ifccsv to web - fix Tabulator dropping columns with non-unique names

src/bonsai/bonsai/bim/module/csv/operator.py

@@ -16,6 +16,7 @@
 # You should have received a copy of the GNU General Public License
 # along with Bonsai.  If not, see <http://www.gnu.org/licenses/>.
 
+from __future__ import annotations
 import os
 import bpy
 import json
@@ -28,6 +29,11 @@ import bonsai.tool as tool
 import bonsai.bim.module.drawing.scheduler as scheduler
 from bonsai.bim.ifc import IfcStore
 from bonsai.bim.handler import refresh_ui_data
+from typing import TYPE_CHECKING
+from collections import Counter
+
+if TYPE_CHECKING:
+    import pandas as pd
 
 
 class AddCsvAttribute(bpy.types.Operator):
@@ -264,11 +270,23 @@ class ExportIfcCsv(bpy.types.Operator):
         if props.format == "web":
             if not context.scene.WebProperties.is_connected:
                 bpy.ops.bim.connect_websocket_server()
-            tool.Web.send_webui_data(data=ifc_csv.dataframe.to_csv(index=False), data_key="csv_data", event="csv_data")
+            df = ifc_csv.dataframe
+            assert df is not None
+            # Tabulator seems to be ignoring columns non-unique columns,
+            # so we ensure they are unique at input.
+            df.columns = self.get_unique_column_names(df)
+            tool.Web.send_webui_data(data=df.to_csv(index=False), data_key="csv_data", event="csv_data")
 
         self.report({"INFO"}, f"Data is exported to {props.format.upper()}.")
         return {"FINISHED"}
 
+    def get_unique_column_names(self, dataframe: pd.DataFrame) -> list[str]:
+        count = Counter()
+        return [
+            f"{col}.{i:03d}" if duped and not count.update([col]) and (i := count[col]) else col
+            for col, duped in zip(dataframe.columns, dataframe.columns.duplicated())
+        ]
+
 
 class ImportIfcCsv(bpy.types.Operator):
     bl_idname = "bim.import_ifccsv"