fix grouped loading, updated size for documents

LaCour.py

@@ -194,6 +194,8 @@ class LaCourDataset(datasets.GeneratorBasedBuilder):
             logger.info("generating examples from = %s", filepaths)
             cols = df.columns.tolist()
             cols.remove('appno')
+            # fix potential null values for judges
+            df['judges'] = df['judges'].fillna('')
             # group appnos to avoid duplicates
             df = df.groupby(cols)['appno'].apply(';'.join).reset_index()
             for _, row in df.iterrows():

README.md

@@ -80,10 +80,10 @@ dataset_info:
     dtype: bool
   splits:
   - name: train
-    num_bytes: 962951
-    num_examples: 369
+    num_bytes: 1007746
+    num_examples: 379
   download_size: 1714597
-  dataset_size: 962951
+  dataset_size: 1007746
 - config_name: transcripts
   features:
   - name: id