thyra 3.7.2__tar.gz → 3.7.3__tar.gz

{thyra-3.7.2 → thyra-3.7.3}/.gitignore

@@ -53,6 +53,8 @@ Archive/
 # ...except documentation notebooks, which are published as part of the docs
 # site and must be committed with their outputs.
 !docs/*.ipynb
+# ...and the shared workflow notebooks (e.g. the Colab validation workflow).
+!notebooks/*.ipynb
 
 # Temporary files
 temp_*

{thyra-3.7.2 → thyra-3.7.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: thyra
-Version: 3.7.2
+Version: 3.7.3
 Summary: A modern Python library for converting Mass Spectrometry Imaging (MSI) data into SpatialData/Zarr format - your portal to spatial omics
 Project-URL: Homepage, https://github.com/M4i-Imaging-Mass-Spectrometry/thyra
 Project-URL: Repository, https://github.com/M4i-Imaging-Mass-Spectrometry/thyra
@@ -60,6 +60,7 @@ Description-Content-Type: text/markdown
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 [![Code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)
 [![Docs](https://img.shields.io/badge/docs-mkdocs-blue)](https://M4i-Imaging-Mass-Spectrometry.github.io/thyra)
+[![Open in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/M4i-Imaging-Mass-Spectrometry/thyra/blob/main/notebooks/Thyra_Validation_Workflow.ipynb)
 
 **Thyra** (from Greek thyra, meaning "door" or "portal") -- a modern Python library for converting Mass Spectrometry Imaging (MSI) data into the standardized **SpatialData/Zarr format**, serving as your portal to spatial omics analysis workflows.
 

{thyra-3.7.2 → thyra-3.7.3}/README.md

@@ -8,6 +8,7 @@
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 [![Code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)
 [![Docs](https://img.shields.io/badge/docs-mkdocs-blue)](https://M4i-Imaging-Mass-Spectrometry.github.io/thyra)
+[![Open in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/M4i-Imaging-Mass-Spectrometry/thyra/blob/main/notebooks/Thyra_Validation_Workflow.ipynb)
 
 **Thyra** (from Greek thyra, meaning "door" or "portal") -- a modern Python library for converting Mass Spectrometry Imaging (MSI) data into the standardized **SpatialData/Zarr format**, serving as your portal to spatial omics analysis workflows.
 

{thyra-3.7.2 → thyra-3.7.3}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "thyra"
-version = "3.7.2"
+version = "3.7.3"
 description = "A modern Python library for converting Mass Spectrometry Imaging (MSI) data into SpatialData/Zarr format - your portal to spatial omics"
 authors = [{ name = "Theodoros Visvikis", email = "t.visvikis@maastrichtuniversity.nl" }]
 maintainers = [{ name = "Theodoros Visvikis", email = "t.visvikis@maastrichtuniversity.nl" }]

{thyra-3.7.2 → thyra-3.7.3}/thyra/__init__.py

@@ -36,7 +36,7 @@ warnings.filterwarnings(
     category=FutureWarning,
 )
 
-__version__ = "3.7.2"
+__version__ = "3.7.3"
 
 # Import key components - avoid wildcard imports
 try:

{thyra-3.7.2 → thyra-3.7.3}/thyra/converters/spatialdata/base_spatialdata_converter.py

@@ -46,6 +46,49 @@ def _suppress_upstream_warnings():
         yield
 
 
+def _numeric_only(value: Any) -> bool:
+    """True when a list round-trips through AnnData/zarr as a numeric array."""
+    if isinstance(value, list):
+        return all(_numeric_only(item) for item in value)
+    if isinstance(value, np.ndarray):
+        return value.dtype.kind in "iufb"
+    return isinstance(value, (bool, int, float, np.integer, np.floating, np.bool_))
+
+
+def _json_fallback(obj: Any) -> Any:
+    """Last-resort encoder for values ``json.dumps`` does not know."""
+    if isinstance(obj, np.ndarray):
+        return obj.tolist()
+    if isinstance(obj, np.generic):
+        return obj.item()
+    return str(obj)
+
+
+def _jsonify_string_lists(obj: Any) -> Any:
+    """Replace any list that is not purely numeric with its JSON encoding.
+
+    AnnData/zarr cannot round-trip such lists: a list of dicts is
+    stringified entry by entry into Python ``repr`` strings, and any list
+    of strings comes back as a numpy string array -- whose ``deepcopy``
+    segfaults outright on numpy 2.1-2.2 (numpy#28609). Since every table
+    copy deepcopies ``uns`` (``AnnData.copy``, spatial queries, joins),
+    one such array in ``uns`` kills the reader's process with no
+    traceback. JSON side-steps both: it stores a single scalar string,
+    and hands consumers the actual structure back through
+    ``json.loads`` instead of ``repr`` output.
+
+    Purely numeric lists (nested included) are kept: they become plain
+    numeric arrays, which are safe and more useful as arrays.
+    """
+    if isinstance(obj, dict):
+        return {key: _jsonify_string_lists(value) for key, value in obj.items()}
+    if isinstance(obj, list):
+        if _numeric_only(obj):
+            return obj
+        return json.dumps(obj, default=_json_fallback)
+    return obj
+
+
 def _resolve_config_enum(raw: Any, by_name: Dict[str, Any], key: str) -> Any:
     """Resolve one resampling-config value to its enum member.
 
@@ -792,15 +835,24 @@ class BaseSpatialDataConverter(BaseMSIConverter, ABC):
         }
 
     def _collect_optional_sections(self, uns: Dict[str, Any], comp_meta: Any) -> None:
-        """Add the vendor sections the reader actually populated."""
+        """Add the vendor sections the reader actually populated.
+
+        Lists that are not purely numeric (imzML ``cvParams``, or any
+        string list a vendor extractor reports) are stored as JSON
+        strings -- see :func:`_jsonify_string_lists` for why letting
+        them reach the writer as lists corrupts them and crashes
+        readers on numpy 2.1-2.2.
+        """
         for key in ("format_specific", "acquisition_params", "instrument_info"):
             value = getattr(comp_meta, key, None)
             if value:
-                uns[key] = self._serialize_for_zarr(value)
+                uns[key] = _jsonify_string_lists(self._serialize_for_zarr(value))
 
         raw_metadata = getattr(comp_meta, "raw_metadata", None)
         if raw_metadata:
-            uns["raw_metadata"] = self._serialize_for_zarr(raw_metadata)
+            uns["raw_metadata"] = _jsonify_string_lists(
+                self._serialize_for_zarr(raw_metadata)
+            )
 
     def _collect_region_info(self, uns: Dict[str, Any]) -> None:
         """Add the acquisition region summary as JSON.

{thyra-3.7.2 → thyra-3.7.3}/thyra/metadata/__init__.py

@@ -53,6 +53,9 @@ from .extractors import (
 # Core data types
 from .types import ComprehensiveMetadata, EssentialMetadata
 
+# Read-side compatibility for stores written by earlier versions
+from .uns_compat import sanitize_uns_string_arrays
+
 # Base classes - import delayed to avoid circular imports
 
 
@@ -67,6 +70,7 @@ __all__ = [
     # Utility functions
     "get_extractor_for_format",
     "list_supported_formats",
+    "sanitize_uns_string_arrays",
 ]
 
 # Version information for the metadata system

thyra-3.7.3/thyra/metadata/uns_compat.py

@@ -0,0 +1,57 @@
+# thyra/metadata/uns_compat.py
+"""Read-side defense for ``uns`` blocks written by earlier Thyra versions.
+
+Stores written before the JSON rule in
+``BaseSpatialDataConverter._collect_optional_sections`` carry string
+lists in ``uns`` (notably ``raw_metadata["cvParams"]``) that materialize
+as numpy ``StringDType`` arrays when read back through AnnData/zarr.
+``copy.deepcopy`` of such an array segfaults the process on numpy
+2.1.x-2.2.x (numpy#28609, fixed in 2.3.0) -- and every table copy
+deepcopies ``uns``: ``AnnData.copy``, ``spatialdata.polygon_query``,
+``bounding_box_query``, and the join paths behind them.
+
+New stores no longer contain these arrays. This module exists for
+already-written stores read in an environment pinned to the affected
+numpy range.
+"""
+
+from collections.abc import Mapping
+from typing import Any, Dict
+
+import numpy as np
+
+__all__ = ["sanitize_uns_string_arrays"]
+
+
+def sanitize_uns_string_arrays(uns: Mapping[str, Any]) -> Dict[str, Any]:
+    """Return ``uns`` with numpy string arrays replaced by plain lists.
+
+    Recurses through nested dicts and lists; every other value is kept
+    as-is, so the result is safe to assign straight back::
+
+        import spatialdata as sd
+        from thyra.metadata import sanitize_uns_string_arrays
+
+        sdata = sd.read_zarr("store.zarr")
+        for table in sdata.tables.values():
+            table.uns = sanitize_uns_string_arrays(table.uns)
+
+    Needed only for stores written by Thyra versions that stored string
+    lists in ``uns`` as arrays, read under numpy 2.1.x-2.2.x, where
+    deepcopying a ``StringDType`` array kills the process (see module
+    docstring). Calling it anywhere else is harmless.
+    """
+    return {key: _sanitize(value) for key, value in uns.items()}
+
+
+def _sanitize(value: Any) -> Any:
+    if isinstance(value, Mapping):
+        return {key: _sanitize(item) for key, item in value.items()}
+    if isinstance(value, list):
+        return [_sanitize(item) for item in value]
+    # ``kind == "T"`` is StringDType, the one dtype whose deepcopy is
+    # broken. On numpy builds without StringDType no array has this
+    # kind, so the walk is a no-op there.
+    if isinstance(value, np.ndarray) and value.dtype.kind == "T":
+        return value.tolist()
+    return value