Update app.py

app.py

@@ -66,6 +66,78 @@ selected_dataset_ai = "Volkswagen Customers"
 df_builder_pivot_str = ""
 
 
+def plot_bucket_fullness(driver_df, title):
+    # Set image path
+    image_path = "./images/image.png"
+
+    # Load background image
+    try:
+        img = Image.open(image_path)
+    except FileNotFoundError:
+        raise FileNotFoundError(f"❌ Error: Background image '{image_path}' not found!")
+
+    # Trust categories (fixed order)
+    categories = [
+        "Vision Trust", "Development Trust", "Benefit Trust",
+        "Competence Trust", "Stability Trust", "Relationship Trust"
+    ]
+    
+    # Extract fullness values from the DataFrame
+    percentages = driver_df.mean().tolist()
+
+    # Colors for each bubble
+    colors = ["#DF8859", "#E3B05B", "#418387", "#6D93AB", "#375570", "#C63F48"]
+
+    # Bubble positions (aligned with background)
+    bubble_positions = [
+        (0.66, 1.20),   # Vision (moved up)
+        (1.4, -0.10),    # Development
+        (0.66, -1.10),  # Benefit
+        (-0.70, -1.20), # Competence
+        (-1.35, 0.0),  # Stability
+        (-0.70, 1.15)   # Relationship (shifted left)
+    ]
+
+    # Scale bubble sizes dynamically based on fullness values
+    max_size = 0.35  # Maximum bubble size
+    min_size = 0.18  # Minimum bubble size
+    min_value, max_value = min(percentages), max(percentages)
+    
+    # Normalize values to fit in the bubble size range
+    bubble_sizes = [
+        min_size + (max_size - min_size) * ((p - min_value) / (max_value - min_value + 1e-5))
+        for p in percentages
+    ]
+
+    # Create the figure
+    fig, ax = plt.subplots(figsize=(8, 8))
+    ax.set_xlim(-2, 2)
+    ax.set_ylim(-2, 2)
+    ax.axis("off")
+
+    # Display background image
+    ax.imshow(img, extent=[-1.5, 1.5, -1.5, 1.5])
+
+    # Draw bubbles with dynamically scaled sizes
+    for i, (x, y) in enumerate(bubble_positions):
+        size = bubble_sizes[i]
+        circle = patches.Circle((x, y), size, facecolor=colors[i], alpha=1.0, lw=1.5)
+        ax.add_patch(circle)
+
+        # **Add percentage inside the bubble**
+        ax.text(
+            x, y, f"{percentages[i]:.1f}%", fontsize=12, fontweight="bold",
+            ha="center", va="center", color="white"
+        )
+
+    # Save plot to a buffer
+    img_buffer = io.BytesIO()
+    plt.savefig(img_buffer, format="png", bbox_inches="tight", facecolor=fig.get_facecolor())
+    img_buffer.seek(0)
+    plt.close(fig)
+
+    # Convert to PIL Image for Gradio display
+    return Image.open(img_buffer)
 def plot_model_results(results_df, average_value, title, model_type):
     """
     Plot model results with specific orders and colors for Trust and NPS models.
@@ -516,20 +588,20 @@ def analyze_excel_single(file_path):
 
         # plot trust
         # Get n_samples from output text
-        n_samples_trust = output_text.split(": Trust")[1]
+        n n_samples_trust = output_text.split(": Trust")[1]
         n_samples_trust = n_samples_trust.split("Analysis based on ")[1]
         n_samples_trust = n_samples_trust.split("observations")[0]
 
-        results_df_trust = None
+       
         results_df_trust = pd.read_csv(csv_output_path_trust)
         results_df_trust["Importance_percent"] = results_df_trust["Importance"] * 100
         average_value_trust = results_df_trust["Importance_percent"].mean()
-
-        img_trust = plot_model_results(
+        
+        # Instead of calling plot_model_results for Trust Drivers,
+        # call the separate bubble plot function:
+        img_trust = plot_trust_driver_bubbles(
             results_df_trust,
-            average_value_trust,
-            f"Trust Drivers: {file_name}",
-            "Trust",
+            f"Trust Drivers: {file_name}"
         )
         display_trust_score_1()