Development

.pre-commit-config.yaml

@@ -2,8 +2,8 @@ fail_fast: false
 default_language_version:
   python: python3
 default_stages:
-  - commit
-  - push
+  - pre-commit
+  - pre-push
 minimum_pre_commit_version: 2.16.0
 repos:
   - repo: https://github.com/pre-commit/mirrors-prettier

docs/index.rst

@@ -63,8 +63,11 @@ Documentation
    pages/example_notebooks.ipynb
 
 .. toctree::
-   :maxdepth: 2
+   :maxdepth: 4
    :includehidden:
    :caption: Module API
 
-   pages/module
+   pages/module/pipeline
+   pages/module/processing
+   pages/module/io
+   pages/module/tools

docs/pages/module/pipeline.rst

@@ -3,71 +3,11 @@ pipeline
 *******************
 
 .. toctree::
-   :maxdepth: 3
-
-project
-#######
-
-Within scPortrait, all operations are centered around the concept of a ``Project``. A ``Project`` is a python class which manages all of the scPortrait processing steps and is the central element through which all operations are performed. Each ``Project`` directly maps to a directory on the file system which contains all of the inputs to a specific scPortrait run as well as the generated outputs. Depending on the structure of the data that is to be processed a different Project class is required. Please see :ref:`here <projects>` for more information.
-
-Project
-=========
-.. autoclass:: scportrait.pipeline.project.Project
-    :members:
-    :show-inheritance:
-
-segmentation
-#############
-
-Segmentation
-==============
-.. autoclass:: scportrait.pipeline.segmentation.Segmentation
-    :members:
-    :show-inheritance:
-
-ShardedSegmentation
-=====================
-.. autoclass:: scportrait.pipeline.segmentation.ShardedSegmentation
-    :members:
-    :show-inheritance:
-
-segmentation workflows
-######################
-.. automodule:: scportrait.pipeline.segmentation.workflows
-    :members:
-
-extraction
-###########
-
-HDF5CellExtraction
-===================
-.. autoclass:: scportrait.pipeline.extraction.HDF5CellExtraction
-    :members:
-    :show-inheritance:
-
-
-Featurization
-#############
-
-MLClusterClassifier
-===================
-.. autoclass:: scportrait.pipeline.featurization.MLClusterClassifier
-    :members:
-
-    .. automethod:: __call__
-
-CellFeaturizer
-==============
-.. autoclass:: scportrait.pipeline.featurization.CellFeaturizer
-    :members:
-
-    .. automethod:: __call__
-
-selection
-###########
-
-LMDSelection
-==============
-.. autoclass:: scportrait.pipeline.selection.LMDSelection
-    :members:
-    :show-inheritance:
+    :maxdepth: 2
+
+    pipeline/project
+    pipeline/segmentation
+    pipeline/segmentation_workflows
+    pipeline/extraction
+    pipeline/featurization
+    pipeline/selection

docs/pages/module/pipeline/extraction.rst

@@ -0,0 +1,8 @@
+extraction
+###########
+
+HDF5CellExtraction
+===================
+.. autoclass:: scportrait.pipeline.extraction.HDF5CellExtraction
+    :members:
+    :show-inheritance:

docs/pages/module/pipeline/featurization.rst

@@ -0,0 +1,16 @@
+Featurization
+#############
+
+MLClusterClassifier
+===================
+.. autoclass:: scportrait.pipeline.featurization.MLClusterClassifier
+    :members:
+
+    .. automethod:: __call__
+
+CellFeaturizer
+==============
+.. autoclass:: scportrait.pipeline.featurization.CellFeaturizer
+    :members:
+
+    .. automethod:: __call__

docs/pages/module/pipeline/project.rst

@@ -0,0 +1,2 @@
+.. automodule:: scportrait.pipeline.project
+    :members:

docs/pages/module/pipeline/segmentation.rst

@@ -0,0 +1,14 @@
+segmentation
+#############
+
+Segmentation
+==============
+.. autoclass:: scportrait.pipeline.segmentation.Segmentation
+    :members:
+    :show-inheritance:
+
+ShardedSegmentation
+=====================
+.. autoclass:: scportrait.pipeline.segmentation.ShardedSegmentation
+    :members:
+    :show-inheritance:

docs/pages/module/pipeline/segmentation_workflows.rst

@@ -0,0 +1,5 @@
+segmentation workflows
+######################
+
+.. automodule:: scportrait.pipeline.segmentation.workflows
+    :members:

docs/pages/module/pipeline/selection.rst

@@ -0,0 +1,8 @@
+selection
+###########
+
+LMDSelection
+==============
+.. autoclass:: scportrait.pipeline.selection.LMDSelection
+    :members:
+    :show-inheritance:

docs/pages/module/tools.rst

@@ -3,7 +3,7 @@ tools
 *******************
 
 .. toctree::
-   :maxdepth: 4
+   :maxdepth: 2
 
    tools/parsing
    tools/stitching

src/scportrait/pipeline/_utils/sdata_io.py

@@ -9,7 +9,7 @@
 import xarray
 from alphabase.io import tempmmap
 from spatialdata import SpatialData
-from spatialdata.models import PointsModel, TableModel
+from spatialdata.models import Image2DModel, PointsModel, TableModel
 from spatialdata.transformations.transformations import Identity
 
 from scportrait.pipeline._base import Logable
@@ -19,7 +19,8 @@
     get_chunk_size,
 )
 
-ChunkSize: TypeAlias = tuple[int, int]
+ChunkSize2D: TypeAlias = tuple[int, int]
+ChunkSize3D: TypeAlias = tuple[int, int, int]
 ObjectType: TypeAlias = Literal["images", "labels", "points", "tables"]
 
 
@@ -143,6 +144,58 @@ def _get_input_image(self, sdata: SpatialData) -> xarray.DataArray:
         return input_image
 
     ## write elements to sdata object
+    def _write_image_sdata(
+        self,
+        image,
+        image_name: str,
+        channel_names: list[str] = None,
+        scale_factors: list[int] = None,
+        chunks: ChunkSize3D = (1, 1000, 1000),
+        overwrite=False,
+    ):
+        """
+        Write the supplied image to the spatialdata object.
+
+        Args:
+            image (dask.array): Image to be written to the spatialdata object.
+            image_name (str): Name of the image to be written to the spatialdata object.
+            channel_names list[str]: List of channel names for the image. Default is None.
+            scale_factors list[int]: List of scale factors for the image. Default is [2, 4, 8]. This will load the image at 4 different resolutions to allow for fluid visualization.
+            chunks (tuple): Chunk size for the image. Default is (1, 1000, 1000).
+            overwrite (bool): Whether to overwrite existing data. Default is False.
+        """
+
+        if scale_factors is None:
+            scale_factors = [2, 4, 8]
+        if scale_factors is None:
+            scale_factors = [2, 4, 8]
+
+        _sdata = self._read_sdata()
+
+        if channel_names is None:
+            channel_names = [f"channel_{i}" for i in range(image.shape[0])]
+
+        # transform to spatialdata image model
+        transform_original = Identity()
+        image = Image2DModel.parse(
+            image,
+            dims=["c", "y", "x"],
+            chunks=chunks,
+            c_coords=channel_names,
+            scale_factors=scale_factors,
+            transformations={"global": transform_original},
+            rgb=False,
+        )
+
+        if overwrite:
+            self._force_delete_object(_sdata, image_name, "images")
+
+        _sdata.images[image_name] = image
+        _sdata.write_element(image_name, overwrite=True)
+
+        self.log(f"Image {image_name} written to sdata object.")
+        self._check_sdata_status()
+
     def _write_segmentation_object_sdata(
         self,
         segmentation_object: spLabels2DModel,
@@ -177,7 +230,7 @@ def _write_segmentation_sdata(
         segmentation: xarray.DataArray | np.ndarray,
         segmentation_label: str,
         classes: set[str] | None = None,
-        chunks: ChunkSize = (1000, 1000),
+        chunks: ChunkSize2D = (1000, 1000),
         overwrite: bool = False,
     ) -> None:
         """Write segmentation data to SpatialData.
@@ -268,6 +321,7 @@ def _add_centers(self, segmentation_label: str, overwrite: bool = False) -> None
         centroids_object = self._get_centers(_sdata, segmentation_label)
         self._write_points_object_sdata(centroids_object, self.centers_name, overwrite=overwrite)
 
+    ## load elements from sdata to a memory mapped array
     def _load_input_image_to_memmap(
         self, tmp_dir_abs_path: str | Path, image: np.typing.NDArray[Any] | None = None
     ) -> str:

src/scportrait/pipeline/extraction.py

@@ -842,8 +842,10 @@ def process(self, partial=False, n_cells=None, seed=42):
         self.log("Loading input images to memory mapped arrays...")
         start_data_transfer = timeit.default_timer()
 
-        self.path_seg_masks = self.project._load_seg_to_memmap(seg_name=self.masks, tmp_dir_abs_path=self._tmp_dir_path)
-        self.path_image_data = self.project._load_input_image_to_memmap(tmp_dir_abs_path=self._tmp_dir_path)
+        self.path_seg_masks = self.filehandler._load_seg_to_memmap(
+            seg_name=self.masks, tmp_dir_abs_path=self._tmp_dir_path
+        )
+        self.path_image_data = self.filehandler._load_input_image_to_memmap(tmp_dir_abs_path=self._tmp_dir_path)
 
         stop_data_transfer = timeit.default_timer()
         time_data_transfer = stop_data_transfer - start_data_transfer