Fix renaming issue in vizgen loader and rename table in vizgen & cosmx loaders (#91)

* Fix renaming issue in vizgen loader

* Rename tables in vizgen and cosmx loaders

Author: LouisK92

src/datasets/loaders/bruker_cosmx/script.py

@@ -227,7 +227,7 @@
     "stitched_labels"    : "cell_labels",
     "points"             : "transcripts",
     "cells_polygons"     : "cell_boundaries",
-    "table"              : "metadata",
+    #"table"              : "metadata",
 }
 
 for old_key, new_key in elements_renaming_map.items():

src/datasets/loaders/bruker_cosmx_nsclc/script.py

@@ -170,7 +170,7 @@ def fixed_get_morphology_coords(images_dir: Path) -> list[str]:
     "stitched_labels"    : "cell_labels",
     "points"             : "transcripts",
     "cells_polygons"     : "cell_boundaries",
-    "table"              : "metadata",
+    #"table"              : "metadata",
 }
 
 for old_key, new_key in elements_renaming_map.items():

src/datasets/loaders/vizgen_merscope/script.py

@@ -188,7 +188,7 @@ def read_boundary_hdf5(folder):
     f"{name}_z3": "morphology_mip",  # TODO: that is actually not the morphology_mip, i.e. either we should rename the label later, or we should actually project over z. But we also want to have 3d at some point anyway
     f"{name}_transcripts": "transcripts",
     f"{name}_polygons": "cell_boundaries",
-    "table": "metadata",
+    #"table": "metadata",
 }
 
 for old_key, new_key in elements_renaming_map.items():
@@ -201,8 +201,10 @@ def read_boundary_hdf5(folder):
 sdata["transcripts"] = sdata["transcripts"].rename(columns={"global_z": "z", "transcript_id": "ensembl_id"})#, "gene": "feature_name"})
 if "gene" in sdata["transcripts"].columns: 
     # No idea why, but somehow dask dataframe renaming for the 'gene' column ends up in a key error when assigning it to sdata["transcripts"].
+    # update: see https://github.com/scverse/spatialdata/issues/996
     sdata["transcripts"]["feature_name"] = sdata["transcripts"]["gene"]
     del sdata["transcripts"]["gene"]
+    sdata['transcripts'].attrs["spatialdata_attrs"]["feature_key"] = "feature_name"
 print(datetime.now() - t0, "Renamed transcripts column 'global_z' -> 'z' and 'gene' -> 'feature_name' and 'transcript_id' -> 'ensembl_id'", flush=True)
 
 print(datetime.now() - t0, "Columns in sdata['transcripts']:", sdata["transcripts"].columns, flush=True)