CI: check for some unused ignore comments (#213)

* CI: check for some unused ignore comments

* pyright ignore

* close file handles

* address comments

* mypy errors?

* comments

* unused import

Author: twoertwein

pandas-stubs/core/groupby/generic.pyi

@@ -141,17 +141,17 @@ class _DataFrameGroupByNonScalar(DataFrameGroupBy):
 class DataFrameGroupBy(GroupBy):
     def any(self, skipna: bool = ...) -> DataFrame: ...
     def all(self, skipna: bool = ...) -> DataFrame: ...
-    # mypy and pyright see these overloads as overlapping
     @overload
-    def apply(  # type: ignore[misc]
+    def apply(
         self, func: Callable[[DataFrame], Scalar | list | dict], *args, **kwargs
     ) -> Series: ...
     @overload
-    def apply(  # type: ignore[misc]
+    def apply(
         self, func: Callable[[DataFrame], Series | DataFrame], *args, **kwargs
     ) -> DataFrame: ...
+    # error: Overload 3 for "apply" will never be used because its parameters overlap overload 1
     @overload
-    def apply(  # type: ignore[misc]
+    def apply(  # pyright: ignore[reportOverlappingOverload]
         self, func: Callable[[Iterable], float], *args, **kwargs
     ) -> DataFrame: ...
     def aggregate(self, arg: AggFuncType = ..., *args, **kwargs) -> DataFrame: ...

pandas-stubs/core/indexes/datetimes.pyi

@@ -72,9 +72,8 @@ class DatetimeIndex(DatetimeTimedeltaMixin, DatetimeIndexProperties):
     def __le__(self, other: Timestamp) -> np_ndarray_bool: ...
     def __gt__(self, other: Timestamp) -> np_ndarray_bool: ...
     def __ge__(self, other: Timestamp) -> np_ndarray_bool: ...
-    # ignore for mypy because we know dtype of a DatetimeIndex
     @property
-    def dtype(self) -> np.dtype | DatetimeTZDtype: ...  # type: ignore[misc]
+    def dtype(self) -> np.dtype | DatetimeTZDtype: ...
 
 def date_range(
     start=...,

pandas-stubs/io/parsers.pyi

@@ -516,6 +516,8 @@ class TextFileReader(abc.Iterator):
     def __init__(self, f, engine=..., **kwds) -> None: ...
     def close(self) -> None: ...
     def __next__(self): ...
+    def __enter__(self) -> TextFileReader: ...
+    def __exit__(self, exc_type, exc_value, traceback) -> None: ...
     def read(self, nrows=...): ...
     def get_chunk(self, size=...): ...
 

pyproject.toml

@@ -138,7 +138,7 @@ no_implicit_optional = true
 strict_optional = true
 # Configuring warnings
 warn_redundant_casts = true
-warn_unused_ignores = false # Change from pandas
+warn_unused_ignores = true
 warn_no_return = true
 warn_return_any = false # TODO
 warn_unreachable = false # GH#27396
@@ -168,6 +168,7 @@ reportMissingParameterType = false
 reportMissingTypeArgument = false
 reportMissingTypeStubs = false
 reportUnknownArgumentType = false
+reportUnusedExpression = false
 reportUnknownLambdaType = false
 reportUnknownMemberType = false
 reportUnknownParameterType = false

tests/__init__.py

@@ -1,21 +1,28 @@
 from __future__ import annotations
 
+from typing import (
+    TYPE_CHECKING,
+    Final,
+)
 
-def check(
-    actual: object, klass: type, dtype: type | None = None, attr: str = "left"
-) -> None:
+from pandas._typing import T
+
+TYPE_CHECKING_INVALID_USAGE: Final = TYPE_CHECKING
+
+
+def check(actual: T, klass: type, dtype: type | None = None, attr: str = "left") -> T:
 
     if not isinstance(actual, klass):
         raise RuntimeError(f"Expected type '{klass}' but got '{type(actual)}'")
     if dtype is None:
-        return None
+        return actual  # type: ignore[return-value]
 
     if hasattr(actual, "__iter__"):
         value = next(iter(actual))  # type: ignore[call-overload]
     else:
         assert hasattr(actual, attr)
-        value = getattr(actual, attr)  # type: ignore[attr-defined]
+        value = getattr(actual, attr)
 
     if not isinstance(value, dtype):
         raise RuntimeError(f"Expected type '{dtype}' but got '{type(value)}'")
-    return None
+    return actual  # type: ignore[return-value]

tests/test_frame.py

@@ -1139,20 +1139,51 @@ def test_types_regressions() -> None:
 
 
 def test_read_csv() -> None:
-    if TYPE_CHECKING:  # skip pytest
-        #  https://github.com/microsoft/python-type-stubs/issues/87
-        df11: pd.DataFrame = pd.read_csv("foo")
-        df12: pd.DataFrame = pd.read_csv("foo", iterator=False)
-        df13: pd.DataFrame = pd.read_csv("foo", iterator=False, chunksize=None)
-        df14: TextFileReader = pd.read_csv("foo", chunksize=0)
-        df15: TextFileReader = pd.read_csv("foo", iterator=False, chunksize=0)
-        df16: TextFileReader = pd.read_csv("foo", iterator=True)
-        df17: TextFileReader = pd.read_csv("foo", iterator=True, chunksize=None)
-        df18: TextFileReader = pd.read_csv("foo", iterator=True, chunksize=0)
-        df19: TextFileReader = pd.read_csv("foo", chunksize=0)
+    with ensure_clean() as path:
+        Path(path).write_text("A,B\n1,2")
+        check(assert_type(pd.read_csv(path), pd.DataFrame), pd.DataFrame)
+        check(
+            assert_type(pd.read_csv(path, iterator=False), pd.DataFrame), pd.DataFrame
+        )
+        check(
+            assert_type(
+                pd.read_csv(path, iterator=False, chunksize=None), pd.DataFrame
+            ),
+            pd.DataFrame,
+        )
+
+        with check(
+            assert_type(pd.read_csv(path, chunksize=1), TextFileReader), TextFileReader
+        ):
+            pass
+        with check(
+            assert_type(pd.read_csv(path, iterator=False, chunksize=1), TextFileReader),
+            TextFileReader,
+        ):
+            pass
+        with check(
+            assert_type(pd.read_csv(path, iterator=True), TextFileReader),
+            TextFileReader,
+        ):
+            pass
+        with check(
+            assert_type(
+                pd.read_csv(path, iterator=True, chunksize=None), TextFileReader
+            ),
+            TextFileReader,
+        ):
+            pass
+        with check(
+            assert_type(pd.read_csv(path, iterator=True, chunksize=1), TextFileReader),
+            TextFileReader,
+        ):
+            pass
 
         # https://github.com/microsoft/python-type-stubs/issues/118
-        pd.read_csv("foo", storage_options=None)
+        check(
+            assert_type(pd.read_csv(path, storage_options=None), pd.DataFrame),
+            pd.DataFrame,
+        )
 
 
 def test_groupby_series_methods() -> None:
@@ -1358,6 +1389,7 @@ def test_set_columns() -> None:
     df = pd.DataFrame({"a": [1, 2, 3], "b": [0.0, 1, 1]})
     # Next line should work, but it is a mypy bug
     # https://github.com/python/mypy/issues/3004
+    # pyright doesn't need the ignore
     df.columns = ["c", "d"]  # type: ignore[assignment]
 
 

tests/test_interval.py

@@ -1,11 +1,12 @@
 from __future__ import annotations
 
-from typing import TYPE_CHECKING
-
 import pandas as pd
 from typing_extensions import assert_type
 
-from tests import check
+from tests import (
+    TYPE_CHECKING_INVALID_USAGE,
+    check,
+)
 
 
 def test_interval_init() -> None:
@@ -49,8 +50,8 @@ def test_interval_length() -> None:
     idres = i1 + pd.Timedelta(seconds=20)
 
     check(assert_type(idres, "pd.Interval[pd.Timestamp]"), pd.Interval, pd.Timestamp)
-    if TYPE_CHECKING:
-        20 in i1  # type: ignore[operator]
+    if TYPE_CHECKING_INVALID_USAGE:
+        20 in i1  # TODO both: ignore[operator]
         i1 + pd.Timestamp("2000-03-03")  # type: ignore[operator]
         i1 * 3  # type: ignore[operator]
         i1 * pd.Timedelta(seconds=20)  # type: ignore[operator]
@@ -68,9 +69,13 @@ def test_interval_length() -> None:
     check(assert_type(i2 * 4, "pd.Interval[int]"), pd.Interval, int)
     check(assert_type(i2 * 4.2, "pd.Interval[float]"), pd.Interval, float)
 
-    if TYPE_CHECKING:
-        pd.Timestamp("2001-01-02") in i2  # type: ignore[operator]
-        i2 + pd.Timedelta(seconds=20)  # type: ignore[operator]
+    if TYPE_CHECKING_INVALID_USAGE:
+        pd.Timestamp(
+            "2001-01-02"
+        ) in i2  # pyright: ignore[reportGeneralTypeIssues] # TODO mypy: ignore[operator]
+        i2 + pd.Timedelta(
+            seconds=20
+        )  # pyright: ignore[reportGeneralTypeIssues] # TODO mypy: ignore[operator]
 
     i3 = pd.Interval(13.2, 19.5)
     check(assert_type(i3.length, float), float)
@@ -82,6 +87,10 @@ def test_interval_length() -> None:
     check(assert_type(i3inres, bool), bool)
     check(assert_type(i3 + 3, "pd.Interval[float]"), pd.Interval, float)
     check(assert_type(i3 * 3, "pd.Interval[float]"), pd.Interval, float)
-    if TYPE_CHECKING:
-        pd.Timestamp("2001-01-02") in i3  # type: ignore[operator]
-        i3 + pd.Timedelta(seconds=20)  # type: ignore[operator]
+    if TYPE_CHECKING_INVALID_USAGE:
+        pd.Timestamp(
+            "2001-01-02"
+        ) in i3  # pyright: ignore[reportGeneralTypeIssues] # TODO mypy: ignore[operator]
+        i3 + pd.Timedelta(
+            seconds=20
+        )  # pyright: ignore[reportGeneralTypeIssues] # TODO mypy: ignore[operator]

tests/test_series.py

@@ -25,7 +25,10 @@
 
 from pandas._typing import Scalar
 
-from tests import check
+from tests import (
+    TYPE_CHECKING_INVALID_USAGE,
+    check,
+)
 
 if TYPE_CHECKING:
     from pandas._typing import np_ndarray_int  # noqa: F401
@@ -661,7 +664,7 @@ def add1(x: int) -> int:
     # inplace
     s6: None = pd.Series([1, 2, 3]).rename("A", inplace=True)
 
-    if TYPE_CHECKING:
+    if TYPE_CHECKING_INVALID_USAGE:
         s7 = pd.Series([1, 2, 3]).rename({1: [3, 4, 5]})  # type: ignore[dict-item]
 
 

tests/test_timefuncs.py

@@ -18,7 +18,10 @@
 from pandas._libs import NaTType
 from pandas._libs.tslibs import BaseOffset
 
-from tests import check
+from tests import (
+    TYPE_CHECKING_INVALID_USAGE,
+    check,
+)
 
 if TYPE_CHECKING:
     from pandas.core.series import (
@@ -201,10 +204,10 @@ def test_iso_calendar() -> None:
 def fail_on_adding_two_timestamps() -> None:
     s1 = pd.Series(pd.to_datetime(["2022-05-01", "2022-06-01"]))
     s2 = pd.Series(pd.to_datetime(["2022-05-15", "2022-06-15"]))
-    if TYPE_CHECKING:
-        ssum: pd.Series = s1 + s2  # type: ignore[operator]
+    if TYPE_CHECKING_INVALID_USAGE:
+        ssum: pd.Series = s1 + s2  # TODO both: ignore[operator]
         ts = pd.Timestamp("2022-06-30")
-        tsum: pd.Series = s1 + ts  # type: ignore[operator]
+        tsum: pd.Series = s1 + ts  # TODO both: ignore[operator]
 
 
 def test_dtindex_tzinfo() -> None: