GH963 Restrict callable to hashableT for read_csv (#1234)

* GH963 Restrict callable to hahableT for read_csv

* GH963 Add test

Author: loicdiridollou

pandas-stubs/io/parsers/readers.pyi

@@ -27,6 +27,7 @@ from pandas._typing import (
     DtypeArg,
     DtypeBackend,
     FilePath,
+    HashableT,
     ListLikeHashable,
     ReadCsvBuffer,
     StorageOptions,
@@ -44,7 +45,7 @@ def read_csv(
     header: int | Sequence[int] | Literal["infer"] | None = ...,
     names: ListLikeHashable | None = ...,
     index_col: int | str | Sequence[str | int] | Literal[False] | None = ...,
-    usecols: UsecolsArgType = ...,
+    usecols: UsecolsArgType[HashableT] = ...,
     dtype: DtypeArg | defaultdict | None = ...,
     engine: CSVEngine | None = ...,
     converters: (
@@ -108,7 +109,7 @@ def read_csv(
     header: int | Sequence[int] | Literal["infer"] | None = ...,
     names: ListLikeHashable | None = ...,
     index_col: int | str | Sequence[str | int] | Literal[False] | None = ...,
-    usecols: UsecolsArgType = ...,
+    usecols: UsecolsArgType[HashableT] = ...,
     dtype: DtypeArg | defaultdict | None = ...,
     engine: CSVEngine | None = ...,
     converters: (
@@ -172,7 +173,7 @@ def read_csv(
     header: int | Sequence[int] | Literal["infer"] | None = ...,
     names: ListLikeHashable | None = ...,
     index_col: int | str | Sequence[str | int] | Literal[False] | None = ...,
-    usecols: UsecolsArgType = ...,
+    usecols: UsecolsArgType[HashableT] = ...,
     dtype: DtypeArg | defaultdict | None = ...,
     engine: CSVEngine | None = ...,
     converters: (
@@ -236,7 +237,7 @@ def read_table(
     header: int | Sequence[int] | Literal["infer"] | None = ...,
     names: ListLikeHashable | None = ...,
     index_col: int | str | Sequence[str | int] | Literal[False] | None = ...,
-    usecols: UsecolsArgType = ...,
+    usecols: UsecolsArgType[HashableT] = ...,
     dtype: DtypeArg | defaultdict | None = ...,
     engine: CSVEngine | None = ...,
     converters: (
@@ -300,7 +301,7 @@ def read_table(
     header: int | Sequence[int] | Literal["infer"] | None = ...,
     names: ListLikeHashable | None = ...,
     index_col: int | str | Sequence[str | int] | Literal[False] | None = ...,
-    usecols: UsecolsArgType = ...,
+    usecols: UsecolsArgType[HashableT] = ...,
     dtype: DtypeArg | defaultdict | None = ...,
     engine: CSVEngine | None = ...,
     converters: (
@@ -364,7 +365,7 @@ def read_table(
     header: int | Sequence[int] | Literal["infer"] | None = ...,
     names: ListLikeHashable | None = ...,
     index_col: int | str | Sequence[str | int] | Literal[False] | None = ...,
-    usecols: UsecolsArgType = ...,
+    usecols: UsecolsArgType[HashableT] = ...,
     dtype: DtypeArg | defaultdict | None = ...,
     engine: CSVEngine | None = ...,
     converters: (

tests/test_io.py

@@ -575,6 +575,11 @@ def test_read_csv():
             DataFrame,
         )
 
+        def cols(x: str) -> bool:
+            return x in ["a", "b"]
+
+        pd.read_csv(path, usecols=cols)
+
 
 def test_read_csv_iterator():
     with ensure_clean() as path:
@@ -727,6 +732,11 @@ def test_types_read_csv() -> None:
             pd.read_csv(path, names="abcd")  # type: ignore[call-overload] # pyright: ignore[reportArgumentType]
             pd.read_csv(path, usecols="abcd")  # type: ignore[call-overload] # pyright: ignore[reportArgumentType]
 
+            def cols2(x: set[float]) -> bool:
+                return sum(x) < 1.0
+
+            pd.read_csv("file.csv", usecols=cols2)  # type: ignore[type-var] # pyright: ignore[reportArgumentType]
+
         tfr1 = pd.read_csv(path, nrows=2, iterator=True, chunksize=3)
         check(assert_type(tfr1, TextFileReader), TextFileReader)
         tfr1.close()