Update app.py

app.py

@@ -1,583 +1,266 @@
+# app.py
+"""
+Streamlit UI for the News Sentiment Analyzer.
+- Calls the in-process FastAPI orchestrator (NewsAnalyzer) directly for zero-latency on Spaces.
+- Lightweight, CPU-safe widgets with progress, charts, tables, and exports (CSV/JSON/PDF + Audio).
+"""
+
+from __future__ import annotations
+
+import io
+import json
+import logging
+from datetime import datetime
+from typing import Any, Dict, List
+
 import streamlit as st
 import pandas as pd
 import plotly.express as px
-import plotly.graph_objects as go
-from wordcloud import WordCloud
-import matplotlib.pyplot as plt
-import asyncio
-import json
-import base64
-from datetime import datetime
-import io
-import os
 
-# Import our modules
-from api import NewsAnalyzer
-from utils import load_config, cache_results
-from report import generate_pdf_report
+# Local modules
+from api import analyzer  # global NewsAnalyzer instance
+from utils import (
+    setup_logging,
+    load_config,
+    calculate_sentiment_distribution,
+    format_number,
+)
+from report import generate_pdf_report  # your existing PDF generator
+
+# ------------------------------------------------------------------------------
+# App setup
+# ------------------------------------------------------------------------------
+
+setup_logging()
+logger = logging.getLogger("app")
 
-# Configure page
 st.set_page_config(
-    page_title="Global Business News Intelligence Dashboard",
-    page_icon="📊",
+    page_title="News Sentiment Analyzer",
+    page_icon="📰",
     layout="wide",
-    initial_sidebar_state="expanded"
 )
 
-# Custom CSS
-st.markdown("""
-<style>
-    .main-header {
-        font-size: 2.5rem;
-        font-weight: bold;
-        text-align: center;
-        color: #2E86AB;
-        margin-bottom: 2rem;
-    }
-    .metric-card {
-        background-color: #f0f2f6;
-        padding: 1rem;
-        border-radius: 10px;
-        border-left: 4px solid #2E86AB;
-    }
-    .sentiment-positive { color: #28a745; font-weight: bold; }
-    .sentiment-negative { color: #dc3545; font-weight: bold; }
-    .sentiment-neutral { color: #6c757d; font-weight: bold; }
-    .audio-container {
-        background-color: #f8f9fa;
-        padding: 10px;
-        border-radius: 5px;
-        margin: 10px 0;
-    }
-</style>
-""", unsafe_allow_html=True)
-
-# Initialize session state
-if 'analyzer' not in st.session_state:
-    st.session_state.analyzer = NewsAnalyzer()
-if 'results' not in st.session_state:
-    st.session_state.results = None
-if 'analysis_complete' not in st.session_state:
-    st.session_state.analysis_complete = False
-
-# Ensure extra session keys exist
-if 'query' not in st.session_state:
-    st.session_state.query = ''
-if 'progress' not in st.session_state:
-    st.session_state.progress = 0
-if 'progress_bar' not in st.session_state:
-    st.session_state.progress_bar = None
-if 'status_text' not in st.session_state:
-    st.session_state.status_text = None
-
-def main():
-    # Header
-    st.markdown('<h1 class="main-header">🌐 Global Business News Intelligence Dashboard</h1>', unsafe_allow_html=True)
-    st.markdown("**Real-time sentiment analysis, multilingual summaries, and audio insights for business intelligence**")
-    
-    # Sidebar
-    with st.sidebar:
-        st.header("⚙️ Configuration")
-        
-        # Input section
-        st.subheader("🎯 Target Analysis")
-        query_type = st.selectbox("Query Type", ["Company", "Stock Ticker", "Keyword", "Industry"])
-        query = st.text_input(f"Enter {query_type}:", placeholder="e.g., Tesla, TSLA, AI technology")
-        
-        st.subheader("📊 Analysis Settings")
-        num_articles = st.slider("Number of Articles", 5, 50, 20)
-        languages = st.multiselect(
-            "Summary Languages", 
-            ["English", "Hindi", "Tamil"], 
-            default=["English"]
-        )
-        include_audio = st.checkbox("Generate Audio Summaries", True)
-        
-        st.subheader("🔧 Model Settings")
-        sentiment_models = st.multiselect(
-            "Sentiment Models",
-            ["VADER", "Loughran-McDonald", "FinBERT"],
-            default=["VADER", "Loughran-McDonald", "FinBERT"]
+# Minimal CSS polish
+st.markdown(
+    """
+    <style>
+    .small { font-size: 0.85rem; color: #666; }
+    .ok { color: #1b8a5a; }
+    .bad { color: #b00020; }
+    .neutral { color: #666; }
+    .stProgress > div > div > div { background-color: #4b8bf4; }
+    .block-container { padding-top: 2rem; }
+    </style>
+    """,
+    unsafe_allow_html=True,
+)
+
+# ------------------------------------------------------------------------------
+# Sidebar controls
+# ------------------------------------------------------------------------------
+
+cfg = load_config()
+
+st.sidebar.header("Settings")
+default_query = st.sidebar.text_input("Company / Keyword", value="Tesla")
+num_articles = st.sidebar.slider("Number of articles", 5, 50, 20, step=1)
+languages = st.sidebar.multiselect(
+    "Summaries in languages",
+    options=["English", "Hindi", "Tamil"],
+    default=["English"],
+)
+include_audio = st.sidebar.checkbox("Generate audio summary", value=True)
+sentiment_models = st.sidebar.multiselect(
+    "Sentiment models",
+    options=["VADER", "Loughran-McDonald", "FinBERT"],
+    default=["VADER", "Loughran-McDonald", "FinBERT"],
+)
+st.sidebar.caption("Tip: disable FinBERT if your Space has < 2GB RAM.")
+
+run_btn = st.sidebar.button("Analyze", use_container_width=True, type="primary")
+
+# ------------------------------------------------------------------------------
+# Header
+# ------------------------------------------------------------------------------
+
+st.title("📰 News Sentiment Analyzer")
+st.caption("Scrape → Summarize → Sentiment → Keywords → Multilingual → Audio — deployed on Hugging Face Spaces")
+
+# ------------------------------------------------------------------------------
+# Helper functions
+# ------------------------------------------------------------------------------
+
+def _articles_to_df(articles: List[Dict[str, Any]]) -> pd.DataFrame:
+    rows = []
+    for a in articles:
+        rows.append(
+            {
+                "title": a.get("title", ""),
+                "source": a.get("source", ""),
+                "date": a.get("date"),
+                "url": a.get("url", ""),
+                "summary": a.get("summary", ""),
+                "sentiment_compound": a.get("sentiment", {}).get("compound", 0.0),
+            }
         )
-        
-        # Analysis button
-        analyze_button = st.button("🚀 Analyze News", type="primary", use_container_width=True)
-    
-    # Main content area
-    if analyze_button and query:
-        st.session_state.analysis_complete = False
-        with st.spinner("🔍 Analyzing news articles... This may take a few minutes."):
-            try:
-                # Create progress UI
-                progress_bar = st.progress(0)
-                status_text = st.empty()
-                # Store UI handles in session state for callbacks
-                st.session_state.progress_bar = progress_bar
-                st.session_state.status_text = status_text
-                
-                # Run analysis
-                config = {
-                    'query': query,
-                    'num_articles': num_articles,
-                    'languages': languages,
-                    'include_audio': include_audio,
-                    'sentiment_models': sentiment_models
-                }
-                
-                # Update progress
-                status_text.text("🔍 Scraping articles...")
-                progress_bar.progress(20)
-                
-                results = st.session_state.analyzer.analyze_news(config, progress_callback=update_progress)
-                st.session_state.results = results
-                st.session_state.analysis_complete = True
-                
-                progress_bar.progress(100)
-                status_text.text("✅ Analysis complete!")
-                
-            except Exception as e:
-                st.error(f"Error during analysis: {str(e)}")
-                st.session_state.analysis_complete = False
-    
-    # Display results
-    if st.session_state.analysis_complete and st.session_state.results:
-        display_results(st.session_state.results)
-    
-    elif not st.session_state.analysis_complete and query:
-        st.info("👆 Click 'Analyze News' to start the analysis")
-    
-    else:
-        show_demo_dashboard()
-
-def update_progress(progress, status):
-    """Callback function for progress updates"""
-    try:
-        st.session_state.progress = progress
-        if st.session_state.progress_bar is not None:
-            st.session_state.progress_bar.progress(int(max(0, min(100, progress))))
-        if st.session_state.status_text is not None:
-            st.session_state.status_text.text(status)
-    except Exception:
-        pass
-
-
-
-def display_results(results):
-    """Display analysis results with interactive dashboard"""
-    st.header(f"📈 Analysis Results for: {results['query']}")
-    
-    # Key metrics
-    col1, col2, col3, col4 = st.columns(4)
-    
-    with col1:
-        st.markdown('<div class="metric-card">', unsafe_allow_html=True)
-        st.metric("Articles Analyzed", len(results['articles']))
-        st.markdown('</div>', unsafe_allow_html=True)
-    
-    with col2:
-        avg_sentiment = results['summary']['average_sentiment']
-        sentiment_color = "sentiment-positive" if avg_sentiment > 0.1 else "sentiment-negative" if avg_sentiment < -0.1 else "sentiment-neutral"
-        st.markdown('<div class="metric-card">', unsafe_allow_html=True)
-        st.metric("Average Sentiment", f"{avg_sentiment:.3f}")
-        st.markdown('</div>', unsafe_allow_html=True)
-    
-    with col3:
-        st.markdown('<div class="metric-card">', unsafe_allow_html=True)
-        st.metric("Sources", len(set([article['source'] for article in results['articles']])))
-        st.markdown('</div>', unsafe_allow_html=True)
-    
-    with col4:
-        st.markdown('<div class="metric-card">', unsafe_allow_html=True)
-        st.metric("Languages", len(results.get('languages', ['English'])))
-        st.markdown('</div>', unsafe_allow_html=True)
-    
-    # Tabs for different views
-    tab1, tab2, tab3, tab4, tab5, tab6 = st.tabs(["📊 Dashboard", "📰 Articles", "🎯 Sentiment", "🗣️ Audio", "📤 Export", "🔌 API"])
-    
-    with tab1:
-        display_dashboard(results)
-    
-    with tab2:
-        display_articles(results)
-    
-    with tab3:
-        display_sentiment_analysis(results)
-    
-    with tab4:
-        display_audio_summaries(results)
-    
-    with tab5:
-        display_export_options(results)
-    
-    with tab6:
-        display_api_info(results)
-
-def display_dashboard(results):
-    """Display main dashboard with charts"""
-    col1, col2 = st.columns(2)
-    
-    with col1:
-        # Sentiment distribution
-        st.subheader("📊 Sentiment Distribution")
-        sentiment_counts = {
-            'Positive': sum(1 for article in results['articles'] if article['sentiment']['compound'] > 0.1),
-            'Negative': sum(1 for article in results['articles'] if article['sentiment']['compound'] < -0.1),
-            'Neutral': sum(1 for article in results['articles'] if -0.1 <= article['sentiment']['compound'] <= 0.1)
+    df = pd.DataFrame(rows)
+    if "date" in df.columns:
+        try:
+            df["date"] = pd.to_datetime(df["date"])
+        except Exception:
+            pass
+    return df
+
+
+def _render_distribution(dist: Dict[str, Any]):
+    cols = st.columns(4)
+    cols[0].metric("Total", dist.get("total", 0))
+    cols[1].metric("Positive", dist.get("positive", 0))
+    cols[2].metric("Negative", dist.get("negative", 0))
+    cols[3].metric("Neutral", dist.get("neutral", 0))
+
+    chart_df = pd.DataFrame(
+        {
+            "Sentiment": ["Positive", "Negative", "Neutral"],
+            "Count": [
+                dist.get("positive", 0),
+                dist.get("negative", 0),
+                dist.get("neutral", 0),
+            ],
         }
-        
-        fig_pie = px.pie(
-            values=list(sentiment_counts.values()),
-            names=list(sentiment_counts.keys()),
-            color_discrete_map={'Positive': '#28a745', 'Negative': '#dc3545', 'Neutral': '#6c757d'}
+    )
+    fig = px.bar(chart_df, x="Sentiment", y="Count", title="Sentiment distribution")
+    st.plotly_chart(fig, use_container_width=True)
+
+
+def _download_buttons(results: Dict[str, Any], df: pd.DataFrame):
+    c1, c2, c3 = st.columns(3)
+
+    # JSON
+    with c1:
+        json_bytes = json.dumps(results, default=str, indent=2).encode("utf-8")
+        st.download_button(
+            "Download JSON",
+            data=json_bytes,
+            file_name=f"news_analysis_{results['query']}_{datetime.now().strftime('%Y%m%d_%H%M%S')}.json",
+            mime="application/json",
+            use_container_width=True,
         )
-        st.plotly_chart(fig_pie, use_container_width=True)
-    
-    with col2:
-        # Source distribution
-        st.subheader("📰 Source Distribution")
-        source_counts = {}
-        for article in results['articles']:
-            source = article['source']
-            source_counts[source] = source_counts.get(source, 0) + 1
-        
-        fig_bar = px.bar(
-            x=list(source_counts.keys()),
-            y=list(source_counts.values()),
-            color=list(source_counts.values()),
-            color_continuous_scale="viridis"
+
+    # CSV
+    with c2:
+        csv_bytes = df.to_csv(index=False).encode("utf-8")
+        st.download_button(
+            "Download CSV",
+            data=csv_bytes,
+            file_name=f"news_analysis_{results['query']}_{datetime.now().strftime('%Y%m%d_%H%M%S')}.csv",
+            mime="text/csv",
+            use_container_width=True,
         )
-        fig_bar.update_layout(xaxis_title="Source", yaxis_title="Article Count")
-        st.plotly_chart(fig_bar, use_container_width=True)
-    
-    # Timeline chart
-    st.subheader("📈 Sentiment Over Time")
-    if results['articles']:
-        df_timeline = pd.DataFrame([
-            {
-                'date': article.get('date', datetime.now()),
-                'sentiment': article['sentiment']['compound'],
-                'title': article['title'][:50] + "..." if len(article['title']) > 50 else article['title']
-            }
-            for article in results['articles']
-            if 'date' in article
-        ])
-        
-        if not df_timeline.empty:
-            fig_timeline = px.scatter(
-                df_timeline,
-                x='date',
-                y='sentiment',
-                hover_data=['title'],
-                color='sentiment',
-                color_continuous_scale=['red', 'gray', 'green'],
-                color_continuous_midpoint=0
-            )
-            fig_timeline.update_layout(
-                xaxis_title="Date",
-                yaxis_title="Sentiment Score",
-                yaxis=dict(range=[-1, 1])
-            )
-            st.plotly_chart(fig_timeline, use_container_width=True)
-    
-    # Keywords word cloud
-    st.subheader("🔤 Key Topics")
-    if 'keywords' in results and results['keywords']:
-        col1, col2 = st.columns([2, 1])
-        
-        with col1:
-            # Create word cloud
-            keywords_text = ' '.join([kw['keyword'] for kw in results['keywords'][:50]])
-            if keywords_text:
-                wordcloud = WordCloud(
-                    width=800, 
-                    height=400, 
-                    background_color='white',
-                    colormap='viridis'
-                ).generate(keywords_text)
-                
-                fig, ax = plt.subplots(figsize=(10, 5))
-                ax.imshow(wordcloud, interpolation='bilinear')
-                ax.axis('off')
-                st.pyplot(fig)
-        
-        with col2:
-            st.write("**Top Keywords:**")
-            for i, kw in enumerate(results['keywords'][:10]):
-                st.write(f"{i+1}. {kw['keyword']} ({kw['score']:.3f})")
-
-def display_articles(results):
-    """Display individual articles with summaries"""
-    st.subheader(f"📰 Articles ({len(results['articles'])})")
-    
-    for i, article in enumerate(results['articles']):
-        with st.expander(f"📄 {article['title']}", expanded=(i < 3)):
-            col1, col2 = st.columns([3, 1])
-            
-            with col1:
-                st.write(f"**Source:** {article['source']}")
-                if 'date' in article:
-                    st.write(f"**Date:** {article['date']}")
-                st.write(f"**URL:** {article.get('url', 'N/A')}")
-                
-                # Sentiment
-                sentiment = article['sentiment']
-                sentiment_label = "Positive" if sentiment['compound'] > 0.1 else "Negative" if sentiment['compound'] < -0.1 else "Neutral"
-                sentiment_color = "sentiment-positive" if sentiment_label == "Positive" else "sentiment-negative" if sentiment_label == "Negative" else "sentiment-neutral"
-                st.markdown(f"**Sentiment:** <span class='{sentiment_color}'>{sentiment_label} ({sentiment['compound']:.3f})</span>", unsafe_allow_html=True)
-            
-            with col2:
-                # Model-specific scores
-                st.write("**Model Scores:**")
-                if 'vader' in sentiment:
-                    st.write(f"VADER: {sentiment['vader']:.3f}")
-                if 'loughran_mcdonald' in sentiment:
-                    st.write(f"L&M: {sentiment['loughran_mcdonald']:.3f}")
-                if 'finbert' in sentiment:
-                    st.write(f"FinBERT: {sentiment['finbert']:.3f}")
-            
-            # Summary
-            if 'summary' in article:
-                st.write("**Summary:**")
-                st.write(article['summary'])
-            
-            # Multilingual summaries
-            if 'summaries' in article:
-                for lang, summary in article['summaries'].items():
-                    if lang != 'English':
-                        st.write(f"**Summary ({lang}):**")
-                        st.write(summary)
-
-def display_sentiment_analysis(results):
-    """Display detailed sentiment analysis"""
-    st.subheader("🎯 Detailed Sentiment Analysis")
-    
-    # Model comparison
-    if results['articles']:
-        model_data = []
-        for article in results['articles']:
-            sentiment = article['sentiment']
-            row = {'title': article['title'][:30] + "..."}
-            if 'vader' in sentiment:
-                row['VADER'] = sentiment['vader']
-            if 'loughran_mcdonald' in sentiment:
-                row['Loughran-McDonald'] = sentiment['loughran_mcdonald']
-            if 'finbert' in sentiment:
-                row['FinBERT'] = sentiment['finbert']
-            row['Final Score'] = sentiment['compound']
-            model_data.append(row)
-        
-        df_models = pd.DataFrame(model_data)
-        st.write("**Model Comparison:**")
-        st.dataframe(df_models, use_container_width=True)
-        
-        # Correlation heatmap
-        numeric_cols = [col for col in df_models.columns if col != 'title']
-        if len(numeric_cols) > 1:
-            corr_matrix = df_models[numeric_cols].corr()
-            fig_heatmap = px.imshow(
-                corr_matrix,
-                text_auto=True,
-                aspect="auto",
-                color_continuous_scale="RdBu_r",
-                color_continuous_midpoint=0
-            )
-            fig_heatmap.update_layout(title="Model Correlation Matrix")
-            st.plotly_chart(fig_heatmap, use_container_width=True)
-    
-    # Top positive and negative articles
-    col1, col2 = st.columns(2)
-    
-    with col1:
-        st.write("**Most Positive Articles:**")
-        positive_articles = sorted(
-            results['articles'], 
-            key=lambda x: x['sentiment']['compound'], 
-            reverse=True
-        )[:5]
-        
-        for article in positive_articles:
-            st.write(f"• {article['title'][:50]}... ({article['sentiment']['compound']:.3f})")
-    
-    with col2:
-        st.write("**Most Negative Articles:**")
-        negative_articles = sorted(
-            results['articles'], 
-            key=lambda x: x['sentiment']['compound']
-        )[:5]
-        
-        for article in negative_articles:
-            st.write(f"• {article['title'][:50]}... ({article['sentiment']['compound']:.3f})")
-
-def display_audio_summaries(results):
-    """Display audio summaries for different languages"""
-    st.subheader("🎵 Audio Summaries")
-    
-    if 'audio_files' in results:
-        for lang, audio_file in results['audio_files'].items():
-            st.write(f"**{lang} Summary:**")
-            
-            # Create audio player
-            if os.path.exists(audio_file):
-                with open(audio_file, 'rb') as audio_file_obj:
-                    audio_bytes = audio_file_obj.read()
-                    st.audio(audio_bytes, format='audio/mp3')
+
+    # PDF
+    with c3:
+        try:
+            pdf_obj = generate_pdf_report(results)  # may return bytes or a file path
+            if isinstance(pdf_obj, (bytes, bytearray)):
+                pdf_bytes = pdf_obj
             else:
-                st.write("Audio file not found")
-    else:
-        st.info("No audio summaries available. Enable audio generation in settings.")
-
-def display_export_options(results):
-    """Display export options"""
-    st.subheader("📤 Export Results")
-    
-    col1, col2, col3 = st.columns(3)
-    
-    with col1:
-        # CSV Export
-        if st.button("📊 Download CSV", use_container_width=True):
-            csv_data = prepare_csv_export(results)
-            st.download_button(
-                label="Click to Download CSV",
-                data=csv_data,
-                file_name=f"news_analysis_{datetime.now().strftime('%Y%m%d_%H%M')}.csv",
-                mime="text/csv"
-            )
-    
-    with col2:
-        # JSON Export
-        if st.button("📋 Download JSON", use_container_width=True):
-            json_data = json.dumps(results, indent=2, default=str)
+                # Assume it's a file path
+                with open(pdf_obj, "rb") as f:
+                    pdf_bytes = f.read()
+
             st.download_button(
-                label="Click to Download JSON",
-                data=json_data,
-                file_name=f"news_analysis_{datetime.now().strftime('%Y%m%d_%H%M')}.json",
-                mime="application/json"
+                "Download PDF",
+                data=pdf_bytes,
+                file_name=f"news_analysis_{results['query']}_{datetime.now().strftime('%Y%m%d_%H%M%S')}.pdf",
+                mime="application/pdf",
+                use_container_width=True,
             )
-    
-    with col3:
-        # PDF Report
-        if st.button("📄 Generate PDF Report", use_container_width=True):
+        except Exception as e:
+            st.info("PDF generator not available or failed. You can still export JSON/CSV.")
+            logger.exception(f"PDF generation failed: {e}")
+
+
+def _render_audio(audio_files: Dict[str, Any]):
+    if not audio_files:
+        return
+    st.subheader("Audio summaries")
+    for lang, path in audio_files.items():
+        if path:
+            st.markdown(f"**{lang}**")
             try:
-                pdf_buffer = generate_pdf_report(results)
-                st.download_button(
-                    label="Click to Download PDF",
-                    data=pdf_buffer,
-                    file_name=f"news_analysis_report_{datetime.now().strftime('%Y%m%d_%H%M')}.pdf",
-                    mime="application/pdf"
-                )
-            except Exception as e:
-                st.error(f"Error generating PDF: {str(e)}")
-
-def display_api_info(results):
-    """Display API information and examples"""
-    st.subheader("🔌 API Access")
-    
-    st.write("**Endpoint:** `/api/analyze`")
-    st.write("**Method:** GET")
-    st.write("**Parameters:**")
-    st.code("""
-    - query: string (required) - Company name, ticker, or keyword
-    - num_articles: integer (default: 20) - Number of articles to analyze
-    - languages: array (default: ["English"]) - Summary languages
-    - include_audio: boolean (default: true) - Generate audio summaries
-    - sentiment_models: array (default: ["VADER", "Loughran-McDonald", "FinBERT"]) - Models to use
-    """)
-    
-    st.write("**Example Request:**")
-    st.code(f"GET /api/analyze?query={results['query']}&num_articles=20")
-    
-    st.write("**Sample Response:**")
-    sample_response = {
-        "query": results['query'],
-        "total_articles": len(results['articles']),
-        "average_sentiment": results['summary']['average_sentiment'],
-        "articles": results['articles'][:2]  # Show first 2 articles as example
-    }
-    st.json(sample_response)
-
-def prepare_csv_export(results):
-    """Prepare CSV data for export"""
-    csv_data = []
-    
-    for article in results['articles']:
-        row = {
-            'title': article['title'],
-            'source': article['source'],
-            'url': article.get('url', ''),
-            'date': article.get('date', ''),
-            'sentiment_compound': article['sentiment']['compound'],
-            'sentiment_label': 'Positive' if article['sentiment']['compound'] > 0.1 else 'Negative' if article['sentiment']['compound'] < -0.1 else 'Neutral',
-            'summary': article.get('summary', '')
-        }
-        
-        # Add model-specific scores
-        if 'vader' in article['sentiment']:
-            row['vader_score'] = article['sentiment']['vader']
-        if 'loughran_mcdonald' in article['sentiment']:
-            row['loughran_mcdonald_score'] = article['sentiment']['loughran_mcdonald']
-        if 'finbert' in article['sentiment']:
-            row['finbert_score'] = article['sentiment']['finbert']
-        
-        csv_data.append(row)
-    
-    df = pd.DataFrame(csv_data)
-    return df.to_csv(index=False)
-
-def show_demo_dashboard():
-    """Show demo dashboard with sample data"""
-    st.header("🚀 Welcome to Global Business News Intelligence")
-    
-    st.markdown("""
-    ### Key Features:
-    - **🔍 Multi-Source News Scraping:** Aggregates news from reliable sources
-    - **🎯 Advanced Sentiment Analysis:** Uses VADER, Loughran-McDonald, and FinBERT models
-    - **🌐 Multilingual Support:** Summaries in English, Hindi, and Tamil
-    - **🎵 Audio Generation:** Text-to-speech for all language summaries
-    - **📊 Interactive Dashboard:** Real-time charts and visualizations
-    - **📤 Multiple Export Formats:** CSV, JSON, and PDF reports
-    - **🔌 API Access:** Programmatic access to all features
-    
-    ### Use Cases:
-    - **📈 Investment Research:** Track sentiment around stocks and companies
-    - **🏢 Brand Monitoring:** Monitor public perception of your brand
-    - **🔍 Market Intelligence:** Stay informed about industry trends
-    - **📰 Media Analysis:** Analyze coverage patterns across sources
-    - **🌍 Global Insights:** Access news in multiple languages
-    
-    ### Get Started:
-    1. Enter a company name, stock ticker, or keyword in the sidebar
-    2. Configure your analysis settings
-    3. Click "Analyze News" to start
-    4. Explore results in the interactive dashboard
-    5. Export your findings in multiple formats
-    """)
-    
-    # Sample visualization
-    st.subheader("📊 Sample Analysis Dashboard")
-    
-    # Create sample data
-    sample_data = {
-        'Sentiment': ['Positive', 'Negative', 'Neutral'],
-        'Count': [45, 15, 40]
+                with open(path, "rb") as f:
+                    st.audio(f.read(), format="audio/mp3")
+            except Exception:
+                # Some Spaces require passing the path directly
+                st.audio(path)
+
+
+# ------------------------------------------------------------------------------
+# Main flow
+# ------------------------------------------------------------------------------
+
+if run_btn:
+    st.info("Starting analysis… this may take ~30–60 seconds on a CPU Space (FinBERT/summarizer/translation are heavy).")
+
+    progress = st.progress(0, text="Initializing…")
+
+    def _cb(p: int, status: str):
+        try:
+            progress.progress(p, text=status)
+        except Exception:
+            pass
+
+    config = {
+        "query": default_query,
+        "num_articles": num_articles,
+        "languages": languages or ["English"],
+        "include_audio": include_audio,
+        "sentiment_models": sentiment_models or ["VADER", "Loughran-McDonald", "FinBERT"],
     }
-    
-    fig = px.pie(
-        values=sample_data['Count'],
-        names=sample_data['Sentiment'],
-        color_discrete_map={'Positive': '#28a745', 'Negative': '#dc3545', 'Neutral': '#6c757d'},
-        title="Sample Sentiment Distribution"
-    )
-    
-    col1, col2 = st.columns([1, 1])
-    with col1:
-        st.plotly_chart(fig, use_container_width=True)
-    
-    with col2:
-        st.write("**Sample Metrics:**")
-        st.metric("Articles Analyzed", "100")
-        st.metric("Average Sentiment", "0.234")
-        st.metric("Sources Covered", "15")
-        st.metric("Languages", "3")
-
-if __name__ == "__main__":
-    main()
+
+    try:
+        results: Dict[str, Any] = analyzer.analyze_news(config, progress_callback=_cb)
+    except Exception as e:
+        progress.empty()
+        st.error(f"Analysis failed: {e}")
+        st.stop()
+
+    progress.empty()
+
+    # Handle empty gracefully
+    if not results.get("articles"):
+        st.warning("No articles found or scraping failed. Try a different query or reduce filters.")
+        st.stop()
+
+    # Header summary
+    st.subheader(f"Results — {results['query']}")
+    dist = results["summary"]["distribution"]
+    _render_distribution(dist)
+
+    # Keywords
+    if results.get("keywords"):
+        top_kw = ", ".join(kw["keyword"] for kw in results["keywords"][:12])
+        st.markdown(f"**Top keywords:** {top_kw}")
+
+    # Articles table
+    df = _articles_to_df(results["articles"])
+    st.dataframe(df, use_container_width=True, hide_index=True)
+
+    # Audio (optional)
+    if results.get("audio_files"):
+        _render_audio(results["audio_files"])
+
+    # Exports
+    st.divider()
+    _download_buttons(results, df)
+
+else:
+    st.info("Enter a company/keyword on the left and click Analyze. Example: Tesla, Nvidia, Reliance, HDFC, Adani, BYD.")
+
+# Footer
+st.markdown(
+    "<p class='small'>Built with Streamlit + FastAPI · CPU-only · "
+    "FinBERT/VADER/LM sentiment · BART/T5 summarization · YAKE keywords · gTTS audio.</p>",
+    unsafe_allow_html=True,
+)