formatting + minor display fix in tutorials

docs/tutorials/comseg.ipynb

@@ -35,7 +35,7 @@
     "\n",
     "First, follow the original [CLI tutorial](https://gustaveroussy.github.io/sopa/tutorials/cli_usage/) until you finished the \"Cellpose segmentation\" section, and then, continue below.\n",
     "\n",
-    "####  1. Save a ComSeg config file as a `config.json` file\n",
+    "####  1. Save a ComSeg config file as a config.json file\n",
     "\n",
     "We display below a minimal example of a ComSeg `config.json` file\n",
     "\n",

docs/tutorials/xenium_explorer/explorer.ipynb

@@ -18,8 +18,6 @@
    "outputs": [],
    "source": [
     "import sopa\n",
-    "import sopa.io\n",
-    "import sopa.segmentation\n",
     "import spatialdata"
    ]
   },
@@ -175,7 +173,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Use the coordinates of a lasso selection in `SpatialData`\n",
+    "## Use the coordinates of a lasso selection in SpatialData\n",
     "\n",
     "On the Xenium Explorer, you can use the Lasso or Rectangular selection tools to select some regions of interest. Then, you'll be able to analyze back this region of interest using `spatialdata`.\n",
     "\n",
@@ -252,7 +250,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "### Cropping a `SpatialData` object from a selection\n",
+    "### Cropping a SpatialData object from a selection\n",
     "\n",
     "You can also export the whole selection as a polygon and use it to crop the `spatialdata` object. For that, click on \"Download Selection Coordinates as CSV\", as below. It will create a file called `\"Selection_1_coordinates.csv\"`.\n",
     "\n",

sopa/cli/app.py

@@ -178,7 +178,7 @@ def aggregate(
 ):
     """Create an `anndata` table containing the transcript count and/or the channel intensities per cell"""
     from sopa.io.standardize import read_zarr_standardized
-    from sopa.segmentation.aggregate import Aggregator
+    from sopa.segmentation import Aggregator
 
     sdata = read_zarr_standardized(sdata_path, warn=True)
 

sopa/segmentation/__init__.py

@@ -1,5 +1,5 @@
-from . import shapes, aggregate, methods, stainings
+from . import aggregation, shapes, methods, stainings
 from ..patches import Patches2D
-from .aggregate import Aggregator, overlay_segmentation
+from .aggregation import Aggregator, overlay_segmentation
 from .stainings import StainingSegmentation
 from .tissue import tissue_segmentation

sopa/segmentation/aggregate.py → sopa/segmentation/aggregation.py

@@ -73,6 +73,24 @@ def overlay_segmentation(
     )
 
 
+def aggregate(
+    sdata: SpatialData,
+    average_intensities: bool = True,
+    expand_radius_ratio: float = 0,
+    min_transcripts: int = 0,
+    min_intensity_ratio: float = 0,
+    **kwargs: int,
+):
+    aggr = Aggregator(sdata, **kwargs)
+
+    aggr.compute_table(
+        average_intensities=average_intensities,
+        expand_radius_ratio=expand_radius_ratio,
+        min_transcripts=min_transcripts,
+        min_intensity_ratio=min_intensity_ratio,
+    )
+
+
 class Aggregator:
     """Perform transcript count and channel averaging over a `SpatialData` object"""
 
@@ -151,9 +169,9 @@ def overlay_segmentation(
         self.geo_df = pd.concat([geo_df_cropped, geo_df], join="outer", axis=0)
         self.geo_df.attrs = old_geo_df.attrs
 
-        self.standardized_table(save_table=save_table)
+        self.add_standardized_table(save_table=save_table)
 
-    def standardized_table(self, save_table: bool = True):
+    def add_standardized_table(self, save_table: bool = True):
         self.table.obs_names = list(map(str_cell_id, range(self.table.n_obs)))
 
         self.geo_df.index = list(self.table.obs_names)
@@ -269,7 +287,7 @@ def compute_table(
             SopaKeys.UNS_HAS_INTENSITIES: average_intensities,
         }
 
-        self.standardized_table(save_table=save_table)
+        self.add_standardized_table(save_table=save_table)
 
 
 def _overlap_area_ratio(row) -> float:

sopa/segmentation/transcripts.py

@@ -17,7 +17,7 @@
 
 from .._constants import SopaKeys
 from .._sdata import get_spatial_element, get_spatial_image
-from . import aggregate, shapes
+from . import aggregation, shapes
 
 log = logging.getLogger(__name__)
 
@@ -58,7 +58,7 @@ def resolve(
         geo_df_new = ShapesModel.parse(geo_df_new, transformations=transformations)
 
         log.info("Aggregating transcripts on merged cells")
-        table_conflicts = aggregate.count_transcripts(sdata, gene_column, geo_df=geo_df_new)
+        table_conflicts = aggregation.count_transcripts(sdata, gene_column, geo_df=geo_df_new)
         table_conflicts.obs_names = new_ids
         table_conflicts = [table_conflicts]
 

sopa/utils/data.py

@@ -158,7 +158,7 @@ def uniform(
 
 
 def _add_table(sdata: SpatialData):
-    from ..segmentation.aggregate import Aggregator
+    from ..segmentation import Aggregator
 
     aggregator = Aggregator(sdata, shapes_key=SopaKeys.CELLPOSE_BOUNDARIES)
 

tests/test_aggregation.py

@@ -6,7 +6,7 @@
 import xarray as xr
 from shapely.geometry import Polygon, box
 
-from sopa.segmentation import aggregate
+from sopa.segmentation import aggregation
 
 dask.config.set({"dataframe.query-planning": False})
 import dask.dataframe as dd  # noqa
@@ -23,7 +23,7 @@ def test_average_channels_aligned():
     # One cell is on the first block, one is overlapping on both blocks, and one is on the last block
     cells = [box(x, y, x + cell_size - 1, y + cell_size - 1) for x, y in cell_start]
 
-    means = aggregate._average_channels_aligned(xarr, cells)
+    means = aggregation._average_channels_aligned(xarr, cells)
 
     true_means = np.stack([image[:, y : y + cell_size, x : x + cell_size].mean(axis=(1, 2)) for x, y in cell_start])
 
@@ -50,7 +50,7 @@ def test_count_transcripts():
 
     gdf = gpd.GeoDataFrame(geometry=polygons)
 
-    adata = aggregate._count_transcripts_aligned(gdf, points, "gene")
+    adata = aggregation._count_transcripts_aligned(gdf, points, "gene")
     expected = np.array([[0, 3, 1], [1, 0, 1], [1, 3, 1]])
 
     assert (adata.X.toarray() == expected).all()