added app.py

app.py

@@ -0,0 +1,200 @@
+# app.py (in project root)
+import os
+import sys
+from pathlib import Path
+import gradio as gr
+import pandas as pd
+
+# Add project to path
+PROJECT_ROOT = Path(__file__).parent
+sys.path.insert(0, str(PROJECT_ROOT))
+
+# Disable HF progress bars
+os.environ["HF_HUB_DISABLE_PROGRESS_BARS"] = "1"
+os.environ["HF_HUB_DISABLE_TELEMETRY"] = "1"
+
+from core.config import EvolutionConfig
+from core.evolution.evolution import MolecularEvolution
+
+def generate_molecules(target_cn, minimize_ysi, optimization_mode, generations, population_size):
+    """Run the genetic algorithm and return results."""
+    
+    try:
+        # Create config
+        config = EvolutionConfig(
+            target_cn=float(target_cn),
+            maximize_cn=(optimization_mode == "Maximize CN"),
+            minimize_ysi=minimize_ysi,
+            generations=int(generations),
+            population_size=int(population_size)
+        )
+        
+        # Run evolution
+        evolution = MolecularEvolution(config)
+        final_df, pareto_df = evolution.evolve()
+        
+        # Prepare outputs
+        if final_df.empty:
+            return "❌ No valid molecules generated", None, None
+        
+        # Format results
+        summary = f"""
+        ✅ **Generation Complete!**
+        
+        - Total molecules: {len(final_df)}
+        - Best CN: {final_df.iloc[0]['cn']:.2f}
+        - Best CN error: {final_df.iloc[0]['cn_error']:.2f}
+        """
+        
+        if minimize_ysi and 'ysi' in final_df.columns:
+            summary += f"\n- Best YSI: {final_df.iloc[0]['ysi']:.2f}"
+            summary += f"\n- Pareto front size: {len(pareto_df) if not pareto_df.empty else 0}"
+        
+        # Return top 20 for display
+        display_cols = ['rank', 'smiles', 'cn', 'cn_error']
+        if 'ysi' in final_df.columns:
+            display_cols.append('ysi')
+        display_cols.extend(['bp', 'density', 'lhv', 'dynamic_viscosity'])
+        
+        available_cols = [c for c in display_cols if c in final_df.columns]
+        
+        top_molecules = final_df.head(20)[available_cols]
+        pareto_molecules = pareto_df.head(20)[available_cols] if not pareto_df.empty else None
+        
+        return summary, top_molecules, pareto_molecules
+        
+    except Exception as e:
+        return f"❌ Error: {str(e)}", None, None
+
+
+# Create Gradio interface
+with gr.Blocks(title="Biofuel Molecule Generator", theme=gr.themes.Soft()) as demo:
+    
+    gr.Markdown("""
+    # 🔬 Biofuel Molecule Generator
+    
+    Generate optimal fuel molecules using AI-powered genetic algorithms.
+    
+    **Features:**
+    - Multi-objective optimization (Cetane Number + Yield Sooting Index)
+    - Realistic chemical mutations (CREM)
+    - Constraint satisfaction (BP, density, viscosity, LHV)
+    - Pareto-optimal solutions
+    """)
+    
+    with gr.Row():
+        with gr.Column(scale=1):
+            gr.Markdown("### ⚙️ Configuration")
+            
+            # Optimization mode
+            optimization_mode = gr.Radio(
+                choices=["Target CN", "Maximize CN"],
+                value="Target CN",
+                label="Optimization Mode",
+                info="Target: Match specific CN | Maximize: Find highest CN"
+            )
+            
+            # Target CN (only used in Target mode)
+            target_cn = gr.Slider(
+                minimum=40,
+                maximum=80,
+                value=50,
+                step=1,
+                label="Target Cetane Number",
+                info="Desired CN value (used in Target mode)"
+            )
+            
+            # YSI minimization
+            minimize_ysi = gr.Checkbox(
+                value=True,
+                label="Minimize Yield Sooting Index (YSI)",
+                info="Reduce soot formation (environmental impact)"
+            )
+            
+            gr.Markdown("### 🧬 Algorithm Parameters")
+            
+            generations = gr.Slider(
+                minimum=3,
+                maximum=10,
+                value=6,
+                step=1,
+                label="Generations",
+                info="More generations = better optimization (slower)"
+            )
+            
+            population_size = gr.Slider(
+                minimum=20,
+                maximum=100,
+                value=50,
+                step=10,
+                label="Population Size",
+                info="More molecules = better exploration (slower)"
+            )
+            
+            gr.Markdown("""
+            ⚠️ **Note:** Larger settings take longer (5-15 minutes)
+            
+            **Recommended:**
+            - Quick test: 3 gen, 20 pop (~2 min)
+            - Standard: 6 gen, 50 pop (~8 min)
+            - Best results: 10 gen, 100 pop (~20 min)
+            """)
+            
+            generate_btn = gr.Button("🚀 Generate Molecules", variant="primary", size="lg")
+        
+        with gr.Column(scale=2):
+            gr.Markdown("### 📊 Results")
+            
+            summary_output = gr.Markdown(label="Summary")
+            
+            with gr.Tabs():
+                with gr.Tab("🏆 Best Candidates"):
+                    top_output = gr.Dataframe(
+                        label="Top 20 Molecules",
+                        interactive=False,
+                        wrap=True
+                    )
+                
+                with gr.Tab("📈 Pareto Front"):
+                    pareto_output = gr.Dataframe(
+                        label="Non-dominated Solutions (CN vs YSI trade-offs)",
+                        interactive=False,
+                        wrap=True
+                    )
+    
+    # Button click
+    generate_btn.click(
+        fn=generate_molecules,
+        inputs=[target_cn, minimize_ysi, optimization_mode, generations, population_size],
+        outputs=[summary_output, top_output, pareto_output]
+    )
+    
+    # Examples
+    gr.Markdown("### 💡 Example Configurations")
+    gr.Examples(
+        examples=[
+            [50, True, "Target CN", 6, 50],  # Standard multi-objective
+            [60, False, "Target CN", 5, 30],  # Single objective (CN only)
+            [50, True, "Maximize CN", 10, 100],  # Maximize with YSI constraint
+        ],
+        inputs=[target_cn, minimize_ysi, optimization_mode, generations, population_size],
+        label="Try these presets"
+    )
+    
+    gr.Markdown("""
+    ---
+    ### 📚 About
+    
+    This tool uses:
+    - **Machine Learning**: 6 property prediction models (R² > 0.90)
+    - **Genetic Algorithm**: CREM-based mutations for chemical validity
+    - **Multi-Objective Optimization**: Pareto fronts for trade-off analysis
+    
+    **Citation:** Salvina Za, [University], 2026
+    
+    [GitHub](https://github.com/SalZa2004/Biofuel-Optimiser-ML) | [Paper](link-when-published)
+    """)
+
+# Launch
+if __name__ == "__main__":
+    demo.launch()