concatenate({'a': sdata1, 'b': sdata2}) renames the coordinate systems of the input objects in place

[!NOTE]
This whole message is AI-generated. The issue was automatically discovered and reported by an AI agent (Claude) during an autonomous bug hunt on the spatialdata code base. It has not been verified or triaged by a human yet; the needs: triage label is set so that a maintainer can confirm it. The reproduction script below was executed by the agent in an isolated environment (see Environment) and its output is pasted verbatim.

Summary

A pure combine operation has a hidden side effect on its inputs: subsequent queries/plots on sdata1 with target_coordinate_system="global" fail or silently return nothing.

Severity (agent's assessment): medium/high — after the call sdata1 maps to global-a instead of global; if backed, write_transformations() persists the renamed systems

Where: src/spatialdata/_core/concatenate.py::_fix_ensure_unique_element_names (remove_transformation(el, remove_all=True) / set_transformation(el, t, new_cs) on the original element objects; only tables are optionally copied via modify_tables_inplace)

Expected behaviour

Inputs unchanged (or the in-place behaviour documented and opt-in, as for tables).

Reproduction

Save as repro.py and run uv run repro.py (the PEP 723 header pins spatialdata to the commit the bug was found on; replace the URL fragment with @main to test the current main branch).

# /// script
# requires-python = ">=3.12"
# dependencies = [
#     "spatialdata @ git+https://github.com/scverse/spatialdata.git@ccf1ea048d054b6624214bf618008a9f9ae223e0",
# ]
# ///
"""concatenate({'a': sdata1, 'b': sdata2}) renames the coordinate systems of the INPUT objects in place."""
import warnings
import numpy as np
from spatialdata import SpatialData, concatenate
from spatialdata.models import Image2DModel
from spatialdata.transformations import get_transformation

warnings.simplefilter("ignore")
s1 = SpatialData(images={"img": Image2DModel.parse(np.zeros((1, 4, 4), dtype=np.uint8))})
s2 = SpatialData(images={"img": Image2DModel.parse(np.zeros((1, 4, 4), dtype=np.uint8))})
print("before: s1 coordinate systems:", s1.coordinate_systems)
merged = concatenate({"a": s1, "b": s2})
print("after : s1 coordinate systems:", s1.coordinate_systems, "| s1['img'] transformations:", list(get_transformation(s1["img"], get_all=True)))
print("merged coordinate systems:", sorted(merged.coordinate_systems))
bug = s1.coordinate_systems != ["global"]
print("VERDICT:", "BUG REPRODUCED (input mutated)" if bug else "NOT REPRODUCED")

Observed output

before: s1 coordinate systems: ['global']
after : s1 coordinate systems: ['global-a'] | s1['img'] transformations: ['global-a']
merged coordinate systems: ['global-a', 'global-b']
VERDICT: BUG REPRODUCED (input mutated)

Possible fix direction (unverified)

Shallow-copy each element before rewriting its transformations (xarray/geopandas .copy(); dask frame .copy() plus a copy of the transform dict; for DataTree copy the tree and per-scale attrs), mirroring modify_tables_inplace=False.

Environment

uv run repro.py with the PEP 723 metadata in the script (fresh, isolated environment; spatialdata built from main @ ccf1ea0 (2026-08-28); Python 3.13, latest releases of the dependencies at run time: pandas 3.0, anndata 0.13, zarr 3.3, dask 2026.8, numpy 2.5, geopandas 1.1, shapely 2.1). macOS (arm64). Also reproduced in a second environment with pandas 2.3.3 / anndata 0.12.11 / numpy 2.4.4 / zarr 3.2.1.

Possibly related issues

#998

Automatically generated; discovered by an AI agent (Claude) and not yet reviewed by a human.

Start with repro.py and src/spatialdata/_core/concatenate.py::_fix_ensure_unique_element_names, then trace how transformations are removed and assigned to input elements. Reproduce the mutation with uv run repro.py and ensure concatenate leaves sdata1 and sdata2 unchanged while the merged object still has unique coordinate systems.