Created metadata to store masknames for zarr stores (#134)

* Created metadata to store masknames for zarr stores * Adjusted version and manifest
hubmapconsortium · Sep 16, 2024 · 1202495 · 1202495
1 parent 7bb1807
commit 1202495
Show file tree

Hide file tree

Showing 6 changed files with 31 additions and 3 deletions.
diff --git a/containers/seg-mudata-to-zarr/VERSION b/containers/seg-mudata-to-zarr/VERSION
@@ -1 +1 @@
-0.0.1
+0.0.2
diff --git a/containers/seg-mudata-to-zarr/context/create_zarr_stores.py b/containers/seg-mudata-to-zarr/context/create_zarr_stores.py
@@ -1,6 +1,7 @@
 import mudata as md
 import pandas as pd
 import numpy as np
+import json
 
 VAR_CHUNK_SIZE = 10
 index_mapping = None
@@ -86,5 +87,14 @@ def create_zarr_for_masks(mdata, output_path):
             zarr_store_path = f'{output_path}/{mask_name}.zarr'
             mask_data.write_zarr(zarr_store_path, chunks=chunks)
             print(f'Created Zarr store for the mask: {mask_name}')
+            write_masknames_to_metadata(mask_names, output_path)
     except Exception as e:
         print(f'Error in getting mask names from mudata {str(e)}')
+
+
+def write_masknames_to_metadata(mask_names, output_path):
+    if isinstance(mask_names, pd.Categorical):
+        mask_names = mask_names.categories.tolist()
+    data = {'mask_names': mask_names}
+    with open(f'{output_path}/metadata.json', 'w') as file:
+        json.dump(data, file, indent=4)
diff --git a/containers/seg-mudata-to-zarr/test-output-expected/secondary_analysis.zarr/metadata.json b/containers/seg-mudata-to-zarr/test-output-expected/secondary_analysis.zarr/metadata.json
@@ -0,0 +1,9 @@
+{
+    "mask_names": [
+        "mask_0",
+        "mask_1",
+        "mask_2",
+        "mask_3",
+        "mask_4"
+    ]
+}
diff --git a/seg-mudata-to-zarr-manifest.json b/seg-mudata-to-zarr-manifest.json
@@ -1,6 +1,6 @@
 [
   {
-    "pattern": "seg-to-mudata-zarr/multiome_downstream_tfidf.zarr/(.*)\\.zgroup",
+    "pattern": "seg-to-mudata-zarr/secondary_analysis.zarr/(.*)/(.*)\\.zgroup",
     "description": "MuData Zarr store for storing annotations for segmentation masks",
     "edam_ontology_term": "EDAM_1.24.format_2333"
   }

diff --git a/seg-mudata-to-zarr.cwl b/seg-mudata-to-zarr.cwl
@@ -6,7 +6,7 @@ class: CommandLineTool
 baseCommand: ['python', '/main.py', '--output_dir', './seg-to-mudata-zarr', '--input_dir']
 hints:
   DockerRequirement:
-    dockerPull: hubmap/portal-container-seg-mudata-to-zarr:0.0.1
+    dockerPull: hubmap/portal-container-seg-mudata-to-zarr:0.0.2
 inputs:
   input_directory:
     type: Directory

diff --git a/...ata-to-zarr/test-output-expected/seg-to-mudata-zarr/secondary_analysis.zarr/metadata.json b/...ata-to-zarr/test-output-expected/seg-to-mudata-zarr/secondary_analysis.zarr/metadata.json
@@ -0,0 +1,9 @@
+{
+    "mask_names": [
+        "mask_0",
+        "mask_1",
+        "mask_2",
+        "mask_3",
+        "mask_4"
+    ]
+}