thyra 3.0.0__tar.gz → 3.0.2__tar.gz

{thyra-3.0.0 → thyra-3.0.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: thyra
-Version: 3.0.0
+Version: 3.0.2
 Summary: A modern Python library for converting Mass Spectrometry Imaging (MSI) data into SpatialData/Zarr format - your portal to spatial omics
 License: MIT
 License-File: LICENSE
@@ -42,8 +42,8 @@ Requires-Dist: tqdm (>=4.50.0)
 Requires-Dist: zarr (>=3.1.4,<3.2)
 Project-URL: Bug Tracker, https://github.com/M4i-Imaging-Mass-Spectrometry/thyra/issues
 Project-URL: Changelog, https://github.com/M4i-Imaging-Mass-Spectrometry/thyra/blob/main/CHANGELOG.md
-Project-URL: Contributing, https://github.com/M4i-Imaging-Mass-Spectrometry/thyra/blob/main/CONTRIBUTING.md
-Project-URL: Documentation, https://github.com/M4i-Imaging-Mass-Spectrometry/thyra#readme
+Project-URL: Contributing, https://M4i-Imaging-Mass-Spectrometry.github.io/thyra/contributing/
+Project-URL: Documentation, https://M4i-Imaging-Mass-Spectrometry.github.io/thyra/
 Project-URL: Discussions, https://github.com/M4i-Imaging-Mass-Spectrometry/thyra/discussions
 Project-URL: Homepage, https://github.com/M4i-Imaging-Mass-Spectrometry/thyra
 Project-URL: Repository, https://github.com/M4i-Imaging-Mass-Spectrometry/thyra
@@ -83,7 +83,7 @@ for a full walkthrough, including the published example dataset
 - **Memory Efficient**: Handles large datasets (100+ GB) through streaming processing
 - **Optical Alignment**: Automatic MSI-to-optical image registration for Bruker data
 - **Multi-Region Support**: Handles slides with multiple tissue sections
-- **Resampling**: Physics-aware mass axis resampling (enabled by default)
+- **Resampling**: Physics-aware mass axis resampling (on by default in the CLI; opt-in from the Python API)
 - **3D Support**: Process volume data or treat as 2D slices
 - **Cross-Platform**: Windows, macOS, and Linux
 

{thyra-3.0.0 → thyra-3.0.2}/README.md

@@ -32,7 +32,7 @@ for a full walkthrough, including the published example dataset
 - **Memory Efficient**: Handles large datasets (100+ GB) through streaming processing
 - **Optical Alignment**: Automatic MSI-to-optical image registration for Bruker data
 - **Multi-Region Support**: Handles slides with multiple tissue sections
-- **Resampling**: Physics-aware mass axis resampling (enabled by default)
+- **Resampling**: Physics-aware mass axis resampling (on by default in the CLI; opt-in from the Python API)
 - **3D Support**: Process volume data or treat as 2D slices
 - **Cross-Platform**: Windows, macOS, and Linux
 

{thyra-3.0.0 → thyra-3.0.2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "poetry.core.masonry.api"
 
 [tool.poetry]
 name = "thyra"
-version = "3.0.0"
+version = "3.0.2"
 description = "A modern Python library for converting Mass Spectrometry Imaging (MSI) data into SpatialData/Zarr format - your portal to spatial omics"
 authors = ["Theodoros Visvikis <t.visvikis@maastrichtuniversity.nl>"]
 maintainers = ["Theodoros Visvikis <t.visvikis@maastrichtuniversity.nl>"]
@@ -12,7 +12,7 @@ license = "MIT"
 readme = "README.md"
 homepage = "https://github.com/M4i-Imaging-Mass-Spectrometry/thyra"
 repository = "https://github.com/M4i-Imaging-Mass-Spectrometry/thyra"
-documentation = "https://github.com/M4i-Imaging-Mass-Spectrometry/thyra#readme"
+documentation = "https://M4i-Imaging-Mass-Spectrometry.github.io/thyra/"
 keywords = [
     "mass-spectrometry",
     "imaging",
@@ -50,7 +50,9 @@ packages = [{include = "thyra"}]
 "Bug Tracker" = "https://github.com/M4i-Imaging-Mass-Spectrometry/thyra/issues"
 "Discussions" = "https://github.com/M4i-Imaging-Mass-Spectrometry/thyra/discussions"
 "Changelog" = "https://github.com/M4i-Imaging-Mass-Spectrometry/thyra/blob/main/CHANGELOG.md"
-"Contributing" = "https://github.com/M4i-Imaging-Mass-Spectrometry/thyra/blob/main/CONTRIBUTING.md"
+# There is no CONTRIBUTING.md at the repository root; the guide lives at
+# docs/contributing.md and is published as this page.
+"Contributing" = "https://M4i-Imaging-Mass-Spectrometry.github.io/thyra/contributing/"
 
 [tool.poetry.dependencies]
 python = ">=3.12, <3.14"

{thyra-3.0.0 → thyra-3.0.2}/thyra/__init__.py

@@ -36,7 +36,7 @@ warnings.filterwarnings(
     category=FutureWarning,
 )
 
-__version__ = "3.0.0"
+__version__ = "3.0.2"
 
 # Import key components - avoid wildcard imports
 try:

{thyra-3.0.0 → thyra-3.0.2}/thyra/converters/spatialdata/spatialdata_3d_converter.py

@@ -272,12 +272,17 @@ class SpatialData3DConverter(BaseSpatialDataConverter):
         n_x, n_y, n_z = self._dimensions
 
         if n_z > 1:
-            # True 3D TIC image
+            # True 3D TIC image. The accumulator is indexed [y, x, z] (see
+            # _create_data_structures), while Image3DModel declares
+            # (c, z, y, x) -- so the axes have to be *moved*. This used to
+            # relabel the shape instead, which reshapes nothing and left
+            # every volume whose extents were not all equal permuted: a
+            # 5(x) x 3(y) x 2(z) acquisition was stored as (1, 3, 5, 2).
             tic_values = data_structures["tic_values"]
-            z_size, y_size, x_size = tic_values.shape
+            tic_values_zyx = np.transpose(tic_values, (2, 0, 1))
 
             # Add channel dimension for 3D image
-            tic_values_with_channel = tic_values.reshape(1, z_size, y_size, x_size)
+            tic_values_with_channel = tic_values_zyx[np.newaxis, ...]
 
             # The image is intrinsically in raster pixel indices.
             # Scale into physical micrometers so "global" agrees with

{thyra-3.0.0 → thyra-3.0.2}/thyra/converters/spatialdata/streaming_converter.py

@@ -461,6 +461,7 @@ class StreamingSpatialDataConverter(BaseSpatialDataConverter):
         nnz_per_row = np.zeros(n_rows, dtype=np.int64)
         total_nnz = 0
         pixel_count = 0
+        n_out_of_bounds = 0
 
         region_map, region_total, region_count = self._init_region_accumulators(n_cols)
 
@@ -476,11 +477,22 @@ class StreamingSpatialDataConverter(BaseSpatialDataConverter):
                 mz_indices, resampled_ints = self._process_spectrum(mzs, intensities)
                 nnz = len(mz_indices)
 
-                nnz_per_row[pixel_idx] = nnz
-                total_nnz += nnz
-
+                # The row count and the TIC are both indexed by grid
+                # position, so both need the bounds check: a coordinate
+                # outside the declared dimensions would otherwise wrap
+                # round and land on an unrelated pixel. Unlike the PCS
+                # pre-scan the guard sits outside the `nnz > 0` test,
+                # because this assignment is unconditional -- a wrapped
+                # empty spectrum would overwrite a real row's count with
+                # zero. total_nnz travels with nnz_per_row so that
+                # indptr[-1] keeps matching the array size derived from
+                # it; pass 2 skips the same spectra.
                 if 0 <= y < n_y and 0 <= x < n_x:
+                    nnz_per_row[pixel_idx] = nnz
+                    total_nnz += nnz
                     tic_values[y, x] = float(np.sum(resampled_ints))
+                else:
+                    n_out_of_bounds += 1
 
                 if nnz > 0:
                     np.add.at(total_intensity, mz_indices, resampled_ints)
@@ -500,6 +512,15 @@ class StreamingSpatialDataConverter(BaseSpatialDataConverter):
                 pbar.update(1)
 
         logger.info(f"Pass 1 complete: {total_nnz:,} total non-zeros")
+        if n_out_of_bounds:
+            logger.warning(
+                "%d spectra sat outside the declared %dx%d grid and were "
+                "skipped. Previously they were written to a wrapped-round "
+                "row index, silently overwriting an unrelated pixel.",
+                n_out_of_bounds,
+                n_x,
+                n_y,
+            )
 
         # Build indptr from nnz counts
         indptr = np.zeros(n_rows + 1, dtype=np.int64)
@@ -628,7 +649,11 @@ class StreamingSpatialDataConverter(BaseSpatialDataConverter):
                 mz_indices, resampled_ints = self._process_spectrum(mzs, intensities)
                 nnz = len(mz_indices)
 
-                if nnz > 0:
+                # Same bounds check as pass 1, and it has to be the same
+                # one: that pass reserved no room for a spectrum outside
+                # the grid, so writing it here would overrun the row it
+                # wrapped onto and displace that row's own entries.
+                if nnz > 0 and 0 <= y < n_y and 0 <= x < n_x:
                     pos = write_pos[pixel_idx]
                     positions = np.arange(pos, pos + nnz)
                     buf_positions.append(positions)