manual ruff fixes

pyproject.toml

@@ -54,7 +54,10 @@ ignore = [
 ]
 
 [tool.ruff.lint.per-file-ignores]
-"**/tests/*" = ["D103"] # Ignore method docstring errors in tests
+"**/tests/*" = [
+    "D103", # Ignore method docstring errors in tests
+    "PD901", # Allow `df` variable name in tests
+]
 
 [tool.ruff.lint.pydocstyle]
 convention = "google"

src/extractor/extractors/io.py

@@ -110,13 +110,13 @@ def _remove_nan(value: Any) -> Any:
     Returns:
         The value with NA and NaN replaced.
     """
-    if type(value) == dict:
+    if isinstance(value, dict):
         for dkey, dvalue in value.items():
             value[dkey] = _remove_nan(dvalue)
-    elif type(value) == list:
+    elif isinstance(value, list):
         for i, item in enumerate(value):
             value[i] = _remove_nan(item)
-    elif type(value) == float and np.isnan(value):
+    elif isinstance(value, float) and np.isnan(value):
         value = None
     elif value is pd.NA:
         value = None

src/extractor/extractors/posts.py

@@ -87,7 +87,7 @@ def load_posts(
     # Get the "url" property if there is an image
     posts_df["og_image_url"] = posts_df["yoast_head_json.og_image"].apply(
         lambda image: image[0]["url"]
-        if type(image) != float and len(image) > 0
+        if not isinstance(image, float) and len(image) > 0
         else None
     )
 

src/extractor/parse/translations/_pickers.py

@@ -84,8 +84,9 @@ def add_translation(self, href: str, lang: str) -> None:
 
 
 class Polylang(LangPicker):
-    """Language picker from the plugin \
-    `Polylang <https://wordpress.org/plugins/polylang/>`.
+    """Language picker from the plugin Polylang.
+
+     `WordPress plugin page <https://wordpress.org/plugins/polylang/>`.
 
     Has the structure::
 

tests/extractors/test_media.py

@@ -20,8 +20,8 @@ def media_df(media_df_and_registry):
 
 def test_media_times(media_df):
     media_1 = media_df.loc[1]
-    assert type(media_1.date_gmt) == pd.Timestamp
-    assert type(media_1.modified_gmt) == pd.Timestamp
+    assert isinstance(media_1.date_gmt, pd.Timestamp)
+    assert isinstance(media_1.modified_gmt, pd.Timestamp)
 
     assert media_1.date_gmt.tzinfo is None, "date_gmt had timezone information"
     assert media_1.modified_gmt.tzinfo is None, "modified_gmt had timezone information"

tests/extractors/test_pages.py

@@ -27,8 +27,8 @@ def test_equals_expected(datadir, pages_df):
 
 def test_post_times(pages_df):
     post_1 = pages_df.loc[1]
-    assert type(post_1.date_gmt) == pd.Timestamp
-    assert type(post_1.modified_gmt) == pd.Timestamp
+    assert isinstance(post_1.date_gmt, pd.Timestamp)
+    assert isinstance(post_1.modified_gmt, pd.Timestamp)
 
     assert post_1.date_gmt.tzinfo is None, "date_gmt had timezone information"
     assert post_1.modified_gmt.tzinfo is None, "modified_gmt had timezone information"

tests/extractors/test_posts.py

@@ -86,8 +86,8 @@ def posts_df(posts_df_and_registry):
 
 def test_post_times(posts_df):
     post_1 = posts_df.loc[1]
-    assert type(post_1.date_gmt) == pd.Timestamp
-    assert type(post_1.modified_gmt) == pd.Timestamp
+    assert isinstance(post_1.date_gmt, pd.Timestamp)
+    assert isinstance(post_1.modified_gmt, pd.Timestamp)
 
     assert post_1.date_gmt.tzinfo is None, "date_gmt had timezone information"
     assert post_1.modified_gmt.tzinfo is None, "modified_gmt had timezone information"
@@ -172,7 +172,7 @@ def test_translations_bidirectional(posts_df_and_registry):
     posts_df, registry = posts_df_and_registry
     posts_df = resolve_post_translations(registry, posts_df)
     # Currently 1 <-> 2, let's remove 1 <- 2
-    posts_df.at[2, "translations"] = []
+    posts_df.loc[2, "translations"] = []
 
     posts_df = ensure_translations_undirected(posts_df)