Update app.py

app.py

@@ -108,35 +108,67 @@ def plot_scatter_tab4(cat, subcat, x, y, col):
     return fig
 
 # Tab 5
+# def plot_scatter_tab5(cat, x, y, z, col):
+#     if cat != "All":
+#         data = raw_data[raw_data["Category"] == cat]
+#     else:
+#         data = raw_data
+#     # Group and normalize the data
+#     grouped_cat = data.groupby(["model", "tag"]).size().reset_index(name="count").sort_values(by="count", ascending=False)
+#     grouped_cat["count"] = grouped_cat.groupby(["model"])["count"].transform(lambda x: x / x.sum())
+
+#     pivot_df = grouped_cat.pivot(index='model', columns='tag', values='count').fillna(0).reset_index()
+
+#     if col == "Size":
+#         pivot_df[col] = pivot_df["model"].map(size_map)
+#     else:
+#         pivot_df[col] = pivot_df["model"].str.split("/").str[0]
+
+#     print("\nDEBUG: pivot_df.head():\n", pivot_df.head())
+#     print("\nDEBUG: pivot_df shape", pivot_df.shape)
+#     print("\nDEBUG: pivot_df columns", pivot_df.columns)
+#     print("\nDEBUG: Unique values x/y/z", pivot_df[x].unique(), pivot_df[y].unique(), pivot_df[z].unique())
+    
+#     fig = px.scatter_3d(pivot_df, x=x, y=y, z=z,
+#         hover_name="model",
+#         title=f'{x} vs {y} vs {z}',
+#         color=col,
+#         color_continuous_scale="agsunset")
+    
+#     return fig
+
 def plot_scatter_tab5(cat, x, y, z, col):
+    print("DEBUG:", cat, x, y, z, col)
     if cat != "All":
-        data = raw_data[raw_data["Category"] == cat]
+        data = raw_data[raw_data["Category"].str.strip().str.lower() == cat.strip().lower()]
     else:
         data = raw_data
-    # Group and normalize the data
+    print("DEBUG: data rows after cat filter:", data.shape[0])
+    if data.empty:
+        return px.scatter_3d(title="No data left after category filtering!")
     grouped_cat = data.groupby(["model", "tag"]).size().reset_index(name="count").sort_values(by="count", ascending=False)
     grouped_cat["count"] = grouped_cat.groupby(["model"])["count"].transform(lambda x: x / x.sum())
 
+    print("DEBUG: grouped_cat head:", grouped_cat.head())
     pivot_df = grouped_cat.pivot(index='model', columns='tag', values='count').fillna(0).reset_index()
+    print("DEBUG: pivot_df head:", pivot_df.head())
+
+    # Ensure chosen columns exist
+    for k in [x, y, z]:
+        if k not in pivot_df.columns:
+            print(f"DEBUG: Axis {k} not found in data columns: {list(pivot_df.columns)}")
+            return px.scatter_3d(title=f"No {k} tag data for models!")
 
     if col == "Size":
         pivot_df[col] = pivot_df["model"].map(size_map)
     else:
         pivot_df[col] = pivot_df["model"].str.split("/").str[0]
 
-    print("\nDEBUG: pivot_df.head():\n", pivot_df.head())
-    print("\nDEBUG: pivot_df shape", pivot_df.shape)
-    print("\nDEBUG: pivot_df columns", pivot_df.columns)
-    print("\nDEBUG: Unique values x/y/z", pivot_df[x].unique(), pivot_df[y].unique(), pivot_df[z].unique())
-    
-    fig = px.scatter_3d(pivot_df, x=x, y=y, z=z,
-        hover_name="model",
-        title=f'{x} vs {y} vs {z}',
-        color=col,
-        color_continuous_scale="agsunset")
-    
+    print("DEBUG: unique color values:", pivot_df[col].unique())
+    fig = px.scatter_3d(pivot_df, x=x, y=y, z=z, hover_name="model", title=f'{x} vs {y} vs {z}', color=col)
     return fig
 
+
 # Tab 6
 data_with_text = pd.read_csv("./tagged_data_with_text.csv")
 def random_sample(r: gr.Request):