spatialdata 0.7.3a2__tar.gz → 0.8.0__tar.gz

spatialdata-0.8.0/.claude/skills/memray/SKILL.md

@@ -0,0 +1,37 @@
+---
+name: memray
+description: Profile the memory usage of a Python script using memray and visualize a temporal flamegraph in the browser. Use when the user wants to investigate memory consumption, find leaks, or understand allocation patterns.
+compatibility: Requires the pixi profiling environment (pixi run -e profiling). Supports Linux and macOS.
+allowed-tools: Bash(pixi run -e profiling memray-run:*) Bash(pixi run -e profiling memray-flame:*) Bash(open:*) Bash(xdg-open:*) Bash(python -m webbrowser:*)
+---
+
+## Steps
+
+1. Ask the user which script to profile (full or relative path).
+
+2. Run the script under memray:
+
+    ```bash
+    pixi run -e profiling memray-run script.py
+    ```
+
+    This produces a binary file named `memray-script.py.<pid>.bin` in the current directory.
+
+3. Generate the flamegraph HTML report from the `.bin` file:
+
+    ```bash
+    pixi run -e profiling memray-flame memray-script.py.<pid>.bin
+    ```
+
+    Replace `<pid>` with the actual PID shown in the filename. This writes `memray-flamegraph-script.py.<pid>.html`.
+
+4. Open the report in the browser:
+    - macOS: `open memray-flamegraph-script.py.<pid>.html`
+    - Linux: `xdg-open memray-flamegraph-script.py.<pid>.html`
+    - Either: `python -m webbrowser memray-flamegraph-script.py.<pid>.html`
+
+## Notes
+
+- The `--temporal` flag (included in `memray-flame`) shows memory over time, not just peak — use this to spot leaks and allocation bursts.
+- To find the `.bin` file if unsure of the name: `ls memray-*.bin`
+- To compare runs, save the previous report: `cp memray-flamegraph-script.py.<pid>.html memray-flamegraph-before.html`

spatialdata-0.8.0/.claude/skills/profimp/SKILL.md

@@ -0,0 +1,30 @@
+---
+name: profimp
+description: Profile Python import time using profimp and open a waterfall HTML report. Use when investigating slow startup or wanting to identify which imports are most expensive.
+compatibility: Requires profimp (available as a pixi dependency). macOS or Linux.
+allowed-tools: Bash(profimp:*) Bash(open:*) Bash(xdg-open:*) Bash(python -m webbrowser:*)
+---
+
+## Steps
+
+1. Ask what to profile. Suggest common patterns for this repo:
+    - `import spatialdata`
+    - `from spatialdata import SpatialData`
+    - `from spatialdata_io import xenium`
+
+2. Run:
+
+```bash
+profimp --html "<import_stmt>" > /tmp/profimp.html
+```
+
+3. Open the report:
+    - macOS: `open /tmp/profimp.html`
+    - Linux: `xdg-open /tmp/profimp.html`
+    - Either: `python -m webbrowser /tmp/profimp.html`
+
+## Notes
+
+- The report is a waterfall chart showing every sub-import and its timing.
+- To compare before/after: `cp /tmp/profimp.html /tmp/profimp-before.html` before re-running.
+- `pixi run python -m profimp` also works if `profimp` is not on PATH.

spatialdata-0.8.0/.claude/skills/pyspy/SKILL.md

@@ -0,0 +1,45 @@
+---
+name: pyspy
+description: Profile the execution time of a Python script using py-spy and visualize the result with speedscope. Use when the user wants to benchmark performance, find slow code paths, or profile CPU time.
+compatibility: Requires the pixi profiling environment (pixi run -e profiling). Speedscope must be installed separately (npm install -g speedscope). sudo is required on macOS.
+allowed-tools: Bash(pixi run -e profiling pyspy:*) Bash(pixi run -e profiling speedscope:*) Bash(sudo pixi run -e profiling pyspy:*)
+---
+
+## Steps
+
+1. Ask the user which script to profile (full or relative path).
+
+2. Run py-spy to record the profile. The output is always written to `profile.speedscope.json` in the current directory.
+
+    **Linux** (no sudo needed):
+
+    ```bash
+    pixi run -e profiling pyspy script.py
+    ```
+
+    **macOS** (sudo required — py-spy needs to attach to the process):
+
+    ```bash
+    sudo pixi run -e profiling pyspy script.py
+    ```
+
+    If sudo fails to find the pixi environment, use absolute paths:
+
+    ```bash
+    sudo /path/to/.pixi/envs/profiling/bin/py-spy record --gil \
+      -o profile.speedscope.json --format speedscope \
+      -- /path/to/.pixi/envs/profiling/bin/python script.py
+    ```
+
+3. Open the result in speedscope:
+    ```bash
+    pixi run -e profiling speedscope
+    ```
+    This opens `profile.speedscope.json` in the browser via the local speedscope CLI.
+
+## Notes
+
+- If the speedscope view is blank, switch threads using the thread selector in the top-right corner.
+- To save a profile before overwriting: `cp profile.speedscope.json profile-before.speedscope.json`
+- `--gil` records only time when the GIL is held (Python-level CPU time). Drop it to include C extension time.
+- speedscope must be installed globally: `npm install -g speedscope`

{spatialdata-0.7.3a2 → spatialdata-0.8.0}/.github/workflows/test.yaml

@@ -19,11 +19,12 @@ jobs:
             fail-fast: false
             matrix:
                 include:
-                    - {os: windows-latest, python: "3.11", dask-version: "2025.12.0", name: "min dask"}
+                    - {os: windows-latest, python: "3.12", dask-version: "2026.3.0", name: "min dask"}
                     - {os: windows-latest, python: "3.14", dask-version: "latest"}
-                    - {os: ubuntu-latest, python: "3.11", dask-version: "latest"}
+                    - {os: ubuntu-latest, python: "3.12", dask-version: "latest"}
+                    - {os: ubuntu-latest, python: "3.13", dask-version: "latest"}
                     - {os: ubuntu-latest, python: "3.14", dask-version: "latest"}
-                    - {os: macos-latest, python: "3.11", dask-version: "latest"}
+                    - {os: macos-latest, python: "3.12", dask-version: "latest"}
                     - {os: macos-latest, python: "3.14", prerelease: "allow", name: "prerelease"}
         env:
             OS: ${{ matrix.os }}
@@ -41,7 +42,6 @@ jobs:
             - name: Install dependencies
               run: |
                   if [[ "${PRERELEASE}" == "allow" ]]; then
-                    sed -i '' 's/requires-python.*//' pyproject.toml # otherwise uv complains that anndata requires python>=3.12 and we only do >=3.11 😱
                     uv add git+https://github.com/scverse/anndata.git
                     uv add pandas>=3.dev0
                   fi

{spatialdata-0.7.3a2 → spatialdata-0.8.0}/.gitignore

@@ -46,11 +46,18 @@ spatialdata-sandbox
 # version file
 _version.py
 
-# other
-node_modules/
+# agents configurations
+.claude/settings.local.json
 
+# benchmarking and profiling
 .asv/
+profile.speedscope.json
+
+# other
+node_modules/
 
 .mypy_cache
 .ruff_cache
 uv.lock
+pixi.lock
+

{spatialdata-0.7.3a2 → spatialdata-0.8.0}/.mypy.ini

@@ -1,5 +1,5 @@
 [mypy]
-python_version = 3.11
+python_version = 3.12
 
 ignore_errors = False
 warn_redundant_casts = True

{spatialdata-0.7.3a2 → spatialdata-0.8.0}/.pre-commit-config.yaml

@@ -9,18 +9,18 @@ ci:
     skip: []
 repos:
     - repo: https://github.com/rbubley/mirrors-prettier
-      rev: v3.8.3
+      rev: v3.8.4
       hooks:
           - id: prettier
             exclude: ^.github/workflows/test.yaml
     - repo: https://github.com/pre-commit/mirrors-mypy
-      rev: v1.20.2
+      rev: v2.1.0
       hooks:
           - id: mypy
             additional_dependencies: [numpy, types-requests]
             exclude: tests/|docs/
     - repo: https://github.com/astral-sh/ruff-pre-commit
-      rev: v0.15.12
+      rev: v0.15.18
       hooks:
           - id: ruff
             args: [--fix, --exit-non-zero-on-fix]

{spatialdata-0.7.3a2 → spatialdata-0.8.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: spatialdata
-Version: 0.7.3a2
+Version: 0.8.0
 Summary: Spatial data format.
 Project-URL: Documentation, https://spatialdata.scverse.org/en/latest
 Project-URL: Source, https://github.com/scverse/spatialdata.git
@@ -37,21 +37,21 @@ License: BSD 3-Clause License
         OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
         OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 License-File: LICENSE
-Requires-Python: >=3.11
+Requires-Python: >=3.12
 Requires-Dist: anndata>=0.9.1
 Requires-Dist: annsel>=0.1.2
 Requires-Dist: click
 Requires-Dist: dask-image
-Requires-Dist: dask<2026.1.2,>=2025.12.0
+Requires-Dist: dask>=2026.3.0
 Requires-Dist: datashader
-Requires-Dist: distributed<2026.1.2
+Requires-Dist: distributed>=2026.3.0
 Requires-Dist: fsspec[http,s3]
 Requires-Dist: geopandas>=0.14
 Requires-Dist: multiscale-spatial-image==2.0.3
 Requires-Dist: networkx
 Requires-Dist: numba>=0.55.0
 Requires-Dist: numpy
-Requires-Dist: ome-zarr>=0.14.0
+Requires-Dist: ome-zarr>=0.16.0
 Requires-Dist: pandas
 Requires-Dist: pooch
 Requires-Dist: pyarrow
@@ -82,7 +82,7 @@ Description-Content-Type: text/markdown
 [![pre-commit.ci status](https://results.pre-commit.ci/badge/github/scverse/spatialdata/main.svg)](https://results.pre-commit.ci/latest/github/scverse/spatialdata/main)
 [![codecov](https://codecov.io/gh/scverse/spatialdata/branch/main/graph/badge.svg?token=X19DRSIMCU)](https://codecov.io/gh/scverse/spatialdata)
 [![documentation badge](https://readthedocs.org/projects/scverse-spatialdata/badge/?version=latest)](https://spatialdata.scverse.org/en/latest/)
-[![DOI](https://zenodo.org/badge/487366481.svg)](https://zenodo.org/badge/latestdoi/487366481)
+[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.20056407.svg)](https://doi.org/10.5281/zenodo.20056407)
 [![Downloads](https://static.pepy.tech/badge/spatialdata)](https://pepy.tech/project/spatialdata)
 [![Release](https://github.com/scverse/spatialdata/actions/workflows/release.yaml/badge.svg?event=release)](https://github.com/scverse/spatialdata/actions/workflows/release.yaml)
 [![Documentation][badge-pypi]][link-pypi]
@@ -91,11 +91,15 @@ Description-Content-Type: text/markdown
 [badge-pypi]: https://badge.fury.io/py/spatialdata.svg
 [link-pypi]: https://pypi.org/project/spatialdata/
 
-SpatialData is a data framework that comprises a FAIR storage format and a collection of python libraries for performant access, alignment, and processing of uni- and multi-modal spatial omics datasets. This repository contains the core spatialdata library. See the links below to learn more about other packages in the SpatialData ecosystem.
+SpatialData is a data framework, a schema and a serialization format for uni- and multi-modal spatial omics datasets. It is supported by a growing suite of software libraries for performant access, manipulation and analysis in multiple languages, including Python, R and JavaScript.
+
+This repository contains the core spatialdata library. See the links below to learn more about other packages in the SpatialData ecosystem.
 
 - [spatialdata-io](https://github.com/scverse/spatialdata-io): load data from common spatial omics technologies into spatialdata.
 - [spatialdata-plot](https://github.com/scverse/spatialdata-plot): Static plotting library for spatialdata.
 - [napari-spatialdata](https://github.com/scverse/napari-spatialdata): napari plugin for interactive exploration and annotation of spatial data.
+- [SpatialData (R)](https://github.com/HelenaLC/SpatialData): R implementation of the SpatialData framework.
+- [SpatialData.js](https://github.com/Taylor-CCB-Group/SpatialData.js): JavaScript/TypeScript implementation of the SpatialData framework.
 
 [//]: # "numfocus-fiscal-sponsor-attribution"
 

{spatialdata-0.7.3a2 → spatialdata-0.8.0}/README.md

@@ -6,7 +6,7 @@
 [![pre-commit.ci status](https://results.pre-commit.ci/badge/github/scverse/spatialdata/main.svg)](https://results.pre-commit.ci/latest/github/scverse/spatialdata/main)
 [![codecov](https://codecov.io/gh/scverse/spatialdata/branch/main/graph/badge.svg?token=X19DRSIMCU)](https://codecov.io/gh/scverse/spatialdata)
 [![documentation badge](https://readthedocs.org/projects/scverse-spatialdata/badge/?version=latest)](https://spatialdata.scverse.org/en/latest/)
-[![DOI](https://zenodo.org/badge/487366481.svg)](https://zenodo.org/badge/latestdoi/487366481)
+[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.20056407.svg)](https://doi.org/10.5281/zenodo.20056407)
 [![Downloads](https://static.pepy.tech/badge/spatialdata)](https://pepy.tech/project/spatialdata)
 [![Release](https://github.com/scverse/spatialdata/actions/workflows/release.yaml/badge.svg?event=release)](https://github.com/scverse/spatialdata/actions/workflows/release.yaml)
 [![Documentation][badge-pypi]][link-pypi]
@@ -15,11 +15,15 @@
 [badge-pypi]: https://badge.fury.io/py/spatialdata.svg
 [link-pypi]: https://pypi.org/project/spatialdata/
 
-SpatialData is a data framework that comprises a FAIR storage format and a collection of python libraries for performant access, alignment, and processing of uni- and multi-modal spatial omics datasets. This repository contains the core spatialdata library. See the links below to learn more about other packages in the SpatialData ecosystem.
+SpatialData is a data framework, a schema and a serialization format for uni- and multi-modal spatial omics datasets. It is supported by a growing suite of software libraries for performant access, manipulation and analysis in multiple languages, including Python, R and JavaScript.
+
+This repository contains the core spatialdata library. See the links below to learn more about other packages in the SpatialData ecosystem.
 
 - [spatialdata-io](https://github.com/scverse/spatialdata-io): load data from common spatial omics technologies into spatialdata.
 - [spatialdata-plot](https://github.com/scverse/spatialdata-plot): Static plotting library for spatialdata.
 - [napari-spatialdata](https://github.com/scverse/napari-spatialdata): napari plugin for interactive exploration and annotation of spatial data.
+- [SpatialData (R)](https://github.com/HelenaLC/SpatialData): R implementation of the SpatialData framework.
+- [SpatialData.js](https://github.com/Taylor-CCB-Group/SpatialData.js): JavaScript/TypeScript implementation of the SpatialData framework.
 
 [//]: # "numfocus-fiscal-sponsor-attribution"
 

{spatialdata-0.7.3a2 → spatialdata-0.8.0}/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '0.7.3a2'
-__version_tuple__ = version_tuple = (0, 7, 3, 'a2')
+__version__ = version = '0.8.0'
+__version_tuple__ = version_tuple = (0, 8, 0)
 
 __commit_id__ = commit_id = None

{spatialdata-0.7.3a2 → spatialdata-0.8.0}/benchmarks/README.md

@@ -65,6 +65,32 @@ git checkout - && git stash pop
 asv compare main HEAD
 ```
 
+### Dataloader benchmarks
+
+Dataloader benchmarks live in `benchmarks/benchmark_dataloader.py`. They use a synthetic in-memory `SpatialData` (2048×2048 image, 500 circle regions) and compute two metrics:
+
+- `time_init` — constructing `ImageTilesDataset` (includes bounding-box pre-computation).
+- `time_fetch` — iterating over all 500 tiles once (pure `__getitem__` calls, no `DataLoader` overhead).
+
+Run both in your current environment:
+
+```bash
+asv run --python=same --show-stderr -b TimeDataloader
+```
+
+Or a single method:
+
+```bash
+asv run --python=same --show-stderr -b TimeDataloader.time_init
+asv run --python=same --show-stderr -b TimeDataloader.time_fetch
+```
+
+Compare against `main` in one shot:
+
+```bash
+asv continuous --show-stderr -v -b TimeDataloader main HEAD
+```
+
 ### Querying benchmarks
 
 Querying using a bounding box without a spatial index is highly impacted by large amounts of points (transcripts), more than table rows (cells).

spatialdata-0.8.0/benchmarks/benchmark_dataloader.py

@@ -0,0 +1,75 @@
+# type: ignore
+"""Benchmarks for ImageTilesDataset: init time and iteration time."""
+
+from __future__ import annotations
+
+import anndata as ad
+import geopandas as gpd
+import numpy as np
+import pandas as pd
+from shapely.geometry import Point
+
+import spatialdata as sd
+from spatialdata.dataloader import ImageTilesDataset
+from spatialdata.models import Image2DModel, ShapesModel, TableModel
+from spatialdata.transformations import Identity
+
+_RNG = np.random.default_rng(42)
+
+_IMAGE_SIZE = 2048
+_N_CIRCLES = 500
+_N_CHANNELS = 3
+
+_DATASET_KWARGS = {
+    "regions_to_images": {"circles": "image"},
+    "regions_to_coordinate_systems": {"circles": "global"},
+    "table_name": "table",
+    "return_annotations": "instance_id",
+}
+
+
+def _make_sdata() -> sd.SpatialData:
+    img_data = _RNG.integers(0, 256, size=(_N_CHANNELS, _IMAGE_SIZE, _IMAGE_SIZE), dtype=np.uint8).astype(np.float32)
+    image = Image2DModel.parse(img_data, dims=["c", "y", "x"], transformations={"global": Identity()})
+
+    radius = 32.0
+    cx = _RNG.uniform(radius, _IMAGE_SIZE - radius, size=_N_CIRCLES)
+    cy = _RNG.uniform(radius, _IMAGE_SIZE - radius, size=_N_CIRCLES)
+    geom = gpd.GeoDataFrame({"geometry": [Point(x, y) for x, y in zip(cx, cy, strict=True)]})
+    geom["radius"] = radius
+    circles = ShapesModel.parse(geom, transformations={"global": Identity()})
+
+    table = ad.AnnData(
+        _RNG.random((_N_CIRCLES, 10)).astype(np.float32),
+        obs=pd.DataFrame(
+            {
+                "region": pd.Categorical(["circles"] * _N_CIRCLES),
+                "instance_id": np.arange(_N_CIRCLES, dtype=np.int64),
+            },
+            index=[str(i) for i in range(_N_CIRCLES)],
+        ),
+    )
+    table = TableModel.parse(table, region="circles", region_key="region", instance_key="instance_id")
+
+    return sd.SpatialData(images={"image": image}, shapes={"circles": circles}, tables={"table": table})
+
+
+class TimeDataloader:
+    """Time ImageTilesDataset construction and tile iteration."""
+
+    def setup(self):
+        self.sdata = _make_sdata()
+        self.ds = ImageTilesDataset(sdata=self.sdata, **_DATASET_KWARGS)
+
+    def teardown(self):
+        del self.ds
+        del self.sdata
+
+    def time_init(self):
+        """Time constructing ImageTilesDataset (bounding-box pre-computation)."""
+        ImageTilesDataset(sdata=self.sdata, **_DATASET_KWARGS)
+
+    def time_fetch(self):
+        """Time iterating over every tile once."""
+        for i in range(len(self.ds)):
+            _ = self.ds[i]

{spatialdata-0.7.3a2 → spatialdata-0.8.0}/docs/api/models_utils.md

@@ -4,6 +4,7 @@
 .. currentmodule:: spatialdata.models
 
 .. autofunction:: get_model
+.. autofunction:: validate_element
 .. autodata:: SpatialElement
 .. autofunction:: get_axes_names
 .. autofunction:: get_spatial_axes

{spatialdata-0.7.3a2 → spatialdata-0.8.0}/docs/index.md

@@ -6,11 +6,15 @@
 
 # An open and universal framework for processing spatial omics data.
 
-SpatialData is a data framework that comprises a FAIR storage format and a collection of python libraries for performant access, alignment, and processing of uni- and multi-modal spatial omics datasets. This page provides documentation on how to install, use, and extend the core `spatialdata` library. See the links below to learn more about other packages in the SpatialData ecosystem.
+SpatialData is a data framework, a schema and a serialization format for uni- and multi-modal spatial omics datasets. It is supported by a growing suite of software libraries for performant access, manipulation and analysis in multiple languages, including Python, R and JavaScript.
+
+This page provides documentation on how to install, use, and extend the core `spatialdata` library. See the links below to learn more about other packages in the SpatialData ecosystem.
 
 - `spatialdata-io`: load data from common spatial omics technologies into `spatialdata` ([repository][spatialdata-io-repo], [documentation][spatialdata-io-docs]).
 - `spatialdata-plot`: Static plotting library for `spatialdata` ([repository][spatialdata-plot-repo], [documentation][spatialdata-plot-docs]).
 - `napari-spatialdata-repo`: napari plugin for interactive exploration and annotation of `spatialdata` ([repository][napari-spatialdata-repo], [documentation][napari-spatialdata-docs]).
+- `SpatialData` (R): R implementation of the SpatialData framework ([repository][spatialdata-r-repo]).
+- `SpatialData.js`: JavaScript/TypeScript implementation of the SpatialData framework ([repository][spatialdata-js-repo]).
 
 Please see our publication {cite}`marconatoSpatialDataOpenUniversal2024` for citation and to learn more.
 
@@ -119,3 +123,5 @@ references.md
 [napari-spatialdata-docs]: https://spatialdata.scverse.org/projects/napari/en/stable/notebooks/spatialdata.html
 [spatialdata-io-docs]: https://spatialdata.scverse.org/projects/io/en/stable/
 [spatialdata-plot-docs]: https://spatialdata.scverse.org/projects/plot/en/stable/api.html
+[spatialdata-r-repo]: https://github.com/HelenaLC/SpatialData
+[spatialdata-js-repo]: https://github.com/Taylor-CCB-Group/SpatialData.js

{spatialdata-0.7.3a2 → spatialdata-0.8.0}/pyproject.toml

@@ -15,7 +15,7 @@ maintainers = [
 urls.Documentation = "https://spatialdata.scverse.org/en/latest"
 urls.Source = "https://github.com/scverse/spatialdata.git"
 urls.Home-page = "https://github.com/scverse/spatialdata.git"
-requires-python = ">=3.11"
+requires-python = ">=3.12"
 dynamic= [
   "version" # allow version to be set by git tags
 ]
@@ -26,8 +26,8 @@ dependencies = [
     "annsel>=0.1.2",
     "click",
     "dask-image",
-    "dask>=2025.12.0,<2026.1.2",
-    "distributed<2026.1.2",
+    "dask>=2026.3.0",
+    "distributed>=2026.3.0",
     "datashader",
     "fsspec[s3,http]",
     "geopandas>=0.14",
@@ -35,7 +35,7 @@ dependencies = [
     "networkx",
     "numba>=0.55.0",
     "numpy",
-    "ome_zarr>=0.14.0",
+    "ome_zarr>=0.16.0",
     "pandas",
     "pooch",
     "pyarrow",
@@ -145,7 +145,7 @@ exclude = [
 
 ]
 line-length = 120
-target-version = "py311"
+target-version = "py312"
 
 [tool.ruff.lint]
 ignore = [
@@ -210,3 +210,37 @@ convention = "numpy"
 
 # pyupgrade typing rewrite TODO: remove at some point from per-file ignore
 # "UP006", "UP007"
+
+[tool.pixi.workspace]
+channels = ["conda-forge"]
+platforms = ["osx-arm64", "linux-64"]
+
+[tool.pixi.dependencies]
+proj = "*"
+
+[tool.pixi.tasks]
+test = "pytest"
+test-parallel = "pytest -n auto --dist worksteal"
+pre-commit = "pre-commit run"
+pre-commit-all = "pre-commit run --all-files"
+
+[tool.pixi.feature.profiling.dependencies]
+py-spy = "*"
+
+[tool.pixi.feature.profiling.pypi-dependencies]
+spatialdata = { path = ".", editable = true }
+profimp = "*"
+memray = "*"
+
+[tool.pixi.feature.profiling.tasks]
+# Usage: pixi run -e profiling pyspy script.py  (prefix with sudo on macOS)
+pyspy = "py-spy record --gil -o profile.speedscope.json --format speedscope -- python"
+# Usage: pixi run -e profiling speedscope
+speedscope = "npx --yes speedscope profile.speedscope.json"
+# Usage: pixi run -e profiling memray-run script.py
+memray-run = "memray run"
+# Usage: pixi run -e profiling memray-flame memray-script.py.<pid>.bin
+memray-flame = "memray flamegraph --temporal"
+
+[tool.pixi.environments]
+profiling = { features = ["profiling"], solve-group = "default" }

{spatialdata-0.7.3a2 → spatialdata-0.8.0}/src/spatialdata/_core/_utils.py

@@ -158,9 +158,9 @@ def sanitize_table(data: AnnData, inplace: bool = True) -> AnnData | None:
     # Handle other attributes
     for attr in ("obsm", "obsp", "varm", "varp", "uns", "layers"):
         d = getattr(sanitized, attr)
-        new_keys = {old: get_unique_name(old, attr) for old in d}
-        # Create new dictionary with sanitized keys
-        new_dict = {new_keys[old]: value for old, value in d.items()}
+        # None is a valid key in layers (anndata >= 0.13: represents X); skip sanitizing it
+        new_keys = {old: get_unique_name(old, attr) for old in d if old is not None}
+        new_dict = {(new_keys[old] if old is not None else old): value for old, value in d.items()}
         setattr(sanitized, attr, new_dict)
 
     return None if inplace else sanitized

{spatialdata-0.7.3a2 → spatialdata-0.8.0}/src/spatialdata/_core/concatenate.py

@@ -219,14 +219,6 @@ def concatenate(
     return sdata
 
 
-def _filter_table_in_coordinate_systems(table: AnnData, coordinate_systems: list[str]) -> AnnData:
-    table_mapping_metadata = table.uns[TableModel.ATTRS_KEY]
-    region_key = table_mapping_metadata[TableModel.REGION_KEY_KEY]
-    new_table = table[table.obs[region_key].isin(coordinate_systems)].copy()
-    new_table.uns[TableModel.ATTRS_KEY][TableModel.REGION_KEY] = new_table.obs[region_key].unique().tolist()
-    return new_table
-
-
 def _fix_ensure_unique_element_names(
     sdatas: dict[str, SpatialData],
     rename_tables: bool,

{spatialdata-0.7.3a2 → spatialdata-0.8.0}/src/spatialdata/_core/operations/rasterize_bins.py

@@ -87,14 +87,10 @@ def rasterize_bins(
     table = sdata.tables[table_name]
     if not isinstance(element, GeoDataFrame | DaskDataFrame | DataArray):
         raise ValueError("The bins should be a GeoDataFrame, a DaskDataFrame or a DataArray.")
-    if isinstance(element, DataArray):
-        if "c" in element.dims:
-            raise ValueError(
-                "If bins is a DataArray, it should hold labels; found a image element instead, with"
-                f" 'c': {element.dims}."
-            )
-        if not np.issubdtype(element.dtype, np.integer):
-            raise ValueError(f"If bins is a DataArray, it should hold integers. Found dtype {element.dtype}.")
+    if isinstance(element, DataArray) and "c" in element.dims:
+        raise ValueError(
+            f"If bins is a DataArray, it should hold labels; found a image element instead, with 'c': {element.dims}."
+        )
 
     _, region_key, instance_key = get_table_keys(table)
     if not table.obs[region_key].dtype == "category":

{spatialdata-0.7.3a2 → spatialdata-0.8.0}/src/spatialdata/_core/operations/transform.py

@@ -6,10 +6,11 @@ import warnings
 from functools import singledispatch
 from typing import TYPE_CHECKING, Any, cast
 
+import dask
 import dask.array as da
+import dask.dataframe as dd
 import dask_image.ndinterp
 import numpy as np
-import pandas as pd
 from dask.array.core import Array as DaskArray
 from dask.dataframe import DataFrame as DaskDataFrame
 from geopandas import GeoDataFrame
@@ -442,20 +443,26 @@ def _(
     axes = get_axes_names(data)
     arrays = []
 
-    # Workaround to prevent partition collaps and missing dependency problem for now.
+    # Dask's expression optimizer can collapse partitions at compute time, making the partition
+    # structure inside vs. outside a disable_dask_tune_optimization() context inconsistent. To avoid
+    # index-alignment failures (e.g. "cannot reindex on an axis with duplicate labels" from parquet
+    # files that each start their index at 0) and length-mismatch errors, we materialise the non-axis
+    # columns and compute the axis arrays inside a single context where the partition structure is
+    # stable, then do plain pandas operations and re-wrap with dd.from_delayed (not dd.from_pandas,
+    # which sorts by index and would scramble rows for non-monotonic or duplicate indices).
     with disable_dask_tune_optimization() if data.npartitions > 1 else contextlib.nullcontext():
+        lengths = [len(part) for part in data.partitions]
         for ax in axes:
             # TODO We have to pass on the lengths explicitly as automatic determination with dask graph optimization
-            #  leads to collaps of the partitions. However this causes a missing dependency problem, which for now is
+            #  leads to collapse of the partitions. However this causes a missing dependency problem, which for now is
             #  prevented by setting the optimization to False when performing this operation.
-            arrays.append(data[ax].to_dask_array(lengths=[len(part) for part in data.partitions]).reshape(-1, 1))
+            arrays.append(data[ax].to_dask_array(lengths=lengths).reshape(-1, 1))
 
-    xdata = DataArray(da.concatenate(arrays, axis=1), coords={"points": range(len(data)), "dim": list(axes)})
-    xtransformed = transformation._transform_coordinates(xdata)
-    transformed = data.drop(columns=list(axes)).copy()
-    # dummy transformation that will be replaced by _adjust_transformation()
-    default_cs = {DEFAULT_COORDINATE_SYSTEM: Identity()}
-    transformed.attrs[TRANSFORM_KEY] = default_cs
+        xdata = DataArray(da.concatenate(arrays, axis=1), coords={"points": range(sum(lengths)), "dim": list(axes)})
+        xtransformed = transformation._transform_coordinates(xdata)
+
+        # Compute non-axis columns while the partition structure is still stable; preserves original index.
+        transformed_pd = data.drop(columns=list(axes)).compute()
 
     for ax in axes:
         indices = xtransformed["dim"] == ax
@@ -463,8 +470,24 @@ def _(
         # TODO: discuss with dask team
         # This is not nice, but otherwise there is a problem with the joint graph of new_ax and transformed, causing
         # a getattr missing dependency of dependent from_dask_array.
-        new_col = pd.Series(new_ax.data.flatten().compute(), index=transformed.index)
-        transformed[ax] = new_col
+        # Assigning a numpy array is positional (no index alignment), so the original index is preserved.
+        transformed_pd[ax] = new_ax.data.flatten().compute()
+
+    # Reconstruct as a dask DataFrame via delayed partitions so that:
+    # (a) row order matches the original (dd.from_pandas sorts by index, which scrambles rows for
+    #     non-monotonic or duplicate indices such as those produced by multi-file parquet reads), and
+    # (b) the original index is preserved exactly.
+    offsets = np.cumsum([0] + lengths)
+    delayed_parts = [dask.delayed(transformed_pd.iloc[offsets[i] : offsets[i + 1]]) for i in range(len(lengths))]
+    transformed = dd.from_delayed(delayed_parts, meta=transformed_pd.iloc[:0])
+    # Preserve spatialdata_attrs (feature_key, instance_key, …) from the original element;
+    # dd.from_delayed starts with empty attrs so we must copy them explicitly.
+    for k, v in data.attrs.items():
+        if k != TRANSFORM_KEY:
+            transformed.attrs[k] = v
+    # dummy transformation that will be replaced by _adjust_transformation()
+    default_cs = {DEFAULT_COORDINATE_SYSTEM: Identity()}
+    transformed.attrs[TRANSFORM_KEY] = default_cs
 
     old_transformations = cast(dict[str, Any], get_transformation(data, get_all=True))