make conll DatasetBuilder classes config based, by accepting descript…

…ion and citation to be specified using code still. PiperOrigin-RevId: 489442803
tensorflow · Nov 18, 2022 · 2800d1d · 2800d1d
1 parent 73e190e
commit 2800d1d
Show file tree

Hide file tree

Showing 4 changed files with 20 additions and 21 deletions.
diff --git a/docs/format_specific_dataset_builders.md b/docs/format_specific_dataset_builders.md
@@ -217,8 +217,6 @@ class MyCoNNLDataset(tfds.dataset_builders.ConllDatasetBuilder):
 
   def _info(self) -> tfds.core.DatasetInfo:
     return self.create_dataset_info(
-      description="My dataset description",
-      citation="My dataset citation",
       # ...
     )
 
@@ -329,8 +327,6 @@ class MyCoNNLUDataset(tfds.dataset_builders.ConllUDatasetBuilder):
 
   def _info(self) -> tfds.core.DatasetInfo:
     return self.create_dataset_info(
-      description="My dataset description",
-      citation="My dataset citation",
       # ...
     )
 

diff --git a/tensorflow_datasets/core/dataset_builder.py b/tensorflow_datasets/core/dataset_builder.py
@@ -881,15 +881,18 @@ def dataset_info_from_configs(self, **kwargs):
     Sub-class should call this and add information not present in config files
     using kwargs directly passed to tfds.core.DatasetInfo object.
 
+    If information is present both in passed arguments and config files, config
+    files will prevail.
+
     Args:
       **kwargs: kw args to pass to DatasetInfo directly.
     """
     metadata = self.get_metadata()
-    return dataset_info.DatasetInfo(
-        builder=self,
-        description=metadata.description,
-        citation=metadata.citation,
-        **kwargs)
+    if metadata.description:
+      kwargs["description"] = metadata.description
+    if metadata.citation:
+      kwargs["citation"] = metadata.citation
+    return dataset_info.DatasetInfo(builder=self, **kwargs)
 
   @abc.abstractmethod
   @utils.docs.doc_private

diff --git a/tensorflow_datasets/core/dataset_builders/conll/conll_dataset_builder.py b/tensorflow_datasets/core/dataset_builders/conll/conll_dataset_builder.py
@@ -90,28 +90,28 @@ def builder_config(self) -> ConllBuilderConfig:
 
   def create_dataset_info(
       self,
-      description: str,
+      description: Optional[str] = None,
       supervised_keys: Optional[dataset_info.SupervisedKeysType] = None,
       homepage: Optional[str] = None,
       citation: Optional[str] = None,
   ) -> dataset_info.DatasetInfo:
     """Initializes `dataset_info.DatasetInfo` for Conll datasets.
 
     Args:
-      description: A short, markdown-formatted description of the dataset.
+      description: [DEPRECATED] A short, markdown-formatted description of the
+        dataset. Prefer placing description in `README.md` file.
       supervised_keys:  Specifies the input structure for supervised learning,
         if applicable for the dataset, used with "as_supervised". Typically this
         is a `(input_key, target_key)` tuple.
       homepage: The homepage of the dataset, if applicable for this dataset.
-      citation: The citation to use for this dataset, if applicable for this
-        dataset.
+      citation: [DEPRECATED] The citation to use for this dataset, if applicable
+        for this dataset. Prefer placing citations in `CITATIONS.bib` file.
 
     Returns:
       `dataset_info.DatasetInfo` for Conll datasets, populated with the values
       from the provided arguments.
     """
-    return dataset_info.DatasetInfo(
-        builder=self,
+    return self.dataset_info_from_configs(
         description=description,
         features=self.builder_config.features_dict,
         supervised_keys=supervised_keys,

diff --git a/tensorflow_datasets/core/dataset_builders/conll/conllu_dataset_builder.py b/tensorflow_datasets/core/dataset_builders/conll/conllu_dataset_builder.py
@@ -155,28 +155,28 @@ def builder_config(self) -> ConllUBuilderConfig:
 
   def create_dataset_info(
       self,
-      description: str,
+      description: Optional[str] = None,
       supervised_keys: Optional[dataset_info.SupervisedKeysType] = None,
       homepage: Optional[str] = None,
       citation: Optional[str] = None,
   ) -> dataset_info.DatasetInfo:
     """Initializes `dataset_info.DatasetInfo` for Conll-U datasets.
 
     Args:
-      description: A short, markdown-formatted description of the dataset.
+      description: [DEPRECATED] A short, markdown-formatted description of the
+        dataset. Prefer placing description in `README.md` file.
       supervised_keys:  Specifies the input structure for supervised learning,
         if applicable for the dataset, used with "as_supervised". Typically this
         is a `(input_key, target_key)` tuple.
       homepage: The homepage of the dataset, if applicable for this dataset.
-      citation: The citation to use for this dataset, if applicable for this
-        dataset.
+      citation: [DEPRECATED] The citation to use for this dataset, if applicable
+        for this dataset. Prefer placing citations in `CITATIONS.bib` file.
 
     Returns:
       `dataset_info.DatasetInfo` for Conll-U datasets, populated with the values
       from the provided arguments.
     """
-    return dataset_info.DatasetInfo(
-        builder=self,
+    return self.dataset_info_from_configs(
         description=description,
         features=self.builder_config.features_dict,
         supervised_keys=supervised_keys,