Update app.py

app.py

@@ -5,37 +5,118 @@ import numpy as np
 from word2number import w2n
 import re
 from typing import Tuple, List, Dict
+from io import BytesIO
+import base64
 
 # Custom CSS for styling
 css = """
-.risk-low { color: #28a745; font-weight: bold; }
-.risk-medium { color: #ffc107; font-weight: bold; }
-.risk-high { color: #dc3545; font-weight: bold; }
-.result-box { padding: 20px; border-radius: 5px; margin-bottom: 20px; }
-.penalty-box { background-color: #f8f9fa; }
-.obligation-box { background-color: #f8f9fa; }
-.delay-box { background-color: #f8f9fa; }
-.risk-container { 
-    display: flex; 
-    justify-content: space-between; 
-    margin-bottom: 20px;
-    gap: 10px;
+:root {
+    --low-color: #28a745;
+    --medium-color: #ffc107;
+    --high-color: #dc3545;
+    --inactive-color: #e9ecef;
 }
-.risk-item {
-    text-align: center;
+.risk-container {
+    display: flex;
+    flex-direction: column;
+    gap: 12px;
+    margin-bottom: 25px;
+}
+.risk-row {
+    display: flex;
+    align-items: center;
+    background: white;
+    border-radius: 8px;
     padding: 15px;
-    border-radius: 5px;
-    width: 100%;
-    background-color: #f8f9fa;
+    box-shadow: 0 2px 4px rgba(0,0,0,0.1);
+    transition: all 0.3s ease;
+}
+.risk-row.active {
+    transform: scale(1.02);
+    box-shadow: 0 4px 8px rgba(0,0,0,0.15);
+}
+.risk-label {
+    width: 100px;
+    font-weight: 600;
+    font-size: 16px;
+    color: #495057;
 }
+.risk-score {
+    width: 80px;
+    font-size: 20px;
+    font-weight: 700;
+    text-align: center;
+}
+.risk-low { color: var(--low-color); }
+.risk-medium { color: var(--medium-color); }
+.risk-high { color: var(--high-color); }
 .heatmap-container {
-    display: flex;
-    justify-content: space-between;
-    gap: 10px;
+    flex-grow: 1;
+    height: 30px;
+    border-radius: 15px;
+    overflow: hidden;
+    position: relative;
+}
+.heatmap-bar {
+    height: 100%;
+    border-radius: 15px;
+    transition: width 0.5s ease;
+}
+.risk-meter {
+    position: absolute;
+    right: 10px;
+    top: 50%;
+    transform: translateY(-50%);
+    font-size: 12px;
+    font-weight: 600;
+    color: white;
+    text-shadow: 0 1px 2px rgba(0,0,0,0.3);
+}
+.result-section {
+    background: white;
+    border-radius: 8px;
+    padding: 20px;
     margin-bottom: 20px;
+    box-shadow: 0 2px 4px rgba(0,0,0,0.1);
+}
+.result-title {
+    font-size: 18px;
+    font-weight: 600;
+    margin-bottom: 15px;
+    color: #343a40;
+    display: flex;
+    align-items: center;
+    gap: 8px;
+}
+.clause-item {
+    margin-bottom: 8px;
+    padding-left: 15px;
+    position: relative;
+}
+.clause-item:before {
+    content: "•";
+    position: absolute;
+    left: 0;
+    color: #6c757d;
+}
+.penalty-amount {
+    font-family: monospace;
+    background: #f8f9fa;
+    padding: 2px 6px;
+    border-radius: 4px;
+    margin-left: 5px;
 }
-.heatmap-item {
-    width: 100%;
+.example-clause {
+    background: #f8f9fa;
+    padding: 12px;
+    border-radius: 6px;
+    margin-bottom: 10px;
+    border-left: 3px solid #6c757d;
+}
+.example-number {
+    font-weight: 600;
+    margin-right: 8px;
+    color: #6c757d;
 }
 """
 
@@ -105,66 +186,59 @@ def calculate_risk_score(penalty_count: int, penalty_values: List[float], obliga
     else:
         return score, "High"
 
-def generate_heatmap(score: float, level: str) -> plt.Figure:
-    """Generate a heatmap for a specific risk level"""
-    fig, ax = plt.subplots(figsize=(6, 1))
-    
-    if level == "Low":
-        cmap = plt.cm.Greens
-        color = '#28a745'
-    elif level == "Medium":
-        cmap = plt.cm.Oranges
-        color = '#ffc107'
-    else:
-        cmap = plt.cm.Reds
-        color = '#dc3545'
-    
-    gradient = np.linspace(0, 1, 256).reshape(1, -1)
-    gradient = np.vstack((gradient, gradient))
-    
-    ax.imshow(gradient, aspect='auto', cmap=cmap)
-    ax.text(128, 0.5, f"{level}: {score:.1f}%" if score > 0 else f"{level}: 0%", 
-            color='white' if level == "High" else 'black', 
-            ha='center', va='center', fontsize=14, fontweight='bold')
-    
-    ax.set_axis_off()
-    plt.tight_layout()
-    return fig
-
-def generate_risk_display(risk_score: float, risk_level: str) -> Tuple[str, plt.Figure, plt.Figure, plt.Figure]:
-    """Generate HTML display and heatmaps for all three risk levels"""
+def create_risk_display(risk_score: float, risk_level: str) -> str:
+    """Create HTML display for all three risk levels"""
     risk_levels = ["Low", "Medium", "High"]
+    colors = {
+        "Low": "var(--low-color)",
+        "Medium": "var(--medium-color)",
+        "High": "var(--high-color)"
+    }
     
-    # Generate heatmaps for all levels
-    heatmaps = []
-    for level in risk_levels:
-        if level == risk_level:
-            score = risk_score
-        else:
-            score = 0
-        heatmaps.append(generate_heatmap(score, level))
+    html_parts = []
+    html_parts.append("<div class='risk-container'>")
     
-    # Generate HTML display
-    risk_html = """
-    <div class='risk-container'>
-    """
     for level in risk_levels:
-        if level == risk_level:
-            value = f"{risk_score:.1f}%"
-            active_class = f"risk-{level.lower()}"
-        else:
-            value = "0%"
-            active_class = ""
+        active = level == risk_level
+        score = risk_score if active else 0
+        color = colors[level] if active else "var(--inactive-color)"
+        opacity = "1" if active else "0.6"
         
-        risk_html += f"""
-        <div class='risk-item {active_class}'>
-            <h3>{level} Risk</h3>
-            <div style='font-size: 24px;'>{value}</div>
+        html_parts.append(f"""
+        <div class='risk-row {'active' if active else ''}'>
+            <div class='risk-label risk-{level.lower()}'>{level} Risk</div>
+            <div class='risk-score risk-{level.lower()}'>{score:.1f}%</div>
+            <div class='heatmap-container'>
+                <div class='heatmap-bar" 
+                     style="width: {score}%; background: {color}; opacity: {opacity}">
+                    <span class='risk-meter'>{score:.1f}%</span>
+                </div>
+            </div>
         </div>
-        """
+        """)
     
-    risk_html += "</div>"
-    return risk_html, heatmaps[0], heatmaps[1], heatmaps[2]
+    html_parts.append("</div>")
+    return "\n".join(html_parts)
+
+def format_clauses(counts: Dict[str, int]) -> str:
+    """Format clause counts into HTML"""
+    return "\n".join([f"<div class='clause-item'>{kw}: <strong>{count}</strong></div>" for kw, count in counts.items()])
+
+def format_penalty_amounts(amounts: List[float]) -> str:
+    """Format penalty amounts into HTML"""
+    if not amounts:
+        return "<div style='color: #6c757d;'>No specific penalty amounts found</div>"
+    return "\n".join([f"<div class='clause-item'><span class='penalty-amount'>${amt:,.2f}</span></div>" for amt in amounts[:5]])
+
+def format_examples(sentences: List[str]) -> str:
+    """Format example sentences into HTML"""
+    if not sentences:
+        return "<div style='color: #6c757d;'>No penalty clauses found</div>"
+    return "\n".join([f"""
+    <div class='example-clause'>
+        <span class='example-number'>{i+1}.</span> {sent}
+    </div>
+    """ for i, sent in enumerate(sentences[:3])])
 
 def analyze_pdf(file_obj) -> List:
     """Main analysis function for Gradio interface"""
@@ -195,15 +269,8 @@ def analyze_pdf(file_obj) -> List:
             total_penalties, penalty_values, total_obligations, total_delays
         )
         
-        # Generate risk display and heatmaps
-        risk_display, low_heatmap, medium_heatmap, high_heatmap = generate_risk_display(risk_score, risk_level)
-        
-        # Prepare results
-        penalty_details = "\n".join([f"- {kw}: {count}" for kw, count in penalty_counts.items()])
-        obligation_details = "\n".join([f"- {kw}: {count}" for kw, count in obligation_counts.items()])
-        delay_details = "\n".join([f"- {kw}: {count}" for kw, count in delay_counts.items()])
-        
-        penalty_amounts = "\n".join([f"- ${amt:,.2f}" for amt in penalty_values[:5]]) if penalty_values else "No specific penalty amounts found"
+        # Generate risk display
+        risk_display = create_risk_display(risk_score, risk_level)
         
         # Find example sentences with penalties
         penalty_sentences = []
@@ -211,65 +278,93 @@ def analyze_pdf(file_obj) -> List:
             if any(kw.lower() in sentence.lower() for kw in penalty_keywords):
                 penalty_sentences.append(sentence.strip())
         
-        penalty_examples = "\n\n".join([f"{i+1}. {sent}" for i, sent in enumerate(penalty_sentences[:3])]) if penalty_sentences else "No penalty clauses found"
+        # Format all results
+        penalty_html = f"""
+        <div class='result-section'>
+            <div class='result-title'>📊 Penalty Clauses: <strong>{total_penalties}</strong> found</div>
+            {format_clauses(penalty_counts)}
+        </div>
+        """
+        
+        amounts_html = f"""
+        <div class='result-section'>
+            <div class='result-title'>💰 Penalty Amounts: <strong>{len(penalty_values)}</strong> found</div>
+            {format_penalty_amounts(penalty_values)}
+        </div>
+        """
+        
+        obligation_html = f"""
+        <div class='result-section'>
+            <div class='result-title'>⚖️ Obligation Clauses: <strong>{total_obligations}</strong> found</div>
+            {format_clauses(obligation_counts)}
+        </div>
+        """
+        
+        delay_html = f"""
+        <div class='result-section'>
+            <div class='result-title'>⏱️ Delay Clauses: <strong>{total_delays}</strong> found</div>
+            {format_clauses(delay_counts)}
+        </div>
+        """
+        
+        examples_html = f"""
+        <div class='result-section'>
+            <div class='result-title'>🔍 Example Penalty Clauses</div>
+            {format_examples(penalty_sentences)}
+        </div>
+        """
         
-        # Return all results
         return [
             risk_display,
-            low_heatmap,
-            medium_heatmap,
-            high_heatmap,
-            f"Total: {total_penalties}\n\n{penalty_details}",
-            f"{len(penalty_values)} amounts found\n\n{penalty_amounts}",
-            f"Total: {total_obligations}\n\n{obligation_details}",
-            f"Total: {total_delays}\n\n{delay_details}",
-            penalty_examples
+            penalty_html,
+            amounts_html,
+            obligation_html,
+            delay_html,
+            examples_html
         ]
     except Exception as e:
-        return [f"Error: {str(e)}"] * 9
+        error_html = f"""
+        <div class='result-section' style='background: #fff3cd;'>
+            <div class='result-title'>❌ Error</div>
+            <div>{str(e)}</div>
+        </div>
+        """
+        return [error_html] * 6
 
 # Create Gradio interface
 with gr.Blocks(css=css, title="PDF Contract Risk Analyzer") as demo:
-    gr.Markdown("# 📄 PDF Contract Risk Analyzer")
-    gr.Markdown("Upload a contract PDF to analyze penalties, obligations, and delays.")
+    gr.Markdown("""
+    <div style='text-align: center; margin-bottom: 30px;'>
+        <h1 style='margin-bottom: 10px;'>📄 PDF Contract Risk Analyzer</h1>
+        <p style='color: #6c757d;'>Upload a contract PDF to analyze penalties, obligations, and delays</p>
+    </div>
+    """)
     
     with gr.Row():
-        with gr.Column():
+        with gr.Column(scale=1):
             file_input = gr.File(label="Upload PDF", file_types=[".pdf"])
-            submit_btn = gr.Button("Analyze PDF", variant="primary")
+            submit_btn = gr.Button("Analyze Contract", variant="primary")
         
-        with gr.Column():
-            gr.Markdown("### 🔍 Overall Risk Assessment")
-            risk_display = gr.HTML(label="Risk Levels")
-            with gr.Row():
-                low_heatmap = gr.Plot(label="Low Risk")
-                medium_heatmap = gr.Plot(label="Medium Risk")
-                high_heatmap = gr.Plot(label="High Risk")
+        with gr.Column(scale=3):
+            gr.Markdown("### 🔍 Risk Assessment Summary")
+            risk_display = gr.HTML()
     
     with gr.Row():
         with gr.Column():
-            gr.Markdown("### 📊 Penalties Analysis")
-            penalty_count = gr.Textbox(label="Penalty Clauses", lines=5)
-            penalty_amounts = gr.Textbox(label="Penalty Amounts", lines=5)
+            penalty_count = gr.HTML()
+            penalty_amounts = gr.HTML()
         
         with gr.Column():
-            gr.Markdown("### ⚖️ Obligations Analysis")
-            obligation_count = gr.Textbox(label="Obligation Clauses", lines=5)
-        
-        with gr.Column():
-            gr.Markdown("### ⏱️ Delays Analysis")
-            delay_count = gr.Textbox(label="Delay Clauses", lines=5)
+            obligation_count = gr.HTML()
+            delay_count = gr.HTML()
     
-    with gr.Row():
-        gr.Markdown("### 🔎 Extracted Penalty Clauses")
-        penalty_examples = gr.Textbox(label="Example Penalty Clauses", lines=5)
+    penalty_examples = gr.HTML()
     
     submit_btn.click(
         fn=analyze_pdf,
         inputs=file_input,
-        outputs=[risk_display, low_heatmap, medium_heatmap, high_heatmap, 
-                penalty_count, penalty_amounts, obligation_count, 
-                delay_count, penalty_examples]
+        outputs=[risk_display, penalty_count, penalty_amounts, 
+                obligation_count, delay_count, penalty_examples]
     )
 
 if __name__ == "__main__":