Improve doc scripts.

PiperOrigin-RevId: 798156419

Author: The TensorFlow Datasets Authors

tensorflow_datasets/scripts/documentation/build_community_catalog.py

@@ -168,7 +168,7 @@ def to_namespace_overview(self) -> str:
     )
 
   def to_details_markdown(self) -> str:
-    """ "Markdown to be shown on the details page for the namespace."""
+    """Markdown to be shown on the details page for the namespace."""
     extra_links = self.format_extra_links(prefix='*   ', infix='\n')
     details = self.templates.dataset_details_template.format(
         name=self.name,
@@ -194,9 +194,6 @@ def dataset_info_per_config(
 
   def documentation(self, keep_short: bool = False) -> str:
     """Returns detailed documentation for all configs of this dataset."""
-    # TODO(weide): if e.g. the description contains markdown chars, then it
-    # messes up the page. Try escaping backticks or using code blocks.
-    # TODO(weide): how to format citation?
     header_template = '## {config_name}'
     template = textwrap.dedent("""
       Use the following command to load this dataset in TFDS:
@@ -207,9 +204,7 @@ def documentation(self, keep_short: bool = False) -> str:
 
       *   **Description**:
 
-      ```
       {description}
-      ```
 
       *   **License**: {license}
       *   **Version**: {version}
@@ -364,6 +359,7 @@ def _parse_dataset_info_proto(
           )
       )
 
+    version = None
     if isinstance(config['version'], dict):
       version = config['version']['version_str']
     elif isinstance(config['version'], str):

tensorflow_datasets/scripts/documentation/dataset_markdown_builder.py

@@ -222,9 +222,9 @@ def _list_versions(self, builder: tfds.core.DatasetBuilder):
     all_versions = set(tfds.core.Version(v) for v in all_versions)
     for v in sorted(all_versions):  # List all available versions
       if v == builder.version:  # Highlight the default version
-        version_name = '**`{}`** (default)'.format(str(v))
+        version_name = f'**`{v}`** (default)'
       else:
-        version_name = '`{}`'.format(str(v))
+        version_name = f'`{v}`'
       if (
           v in curr_versions  # Filter versions only present in RELEASE_NOTES
           and self._nightly_doc_util
@@ -322,14 +322,14 @@ def _build_autocached_info(self, builder: tfds.core.DatasetBuilder):
       autocached_info_parts = []
       if always_cached:
         split_names_str = ', '.join(always_cached)
-        autocached_info_parts.append('Yes ({})'.format(split_names_str))
+        autocached_info_parts.append(f'Yes ({split_names_str})')
       if never_cached:
         split_names_str = ', '.join(never_cached)
-        autocached_info_parts.append('No ({})'.format(split_names_str))
+        autocached_info_parts.append(f'No ({split_names_str})')
       if unshuffle_cached:
         split_names_str = ', '.join(unshuffle_cached)
         autocached_info_parts.append(
-            'Only when `shuffle_files=False` ({})'.format(split_names_str)
+            f'Only when `shuffle_files=False` ({split_names_str})'
         )
       autocached_info = ', '.join(autocached_info_parts)
     return autocached_info
@@ -346,7 +346,7 @@ class SplitInfoSection(Section):
 
   def _get_num_examples(self, split_info):
     if split_info.num_examples:
-      return '{:,}'.format(split_info.num_examples)
+      return f'{split_info.num_examples:,}'
     return 'Not computed'
 
   def get_key(self, builder: tfds.core.DatasetBuilder):