Tokenization fixed

tokenization_nicheformer.py

@@ -290,9 +290,17 @@ class NicheformerTokenizer(PreTrainedTokenizer):
             X = adata.X
 
             # Get metadata for special tokens
-            modality = adata.obs.get('modality', None)
-            species = adata.obs.get('specie', None)
-            technology = adata.obs.get('assay', None)
+            # Print column types
+            print("\nColumn types:")
+            if 'modality' in adata.obs.columns:
+                print(f"modality type: {type(adata.obs['modality'])} with dtype: {adata.obs['modality'].dtype}")
+            if 'specie' in adata.obs.columns:
+                print(f"specie type: {type(adata.obs['specie'])} with dtype: {adata.obs['specie'].dtype}")
+            if 'assay' in adata.obs.columns:
+                print(f"assay type: {type(adata.obs['assay'])} with dtype: {adata.obs['assay'].dtype}")
+            modality = adata.obs['modality'] if 'modality' in adata.obs.columns else None
+            species = adata.obs['specie'] if 'specie' in adata.obs.columns else None
+            technology = adata.obs['assay'] if 'assay' in adata.obs.columns else None
 
             print(f"Modality: {modality}")
             print(f"Species: {species}")