update via advice from RapidFuzz#291

based on rapidfuzz/RapidFuzz#291

dtoolkit/accessor/series/textdistance.py

@@ -1,7 +1,6 @@
 from __future__ import annotations
 
 from functools import lru_cache
-from functools import wraps
 from typing import Callable
 from warnings import warn
 
@@ -21,6 +20,7 @@ def textdistance(
     other: None | str | pd.Series = None,
     method: Callable = None,
     align: bool = True,
+    **kwargs,
 ) -> pd.Series:
     """
     Return a ``Series`` containing the text distance to aligned ``other``.
@@ -60,10 +60,6 @@ def textdistance(
     --------
     textdistance_matrix
 
-    Notes
-    -----
-    The distance of any value compared to nan or None is 0.
-
     Examples
     --------
     >>> import dtoolkit
@@ -87,11 +83,12 @@ def textdistance(
         raise TypeError(f"Expected string dtype, but got {s.dtype!r}.")
 
     if method is None:
-        method = __import__("rapidfuzz.fuzz").fuzz.ratio
-    method = lru_cache(check_none(check_nan(method)))
+        method = __import__("rapidfuzz").fuzz.ratio
+    method = lru_cache(method)
 
     if isinstance(other, str):
-        return s.apply(method, args=(other,))
+        return s.apply(method, args=(other,), **kwargs)
+
     elif isinstance(other, pd.Series):
         if not is_string_dtype(other):
             raise TypeError(f"Expected Series(string), but got {other.dtype!r}.")
@@ -104,37 +101,9 @@ def textdistance(
             raise ValueError(f"{s.size=} != {other.size=}.")
 
         return pd.Series(
-            (method(*xy) for xy in zip(s, other)),
+            (method(*xy, **kwargs) for xy in zip(s, other)),
             name=s.name,
             index=s.index,
         )
-    elif other is None or (not is_list_like(other) and pd.isna(other)):
-        # NOTE:
-        # - pd.na(Series) returns array-like of bool
-        #   to make sure pd.isna(other) returns bool
-        #   need to other is not array-like
-        # - compare to None or nan always returns 0
-        #   the behavior is following rapidfuzz.fuzz.ratio
-        return pd.Series(np.zeros(s.size), name=s.name, index=s.index)
 
     raise TypeError(f"Expected Series(string), but got {type(other).__name__!r}.")
-
-
-def check_none(func):
-    @wraps(func)
-    def decorator(*args, **kwargs):
-        # NOTE: compare to None always returns 0
-        # the behavior is following rapidfuzz.fuzz.ratio
-        return 0 if args[0] is None or args[1] is None else func(*args, **kwargs)
-
-    return decorator
-
-
-def check_nan(func):
-    @wraps(func)
-    def decorator(*args, **kwargs):
-        # NOTE: compare to nan always returns 0
-        # the behavior is following rapidfuzz.fuzz.ratio
-        return 0 if pd.isna(args[0]) or pd.isna(args[1]) else func(*args, **kwargs)
-
-    return decorator

dtoolkit/accessor/series/textdistance_matrix.py

@@ -49,10 +49,6 @@ def textdistance_matrix(
     --------
     textdistance
 
-    Notes
-    -----
-    Can't handle nan or None type value.
-
     Examples
     --------
     >>> import dtoolkit

test/accessor/series/test_textdistance_matrix.py

@@ -36,13 +36,6 @@
             rapidfuzz.string_metric.levenshtein,
             pd.DataFrame([[4, 9], [6, 9]]),
         ),
-        # other elements contain None or nan
-        (
-            pd.Series(["hello", "world", "!"]),
-            pd.Series(["hi!", None, float("nan")]),
-            None,
-            pd.DataFrame([[25, 0, 0], [0, 0, 0], [50, 0, 0]]),
-        ),
     ],
 )
 def test_work(s, other, method, expected):