Merge pull request #224 from berombau/macsima-reader

add macsima reader

Author: LucaMarconato

CHANGELOG.md

@@ -11,6 +11,7 @@ and this project adheres to [Semantic Versioning][].
 ## incoming release
 
 -   (Visium/Visium HD) lowres and hires images now mapped also to the 'global' coordinate system #230
+-   (Macsima) added support @berombau #224
 
 ## [0.1.6] - 2024-11-26
 

README.md

@@ -24,14 +24,15 @@ This package contains reader functions to load common spatial omics formats into
 -   Steinbock (output data)
 -   STOmics Stereo-seq®
 -   Vizgen MERSCOPE® (MERFISH)
+-   MACSima® (MACS® iQ View output)
 
 Note: all mentioned technologies are registered trademarks of their respective companies.
 
 ## Known limitations
 
-Contributions for addressing the below limitations are very welcomed. 
+Contributions for addressing the below limitations are very welcomed.
 
-- Only Stereo-seq 7.x is supported, 8.x is not currently supported. https://github.com/scverse/spatialdata-io/issues/161
+-   Only Stereo-seq 7.x is supported, 8.x is not currently supported. https://github.com/scverse/spatialdata-io/issues/161
 
 ### How to Contribute
 
@@ -41,7 +42,6 @@ Contributions for addressing the below limitations are very welcomed.
 
 3. **Optional Enhancements**: To facilitate reproducibility and ease of data access, consider adding a folder in the [spatialdata-sandbox](https://github.com/giovp/spatialdata-sandbox) repository. Include a `download.py` and `to_zarr.py` script (refer to examples in the repository) to enable others to reproduce your reader by simply running these scripts sequentially.
 
-
 ## Getting started
 
 Please refer to the [documentation][link-docs]. In particular, the

pyproject.toml

@@ -34,6 +34,7 @@ dependencies = [
     "pyarrow",
     "readfcs",
     "tifffile>=2023.8.12",
+    "ome-types",
 ]
 
 [project.optional-dependencies]

src/spatialdata_io/__init__.py

@@ -4,6 +4,7 @@
 from spatialdata_io.readers.cosmx import cosmx
 from spatialdata_io.readers.curio import curio
 from spatialdata_io.readers.dbit import dbit
+from spatialdata_io.readers.macsima import macsima
 from spatialdata_io.readers.mcmicro import mcmicro
 from spatialdata_io.readers.merscope import merscope
 from spatialdata_io.readers.seqfish import seqfish
@@ -32,6 +33,7 @@
     "xenium_explorer_selection",
     "dbit",
     "visium_hd",
+    "macsima",
 ]
 
 __version__ = version("spatialdata-io")

src/spatialdata_io/readers/_utils/_utils.py

@@ -7,6 +7,9 @@
 
 from anndata import AnnData, read_text
 from h5py import File
+from ome_types import from_tiff
+from ome_types.model import Pixels, UnitsLength
+from spatialdata._logging import logger
 
 from spatialdata_io.readers._utils._read_10x_h5 import _read_10x_h5
 
@@ -75,3 +78,48 @@ def _initialize_raster_models_kwargs(
     if "scale_factors" not in labels_models_kwargs:
         labels_models_kwargs["scale_factors"] = [2, 2, 2, 2]
     return image_models_kwargs, labels_models_kwargs
+
+
+def calc_scale_factors(lower_scale_limit: float, min_size: int = 1000, default_scale_factor: int = 2) -> list[int]:
+    """Calculate scale factors based on image size to get lowest resolution under min_size pixels."""
+    # get lowest dimension, ignoring channels
+    scale_factor: int = default_scale_factor
+    scale_factors = [scale_factor]
+    lower_scale_limit /= scale_factor
+    while lower_scale_limit >= min_size:
+        # scale_factors are cumulative, so we don't need to do e.g. scale_factor *= 2
+        scale_factors.append(scale_factor)
+        lower_scale_limit /= scale_factor
+    return scale_factors
+
+
+def parse_channels(path: Path) -> list[str]:
+    """Parse channel names from an OME-TIFF file."""
+    images = from_tiff(path).images
+    if len(images) > 1:
+        logger.warning("Found multiple images in OME-TIFF file. Only the first one will be used.")
+    channels = images[0].pixels.channels
+    logger.debug(channels)
+    names = [c.name for c in channels if c.name is not None]
+    return names
+
+
+def parse_physical_size(path: Path | None = None, ome_pixels: Pixels | None = None) -> float:
+    """Parse physical size from OME-TIFF to micrometer."""
+    pixels = ome_pixels or from_tiff(path).images[0].pixels
+    logger.debug(pixels)
+    if pixels.physical_size_x_unit != pixels.physical_size_y_unit:
+        logger.error("Physical units for x and y dimensions are not the same.")
+        raise NotImplementedError
+    if pixels.physical_size_x != pixels.physical_size_y:
+        logger.error("Physical sizes for x and y dimensions are not the same.")
+        raise NotImplementedError
+    # convert to micrometer if needed
+    if pixels.physical_size_x_unit == UnitsLength.NANOMETER:
+        physical_size = pixels.physical_size_x / 1000
+    elif pixels.physical_size_x_unit == UnitsLength.MICROMETER:
+        physical_size = pixels.physical_size_x
+    else:
+        logger.error(f"Physical unit not recognized: '{pixels.physical_size_x_unit}'.")
+        raise NotImplementedError
+    return float(physical_size)