Refactor entity matching name cleaner to be more efficient

docs/release_notes.rst

@@ -70,11 +70,17 @@ EPA CEMS
 ~~~~~~~~
 * Added 2024 Q3 of CEMS data. See :issue:`3943` and :pr:`3948`.
 
-FERC to EIA Record Linkage
+Record Linkage
 ^^^^^^^^^^^^^^^^^^^^^^^^^^
 * Updated the ``splink`` FERC to EIA development notebook to be compatible with
   the latest version of ``splink``. This notebook is not run in production but
   is helpful for visualizing model weights and what is happening under the hood.
+* Updated ``pudl.analysis.record_linkage.name_cleaner`` company name cleaning
+  module to be more efficient by removing all ``.apply`` and instead use
+  ``pd.Series.replace`` to make regex replacement rules vectorized. Also removed
+  some of the allowed replacement rules to make the cleaner simpler and more
+  effective. This module runs approximately 3x faster now when cleaning a
+  string Series.
 
 .. _release-v2024.10.0:
 

src/pudl/analysis/record_linkage/eia_ferc1_record_linkage.py

@@ -78,11 +78,9 @@
     cleaning_rules_list=[
         "remove_word_the_from_the_end",
         "remove_word_the_from_the_beginning",
-        "replace_amperstand_between_space_by_AND",
+        "replace_ampersand_by_AND",
         "replace_hyphen_by_space",
-        "replace_hyphen_between_spaces_by_single_space",
         "replace_underscore_by_space",
-        "replace_underscore_between_spaces_by_single_space",
         "remove_all_punctuation",
         "remove_numbers",
         "remove_math_symbols",