Automated Code Change

tensorflow_datasets/datasets/patch_camelyon/patch_camelyon_dataset_builder.py

@@ -54,13 +54,13 @@ def _split_generators(self, dl_manager):
     paths = dl_manager.download_and_extract(resources)
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST, gen_kwargs=dict(split='test', paths=paths)
+            name=tfds.Split.TEST, gen_kwargs=dict(split='test', paths=paths)  # pyrefly: ignore[missing-attribute]
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN, gen_kwargs=dict(split='train', paths=paths)
+            name=tfds.Split.TRAIN, gen_kwargs=dict(split='train', paths=paths)  # pyrefly: ignore[missing-attribute]
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs=dict(split='valid', paths=paths),
         ),
     ]

tensorflow_datasets/datasets/paws_wiki/paws_wiki_dataset_builder.py

@@ -44,9 +44,9 @@
     _UNLABELED_FINAL: ["train", "dev"],
 }
 _SPLIT_MAPPINGS = {
-    "train": tfds.Split.TRAIN,
-    "dev": tfds.Split.VALIDATION,
-    "test": tfds.Split.TEST,
+    "train": tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
+    "dev": tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
+    "test": tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
 }
 
 _CLASS_LABELS = ["different_meaning", "paraphrase"]
@@ -101,10 +101,10 @@ def _info(self):
   def _split_generators(self, dl_manager):
     """Returns SplitGenerators."""
     dl_paths = dl_manager.download_and_extract(_DOWNLOAD_URLS)
-    subset = self.builder_config.subset
+    subset = self.builder_config.subset  # pyrefly: ignore[missing-attribute]
     labels_path = os.path.join(dl_paths[subset], _EXTRACTED_FOLDERS[subset])
 
-    if self.builder_config.tokenized:
+    if self.builder_config.tokenized:  # pyrefly: ignore[missing-attribute]
       mappings_path = ""
       tags2texts = {}
     else:
@@ -193,7 +193,7 @@ def _generate_examples(
           key = row["id"]
           label_str = (
               "noisy_label"
-              if self.builder_config.subset == _UNLABELED_FINAL
+              if self.builder_config.subset == _UNLABELED_FINAL  # pyrefly: ignore[missing-attribute]
               else "label"
           )
           example = {

tensorflow_datasets/datasets/paws_x_wiki/paws_x_wiki_dataset_builder.py

@@ -75,32 +75,32 @@ def _split_generators(self, dl_manager):
     base_path = os.path.join(dl_path["x-final"], "x-final")
 
     # Name of file for training for 'en' is different from other languages
-    if self.builder_config.language == "en":
+    if self.builder_config.language == "en":  # pyrefly: ignore[missing-attribute]
       training_path = os.path.join(
           base_path, self.builder_config.language, "train.tsv"
       )
     else:
-      training_path = os.path.join(
+      training_path = os.path.join(  # pyrefly: ignore[no-matching-overload]
           base_path, self.builder_config.language, "translated_train.tsv"
       )
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"path": training_path},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
-                "path": os.path.join(
+                "path": os.path.join(  # pyrefly: ignore[no-matching-overload]
                     base_path, self.builder_config.language, "test_2k.tsv"
                 )
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
-                "path": os.path.join(
+                "path": os.path.join(  # pyrefly: ignore[no-matching-overload]
                     base_path, self.builder_config.language, "dev_2k.tsv"
                 )
             },

tensorflow_datasets/datasets/penguins/penguins_dataset_builder.py

@@ -159,11 +159,11 @@ class Builder(tfds.core.GeneratorBasedBuilder):
   def _info(self) -> tfds.core.DatasetInfo:
     """Returns the dataset metadata."""
     supervised_keys = None
-    features = self.builder_config.features
+    features = self.builder_config.features  # pyrefly: ignore[missing-attribute]
     supervised_features = features.copy()
-    label_name = self.builder_config.label
+    label_name = self.builder_config.label  # pyrefly: ignore[missing-attribute]
 
-    if self.builder_config.name == 'processed':
+    if self.builder_config.name == 'processed':  # pyrefly: ignore[missing-attribute]
       label_feature = supervised_features.pop(label_name, None)
       supervised_keys = ('features', label_name)
       features = {
@@ -187,7 +187,7 @@ def _info(self) -> tfds.core.DatasetInfo:
 
   def _split_generators(self, dl_manager: tfds.download.DownloadManager):
     """Returns SplitGenerators."""
-    path = dl_manager.download(_PENGUINS_PATH + self.builder_config.file_name)
+    path = dl_manager.download(_PENGUINS_PATH + self.builder_config.file_name)  # pyrefly: ignore[missing-attribute]
     return {'train': self._generate_examples(path)}
 
   def _generate_examples(self, path):
@@ -197,8 +197,8 @@ def _generate_examples(self, path):
         row = {f: self._clean_up(f, v) for f, v in row.items()}
 
         # Pack features if requested.
-        if self.builder_config.name == 'processed':
-          label_name = self.builder_config.label
+        if self.builder_config.name == 'processed':  # pyrefly: ignore[missing-attribute]
+          label_name = self.builder_config.label  # pyrefly: ignore[missing-attribute]
           label = row.pop(label_name, None)
           row = list(row.values())
           yield i, {'features': row, label_name: label}
@@ -207,21 +207,21 @@ def _generate_examples(self, path):
 
   def _clean_up(self, field, value):
     """Applies field-level pre-processing, if needed."""
-    if not self.builder_config.cleanup:
+    if not self.builder_config.cleanup:  # pyrefly: ignore[missing-attribute]
       return value
     if field not in self.builder_config.cleanup:
       return value
 
-    feature_type = self.builder_config.features[field]
+    feature_type = self.builder_config.features[field]  # pyrefly: ignore[missing-attribute]
     if feature_type == np.float32:
       # Field is a float. If it won't parse, clean it up.
       try:
         return float(value)
       except ValueError:
-        return self.builder_config.cleanup[field]
+        return self.builder_config.cleanup[field]  # pyrefly: ignore[bad-index]
     elif isinstance(feature_type, tfds.features.ClassLabel):
       # Field is a class. If it's OOV, clean it up.
       if value not in feature_type.names:
-        return self.builder_config.cleanup[field]
+        return self.builder_config.cleanup[field]  # pyrefly: ignore[bad-index]
 
     return value

tensorflow_datasets/datasets/pet_finder/pet_finder_dataset_builder.py

@@ -100,15 +100,15 @@ def _split_generators(self, dl_manager):
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "csv_name": "train.csv",
                 "csv_paths": dl_paths["train"],
                 "img_paths": dl_paths["train_images"],
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "csv_name": "test.csv",
                 "csv_paths": dl_paths["test"],

tensorflow_datasets/datasets/pg19/pg19_dataset_builder.py

@@ -59,21 +59,21 @@ def _split_generators(self, dl_manager):
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'metadata': metadata_dict,
                 'filepath': os.path.join(_DATA_DIR, 'train'),
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'metadata': metadata_dict,
                 'filepath': os.path.join(_DATA_DIR, 'validation'),
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'metadata': metadata_dict,
                 'filepath': os.path.join(_DATA_DIR, 'test'),

tensorflow_datasets/datasets/piqa/piqa_dataset_builder.py

@@ -58,7 +58,7 @@ def _split_generators(self, dl_manager):
     # Specify the splits
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "data_path": os.path.join(
                     extracted_path, "physicaliqa-train-dev/train.jsonl"
@@ -69,7 +69,7 @@ def _split_generators(self, dl_manager):
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "data_path": os.path.join(
                     extracted_path, "physicaliqa-train-dev/dev.jsonl"

tensorflow_datasets/datasets/places365_small/places365_small_dataset_builder.py

@@ -18,7 +18,7 @@
 import os
 
 from etils import epath
-import six.moves.urllib as urllib
+import six.moves.urllib as urllib  # pyrefly: ignore[missing-source-for-stubs]
 import tensorflow_datasets.public_api as tfds
 
 _BASE_URL = "http://data.csail.mit.edu/places/places365/"

tensorflow_datasets/datasets/plant_leaves/plant_leaves_dataset_builder.py

@@ -80,7 +80,7 @@ def _split_generators(self, dl_manager):
     archive_path = dl_manager.download(_DOWNLOAD_URL)
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"archive": dl_manager.iter_archive(archive_path)},
         )
     ]

tensorflow_datasets/datasets/plant_village/plant_village_dataset_builder.py

@@ -84,7 +84,7 @@ def _split_generators(self, dl_manager):
     path = dl_manager.download_and_extract(_URL)
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN, gen_kwargs={"datapath": path}
+            name=tfds.Split.TRAIN, gen_kwargs={"datapath": path}  # pyrefly: ignore[missing-attribute]
         )
     ]
 

tensorflow_datasets/datasets/plantae_k/plantae_k_dataset_builder.py

@@ -91,7 +91,7 @@ def _split_generators(self, dl_manager):
             )
       return [
           tfds.core.SplitGenerator(
-              name=tfds.Split.TRAIN, gen_kwargs={"image_files": image_files}
+              name=tfds.Split.TRAIN, gen_kwargs={"image_files": image_files}  # pyrefly: ignore[missing-attribute]
           )
       ]
 

tensorflow_datasets/datasets/protein_net/protein_net_dataset_builder.py

@@ -34,7 +34,7 @@
 
 def _parse_array(lines: Sequence[str]) -> np.ndarray:
   """Parse lines of tab-separated numbers into an array."""
-  lines = [x.split('\t') for x in lines]
+  lines = [x.split('\t') for x in lines]  # pyrefly: ignore[bad-assignment]
   return np.array(lines, dtype=np.float32)
 
 
@@ -165,16 +165,16 @@ def _split_generators(
       self, dl_manager: tfds.download.DownloadManager
   ) -> Dict[Union[str, tfds.Split], _ExampleIterator]:
     """Returns SplitGenerators."""
-    name = self.builder_config.name  # Configurable dataset (config) name.
+    name = self.builder_config.name  # Configurable dataset (config) name.  # pyrefly: ignore[missing-attribute]
     path = dl_manager.download_and_extract(
         urllib.parse.urljoin(self.URL, self.FILES[name])
     )
 
     splits = {
-        tfds.Split.VALIDATION: self._generate_examples(
+        tfds.Split.VALIDATION: self._generate_examples(  # pyrefly: ignore[missing-attribute]
             os.path.join(path, name, 'validation')
         ),
-        tfds.Split.TEST: self._generate_examples(
+        tfds.Split.TEST: self._generate_examples(  # pyrefly: ignore[missing-attribute]
             os.path.join(path, name, 'testing')
         ),
     }

tensorflow_datasets/datasets/qa4mre/qa4mre_dataset_builder.py

@@ -92,10 +92,10 @@ def _get_question(
       'document_id': document_id,
       'document_str': document_str,
       'question_id': question_id,
-      'question_str': question_str,
+      'question_str': question_str,  # pyrefly: ignore[unbound-name]
       'answer_options': possible_answers,
-      'correct_answer_id': correct_answer_id,
-      'correct_answer_str': correct_answer_str,
+      'correct_answer_id': correct_answer_id,  # pyrefly: ignore[unbound-name]
+      'correct_answer_str': correct_answer_str,  # pyrefly: ignore[unbound-name]
   }
 
   return id_, feats
@@ -247,17 +247,17 @@ def _split_generators(self, dl_manager):
     cfg = self.builder_config
     download_urls = dict()
 
-    if cfg.track == 'main':
-      download_urls['{}.main.{}'.format(cfg.year, cfg.lang)] = os.path.join(
-          _BASE_URL, PATHS[cfg.year]['_PATH_TMPL_MAIN_GS'].format(cfg.lang)
+    if cfg.track == 'main':  # pyrefly: ignore[missing-attribute]
+      download_urls['{}.main.{}'.format(cfg.year, cfg.lang)] = os.path.join(  # pyrefly: ignore[missing-attribute]
+          _BASE_URL, PATHS[cfg.year]['_PATH_TMPL_MAIN_GS'].format(cfg.lang)  # pyrefly: ignore[missing-attribute]
       )  # pytype: disable=attribute-error
 
-    if cfg.year in ['2012', '2013'] and cfg.track == 'alzheimers':
-      download_urls['{}.alzheimers.EN'.format(cfg.year)] = os.path.join(
+    if cfg.year in ['2012', '2013'] and cfg.track == 'alzheimers':  # pyrefly: ignore[missing-attribute]
+      download_urls['{}.alzheimers.EN'.format(cfg.year)] = os.path.join(  # pyrefly: ignore[no-matching-overload]
           _BASE_URL, PATHS[cfg.year]['_PATH_ALZHEIMER']
       )
 
-    if cfg.year == '2013' and cfg.track == 'entrance_exam':
+    if cfg.year == '2013' and cfg.track == 'entrance_exam':  # pyrefly: ignore[missing-attribute]
       download_urls['2013.entrance_exam.EN'] = os.path.join(
           _BASE_URL, PATHS[cfg.year]['_PATH_ENTRANCE_EXAM']
       )
@@ -266,7 +266,7 @@ def _split_generators(self, dl_manager):
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'filepath': downloaded_files[
                     '{}.{}.{}'.format(cfg.year, cfg.track, cfg.lang)
@@ -293,7 +293,7 @@ def _generate_examples(self, filepath):
                 topic_id,
                 topic_name,
                 test_id,
-                document_id,
-                document_str,
+                document_id,  # pyrefly: ignore[unbound-name]
+                document_str,  # pyrefly: ignore[unbound-name]
                 question,
             )

tensorflow_datasets/datasets/qasc/qasc_dataset_builder.py

@@ -55,15 +55,15 @@ def _split_generators(self, dl_manager):
     data_dir = os.path.join(dl_dir["QASC_Dataset"], "QASC_Dataset")
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"filepath": os.path.join(data_dir, "train.jsonl")},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"filepath": os.path.join(data_dir, "test.jsonl")},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"filepath": os.path.join(data_dir, "dev.jsonl")},
         ),
     ]

tensorflow_datasets/datasets/qm9/qm9_dataset_builder.py

@@ -131,7 +131,7 @@ def _process_molecule(
     thermo = 0
     for z in atoms['Z'].values:
       thermo += v[z]
-    example[f'{k}_atomization'] = example[k] - thermo
+    example[f'{k}_atomization'] = example[k] - thermo  # pyrefly: ignore[unsupported-operation]
 
   return example
 
@@ -234,7 +234,7 @@ class Builder(tfds.core.GeneratorBasedBuilder):
   def _info(self) -> tfds.core.DatasetInfo:
     """Returns the dataset metadata."""
     return self.dataset_info_from_configs(
-        disable_shuffling=self.builder_config.permutation_seed is None,
+        disable_shuffling=self.builder_config.permutation_seed is None,  # pyrefly: ignore[missing-attribute]
         features=tfds.features.FeaturesDict({
             'num_atoms': tfds.features.Tensor(shape=(), dtype=np.int64),
             'charges': tfds.features.Tensor(shape=(29,), dtype=np.int64),
@@ -306,9 +306,9 @@ def _split_generators(
 
     split_ids = _get_split_ids(
         uncharacterized,
-        permutation_seed=self.builder_config.permutation_seed,
-        train_size=self.builder_config.train_size,
-        validation_size=self.builder_config.validation_size,
+        permutation_seed=self.builder_config.permutation_seed,  # pyrefly: ignore[missing-attribute]
+        train_size=self.builder_config.train_size,  # pyrefly: ignore[missing-attribute]
+        validation_size=self.builder_config.validation_size,  # pyrefly: ignore[missing-attribute]
     )
 
     return {

tensorflow_datasets/datasets/quality/quality_dataset_builder.py

@@ -88,7 +88,7 @@ def _split_generators(self, dl_manager: tfds.download.DownloadManager):
     """Returns SplitGenerators."""
     path = dl_manager.download_and_extract(_DOWNLOAD_URL)
 
-    if self.builder_config.stripped:
+    if self.builder_config.stripped:  # pyrefly: ignore[missing-attribute]
       return {
           'train': self._generate_examples(
               path / 'QuALITY.v0.9.htmlstripped.train', 'train'