_html=True)
        
            # Add caption
            col1, col2 = st.columns([3, 1])
            with col1:
                st.caption("💡 Tip: Use mouse wheel to zoom, click and drag to pan around")

        # 2. Detailed Token Analysis
        st.header("Token Analysis")
        token_df = pd.DataFrame(tokens)
        
        # Create two columns for token distribution and token details
        col1, col2 = st.columns([1, 2])
        
        with col1:
            # Token distribution visualization
            pos_counts = Counter([token['POS'] for token in tokens])
            fig = px.pie(
                values=list(pos_counts.values()),
                names=list(pos_counts.keys()),
                title="Parts of Speech Distribution"
            )
            fig.update_layout(height=400)
            st.plotly_chart(fig, use_container_width=True)
            
        with col2:
            st.dataframe(token_df, use_container_width=True)
        
        # Additional Analysis in Expanders
        with st.expander("Named Entities"):
            if entities:
                ent_df = pd.DataFrame(entities)
                
                # Visualization of entity distribution
                entity_counts = Counter([ent['Label'] for ent in entities])
                fig = px.bar(
                    x=list(entity_counts.keys()),
                    y=list(entity_counts.values()),
                    title="Distribution of Named Entities",
                    labels={'x': 'Entity Type', 'y': 'Count'}
                )
                st.plotly_chart(fig)
                
                st.table(ent_df)
            else:
                st.info("No named entities found in the text.")
        
        with st.expander("Noun Chunks (Phrases)"):
            if noun_chunks:
                st.table(pd.DataFrame(noun_chunks))
            else:
                st.info("No noun chunks found in the text.")
        
        with st.expander("Text Statistics"):
            col1, col2, col3 = st.columns(3)
            with col1:
                st.metric("Word Count", stats['Word Count'])
            with col2:
                st.metric("Sentence Count", stats['Sentence Count'])
            with col3:
                st.metric("Unique Words", stats['Unique Words'])
            
            st.metric("Average Words per Sentence", stats['Average Words per Sentence'])
            st.metric("Stop Words Percentage", f"{stats['Stop Words %']}%")