loading metadata as well

merfish.py

@@ -1,5 +1,6 @@
 import datasets
 import os
+import glob
 import pandas as pd
 
 class MERFISHConfig(datasets.BuilderConfig):
@@ -9,7 +10,7 @@ class MERFISHConfig(datasets.BuilderConfig):
 class MERFISH(datasets.GeneratorBasedBuilder):
     BUILDER_CONFIGS = [
         MERFISHConfig(name="raw", description="Raw MERFISH data"),
-        MERFISHConfig(name="processed", description="Processed log2 MERFISH data"),
+        MERFISHConfig(name="processed", description="Processed MERFISH data"),
     ]
 
     def _info(self):
@@ -24,28 +25,36 @@ class MERFISH(datasets.GeneratorBasedBuilder):
         )
 
     def _split_generators(self, dl_manager):
-        base_path = os.path.join(self.config.name)
+        data_files = {
+            "expression": os.path.join(self.config.name, "expression", "*.parquet"),
+            "gene_metadata": os.path.join(self.config.name, "gene_metadata.parquet"),
+            "cell_metadata": os.path.join(self.config.name, "cell_metadata.parquet"),
+        }
+        downloaded = dl_manager.download(data_files)
         return [
-            datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"data_dir": base_path}),
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={
+                    "expression_files": sorted(glob.glob(downloaded["expression"])),
+                    "gene_metadata_path": downloaded["gene_metadata"],
+                    "cell_metadata_path": downloaded["cell_metadata"],
+                },
+            ),
         ]
 
-    def _generate_examples(self, data_dir):
-        expr_dir = os.path.join(data_dir, "expression")
-        gene_metadata_path = os.path.join(data_dir, "gene_metadata.parquet")
-
-        # Load gene names from gene_metadata.parquet
-        gene_metadata = pd.read_parquet(gene_metadata_path)
-        gene_names = gene_metadata["gene_id"].tolist() if "gene_id" in gene_metadata.columns else gene_metadata.index.tolist()
+    def _generate_examples(self, expression_files, gene_metadata_path, cell_metadata_path):
+        gene_df = pd.read_parquet(gene_metadata_path)
+        gene_names = gene_df["gene_id"].tolist() if "gene_id" in gene_df.columns else gene_df.index.tolist()
+        cell_df = pd.read_parquet(cell_metadata_path)
 
         idx = 0
-        for fname in sorted(os.listdir(expr_dir)):
-            if fname.endswith(".parquet"):
-                df = pd.read_parquet(os.path.join(expr_dir, fname))
-                for cell_id, row in df.iterrows():
-                    yield idx, {
-                        "cell_id": cell_id,
-                        "expression": row.to_numpy(dtype="float32").tolist(),
-                        "gene_names": gene_names,
-                    }
-                    idx += 1
+        for filepath in expression_files:
+            df = pd.read_parquet(filepath)
+            for cell_id, row in df.iterrows():
+                yield idx, {
+                    "cell_id": cell_id,
+                    "expression": row.to_numpy(dtype="float32").tolist(),
+                    "gene_names": gene_names,
+                }
+                idx += 1