Name conflict in kfactory.kcell::name

[oradwastaken]

Why is it that kfactory still raises this ERROR | kfactory.kcell:name:687 - Name conflict in kfactory.kcell::name.

Even if I'm using rename_duplicated_cells=True?

It seems like everything is compiling correctly, but the cells have $1 appended to them. I find I can strategically add gf.clear_cache() in certain locations to avoid this warning (and the appended $ index) but that sometimes breaks the workflow.

Below is my MWE:

import tempfile
from pathlib import Path

import gdsfactory as gf

c = gf.Component('top')

c << gf.c.rectangle(size=(100, 100), layer=(1, 0))

with tempfile.TemporaryDirectory() as tmp_dir:
    temp_file = Path(tmp_dir) / "temp.oas"

    path = c.write_gds(temp_file)

    # Perform some external operation on the temp_file, and then:
    new_component = gf.import_gds(path, rename_duplicated_cells=True)

Running this yields:

2026-03-11 16:02:06.814 | ERROR    | kfactory.kcell:name:687 - Name conflict in kfactory.kcell::name at line 687
Renaming Unnamed_3 (cell_index=3) to top would cause it to be named the same as:
 - top (cell_index=0), function_name=None, basename=None

[sebastian-goeldi]

because it creates two cells with the same name in the same namespace (internal objects of kfactory). You can ignore this as long as not both get instantiated at the same time

[oradwastaken]

I understand that it's not a problem, but it's certainly noisy. I would expect that enabling the rename_duplicated_cells=True flag would disable these warnings.

[sebastian-goeldi]

Internally it's still naming the cells as duplicate as it seems though. That might be a bug. I'll look into it.