Update column name sent_ids to ids

README.md

@@ -49,16 +49,16 @@ dataset_info:
           '1': semantic-break
   splits:
   - name: train
-    num_bytes: 105236889
+    num_bytes: 99374519
     num_examples: 13679
   - name: validation
-    num_bytes: 15693016
+    num_bytes: 14819645
     num_examples: 1953
   - name: test
-    num_bytes: 31140798
+    num_bytes: 29405354
     num_examples: 3907
   download_size: 94042594
-  dataset_size: 152070703
+  dataset_size: 143599518
 - config_name: en_disease
   features:
   - name: id
@@ -79,15 +79,15 @@ dataset_info:
           '1': semantic-break
   splits:
   - name: train
-    num_bytes: 22409988
+    num_bytes: 21279366
     num_examples: 2513
   - name: validation
-    num_bytes: 3190201
+    num_bytes: 3029140
     num_examples: 359
   - name: test
-    num_bytes: 6088470
+    num_bytes: 5778276
     num_examples: 718
   download_size: 94042594
-  dataset_size: 31688659
+  dataset_size: 30086782
 ---
 # Dataset Card for WikiSection (en_city, en_disease) Dataset

preprocess_util.py

@@ -72,7 +72,7 @@ def parse_split(filepath: str, drop_titles: bool = False, sent_tokenize_method:
         doc = {
             'id': id,
             'title': title,
-            'sent_ids': [],
+            'ids': [],
             'sentences': [],
             'titles_mask': [],
             'labels': [],
@@ -93,14 +93,14 @@ def parse_split(filepath: str, drop_titles: bool = False, sent_tokenize_method:
             # Add the title as a single sentence
             if not drop_titles and sec_title:
             # if not drop_titles and non_empty(sec_title):
-                doc['sent_ids'].append(f'{sec_idx}')
+                doc['ids'].append(f'{sec_idx}')
                 doc['sentences'].append(sec_title)
                 doc['titles_mask'].append(1)
                 doc['labels'].append(0)
 
             # Add the sentences
             for sent_idx, sent in enumerate(sentences):
-                doc['sent_ids'].append(f'{sec_idx}_{sent_idx}')
+                doc['ids'].append(f'{sec_idx}_{sent_idx}')
                 doc['sentences'].append(sent)
                 doc['titles_mask'].append(0)
                 doc['labels'].append(1 if sent_idx == len(sentences) - 1 else 0)

wikisection.py

@@ -110,7 +110,7 @@ class WikiSection(datasets.GeneratorBasedBuilder):
             {
                 "id": datasets.Value("string"),         # document id --> [doc0, doc1, ...]
                 "title": datasets.Value("string"),
-                "sent_ids": datasets.Sequence(          # document sentence ids --> [[doc0_sent0, doc0_sent1, ...], ...]
+                "ids": datasets.Sequence(               # document sentence ids --> [[doc0_sent0, doc0_sent1, ...], ...]
                     datasets.Value("string")
                 ),
                 "sentences": datasets.Sequence(