Update app.py

app.py

@@ -288,6 +288,10 @@ df_combined["chaos_score"] = np.log1p(df_combined.get("diag_srl_gamma", 0)) / (d
 # ---------------------------------------------------------------
 # Core Visualization and Analysis Functions
 # ---------------------------------------------------------------
+# Cache for resolved audio paths and CMT data to avoid repeated computations
+_audio_path_cache = {}
+_cmt_data_cache = {}
+
 def resolve_audio_path(row: pd.Series) -> str:
     """
     Intelligently reconstructs the full path to an audio file
@@ -299,69 +303,70 @@ def resolve_audio_path(row: pd.Series) -> str:
     basename = str(row.get("filepath", ""))
     source = row.get("source", "")
     label = row.get("label", "")
+    
+    # Check cache first
+    cache_key = f"{source}:{label}:{basename}"
+    if cache_key in _audio_path_cache:
+        return _audio_path_cache[cache_key]
 
-    # Debug information
-    print(f"🔍 Resolving audio path for: {basename} ({source}, {label})")
+    resolved_path = basename  # Default fallback
 
     # For "Dog" data, the structure is: combined/{label}/{filename}
     if source == "Dog":
         # Try with label subdirectory first
         expected_path = os.path.join(DOG_AUDIO_BASE_PATH, label, basename)
-        print(f"  Trying dog path: {expected_path} (exists: {os.path.exists(expected_path)})")
         if os.path.exists(expected_path):
-            return expected_path
-        
-        # Try without subdirectory in case files are flat
-        expected_path = os.path.join(DOG_AUDIO_BASE_PATH, basename)
-        print(f"  Trying flat dog path: {expected_path} (exists: {os.path.exists(expected_path)})")
-        if os.path.exists(expected_path):
-            return expected_path
+            resolved_path = expected_path
+        else:
+            # Try without subdirectory in case files are flat
+            expected_path = os.path.join(DOG_AUDIO_BASE_PATH, basename)
+            if os.path.exists(expected_path):
+                resolved_path = expected_path
 
     # For "Human" data, search within all "Actor_XX" subfolders
     elif source == "Human":
         if os.path.isdir(HUMAN_AUDIO_BASE_PATH):
-            print(f"  Searching in human base: {HUMAN_AUDIO_BASE_PATH}")
             for actor_folder in os.listdir(HUMAN_AUDIO_BASE_PATH):
                 if actor_folder.startswith("Actor_"):
                     expected_path = os.path.join(HUMAN_AUDIO_BASE_PATH, actor_folder, basename)
-                    print(f"    Trying: {expected_path} (exists: {os.path.exists(expected_path)})")
                     if os.path.exists(expected_path):
-                        return expected_path
+                        resolved_path = expected_path
+                        break
         
         # Try without subdirectory in case files are flat
-        expected_path = os.path.join(HUMAN_AUDIO_BASE_PATH, basename)
-        print(f"  Trying flat human path: {expected_path} (exists: {os.path.exists(expected_path)})")
-        if os.path.exists(expected_path):
-            return expected_path
-    
-    # Fallback for dummy data or other cases
-    print(f"  Trying basename directly: {basename} (exists: {os.path.exists(basename)})")
-    if os.path.exists(basename):
-        return basename
+        if resolved_path == basename:
+            expected_path = os.path.join(HUMAN_AUDIO_BASE_PATH, basename)
+            if os.path.exists(expected_path):
+                resolved_path = expected_path
     
     # Try in local directories (for dummy data)
-    if source == "Dog":
-        for label_dir in ["bark", "growl", "whine", "pant"]:
-            local_path = os.path.join(DOG_DIR, label_dir, basename)
-            print(f"  Trying local dog path: {local_path} (exists: {os.path.exists(local_path)})")
+    if resolved_path == basename:
+        if source == "Dog":
+            for label_dir in ["bark", "growl", "whine", "pant"]:
+                local_path = os.path.join(DOG_DIR, label_dir, basename)
+                if os.path.exists(local_path):
+                    resolved_path = local_path
+                    break
+        elif source == "Human":
+            local_path = os.path.join(HUMAN_DIR, "Actor_01", basename)
             if os.path.exists(local_path):
-                return local_path
-    elif source == "Human":
-        local_path = os.path.join(HUMAN_DIR, "Actor_01", basename)
-        print(f"  Trying local human path: {local_path} (exists: {os.path.exists(local_path)})")
-        if os.path.exists(local_path):
-            return local_path
+                resolved_path = local_path
     
-    # If all else fails, return the original basename
-    print(f"  ❌ No valid path found, returning basename: {basename}")
-    return basename
+    # Cache the result
+    _audio_path_cache[cache_key] = resolved_path
+    return resolved_path
 
 def get_cmt_data(filepath: str, lens: str):
+    # Check cache first
+    cache_key = f"{filepath}:{lens}"
+    if cache_key in _cmt_data_cache:
+        return _cmt_data_cache[cache_key]
+    
     try:
         y, _ = sf.read(filepath)
         if y.ndim > 1: y = np.mean(y, axis=1)
     except Exception as e:
-        print(f"Error reading audio file {filepath}: {e}")
+        print(f"Error reading audio file {os.path.basename(filepath)}: {e}")
         return None
 
     cmt = ExpandedCMT()
@@ -371,11 +376,15 @@ def get_cmt_data(filepath: str, lens: str):
     # The _apply_lens function now returns additional diagnostic info
     phi, w, z, original_count, final_count = cmt._apply_lens(encoded, lens)
 
-    return {
+    result = {
         "phi": phi, "w": w, "z": z,
         "original_count": original_count,
         "final_count": final_count
     }
+    
+    # Cache the result
+    _cmt_data_cache[cache_key] = result
+    return result
 
 def generate_holographic_field(z: np.ndarray, phi: np.ndarray, resolution: int):
     if z is None or phi is None or len(z) < 4: return None
@@ -725,7 +734,7 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="teal", secondary_hue="cyan")) a
             def update_cross_species_view(species, primary_file, neighbor_file, lens, resolution, wavelength):
                 if not primary_file:
                     empty_fig = go.Figure(layout={"title": "Please select a primary file."})
-                    return empty_fig, empty_fig, "", "", None, None, []
+                    return empty_fig, empty_fig, "", "", None, None
 
                 # Get primary row
                 primary_row = df_combined[
@@ -756,13 +765,16 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="teal", secondary_hue="cyan")) a
                         (df_combined["source"] == opposite_species)
                     ]) > 0 else None
 
-                # Get CMT data for both files
+                # Get CMT data for both files (with progress indication)
                 primary_fp = resolve_audio_path(primary_row)
+                print(f"🎵 Processing primary audio: {os.path.basename(primary_fp)}")
                 primary_cmt = get_cmt_data(primary_fp, lens)
                 
                 neighbor_cmt = None
+                neighbor_fp = None
                 if neighbor_row is not None:
                     neighbor_fp = resolve_audio_path(neighbor_row)
+                    print(f"🎵 Processing neighbor audio: {os.path.basename(neighbor_fp)}")
                     neighbor_cmt = get_cmt_data(neighbor_fp, lens)
 
                 # Create visualizations
@@ -840,12 +852,22 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="teal", secondary_hue="cyan")) a
                                    primary_info_html, neighbor_info_html,
                                    primary_audio_out, neighbor_audio_out]
 
-            for component in cross_species_inputs:
-                component.change(update_cross_species_view, 
-                               inputs=cross_species_inputs, 
-                               outputs=cross_species_outputs)
-
-            # Initialize on load - let the pre-populated dropdowns handle the initial state
+            # Only bind change events, not load events to avoid overwhelming initialization
+            primary_dropdown.change(update_cross_species_view, 
+                                   inputs=cross_species_inputs, 
+                                   outputs=cross_species_outputs)
+            neighbor_dropdown.change(update_cross_species_view, 
+                                   inputs=cross_species_inputs, 
+                                   outputs=cross_species_outputs)
+            holo_lens_dropdown.change(update_cross_species_view, 
+                                    inputs=cross_species_inputs, 
+                                    outputs=cross_species_outputs)
+            holo_resolution_slider.change(update_cross_species_view, 
+                                        inputs=cross_species_inputs, 
+                                        outputs=cross_species_outputs)
+            holo_wavelength_slider.change(update_cross_species_view, 
+                                        inputs=cross_species_inputs, 
+                                        outputs=cross_species_outputs)
 
 if __name__ == "__main__":
     demo.launch(share=True, debug=True)