chore(python): let read_csv take Sequence as columns, remove several type: ignore

py-polars/polars/_html.py

@@ -5,7 +5,10 @@
 import os
 from textwrap import dedent
 from types import TracebackType
-from typing import Iterable
+from typing import TYPE_CHECKING, Iterable
+
+if TYPE_CHECKING:
+    from polars.internals import DataFrame
 
 
 class Tag:
@@ -43,7 +46,7 @@ def __exit__(
 class HTMLFormatter:
     def __init__(
         self,
-        df: DataFrame,  # type: ignore[name-defined] # noqa: F821
+        df: DataFrame,
         max_cols: int = 75,
         max_rows: int = 40,
         from_series: bool = False,
@@ -75,7 +78,7 @@ def __init__(
 
     def write_header(self) -> None:
         """Write the header of an HTML table."""
-        shape = self.df.shape
+        shape: tuple[int, ...] = self.df.shape
         if self.series:
             shape = shape[:1]
 

py-polars/polars/internals/series/datetime.py

@@ -1,7 +1,7 @@
 from __future__ import annotations
 
 from datetime import date, datetime, time, timedelta
-from typing import TYPE_CHECKING
+from typing import TYPE_CHECKING, cast
 
 import polars.internals as pli
 from polars.internals.series.utils import expr_dispatch
@@ -52,8 +52,7 @@ def min(self) -> date | datetime | timedelta:
         datetime.datetime(2001, 1, 1, 0, 0)
 
         """
-        # we can ignore types because we are certain we get a logical type
-        return pli.wrap_s(self._s).min()  # type: ignore[return-value]
+        return cast("date | datetime | timedelta", pli.wrap_s(self._s).min())
 
     def max(self) -> date | datetime | timedelta:
         """
@@ -75,7 +74,7 @@ def max(self) -> date | datetime | timedelta:
         datetime.datetime(2001, 1, 3, 0, 0)
 
         """
-        return pli.wrap_s(self._s).max()  # type: ignore[return-value]
+        return cast("date | datetime | timedelta", pli.wrap_s(self._s).max())
 
     def median(self) -> date | datetime | timedelta | None:
         """

py-polars/polars/io.py

@@ -10,6 +10,7 @@
     BinaryIO,
     Callable,
     Mapping,
+    Sequence,
     TextIO,
     cast,
     overload,
@@ -58,7 +59,7 @@ def _check_arg_is_1byte(
 def read_csv(
     file: str | TextIO | BytesIO | Path | BinaryIO | bytes,
     has_header: bool = True,
-    columns: list[int] | list[str] | None = None,
+    columns: Sequence[int] | Sequence[str] | None = None,
     new_columns: list[str] | None = None,
     sep: str = ",",
     comment_char: str | None = None,
@@ -230,7 +231,7 @@ def read_csv(
         and not low_memory
         and null_values is None
     ):
-        include_columns = None
+        include_columns: Sequence[str] | None = None
 
         if columns:
             if not has_header:
@@ -1235,7 +1236,7 @@ def _read_excel_sheet(
     parser: Any,
     sheet_id: int | None,
     sheet_name: str | None,
-    read_csv_options: dict[str, Any] | None,
+    read_csv_options: dict[str, Any],
 ) -> DataFrame:
     csv_buffer = StringIO()
 
@@ -1246,7 +1247,7 @@ def _read_excel_sheet(
     csv_buffer.seek(0)
 
     # Parse CSV output.
-    return read_csv(csv_buffer, **read_csv_options)  # type: ignore[arg-type]
+    return read_csv(csv_buffer, **read_csv_options)
 
 
 def _get_delta_lake_table(
@@ -1660,7 +1661,7 @@ def scan_ds(ds: pa.dataset.dataset, allow_pyarrow_filter: bool = True) -> LazyFr
 def read_csv_batched(
     file: str | Path,
     has_header: bool = True,
-    columns: list[int] | list[str] | None = None,
+    columns: Sequence[int] | Sequence[str] | None = None,
     new_columns: list[str] | None = None,
     sep: str = ",",
     comment_char: str | None = None,

py-polars/polars/utils.py

@@ -8,7 +8,8 @@
 import sys
 import warnings
 from collections.abc import MappingView, Reversible, Sized
-from datetime import date, datetime, time, timedelta, timezone, tzinfo
+from datetime import date, datetime, time, timedelta, timezone
+from datetime import tzinfo as tzinfo_t
 from pathlib import Path
 from typing import (
     TYPE_CHECKING,
@@ -18,6 +19,7 @@
     Iterable,
     Sequence,
     TypeVar,
+    cast,
     overload,
 )
 
@@ -201,20 +203,22 @@ def range_to_slice(rng: range) -> slice:
 
 def handle_projection_columns(
     columns: Sequence[str] | Sequence[int] | str | None,
-) -> tuple[list[int] | None, list[str] | None]:
+) -> tuple[list[int] | None, Sequence[str] | None]:
     """Disambiguates between columns specified as integers vs. strings."""
     projection: list[int] | None = None
-    if columns:
+    new_columns: Sequence[str] | None = None
+    if columns is not None:
         if isinstance(columns, str):
-            columns = [columns]
+            new_columns = [columns]
         elif is_int_sequence(columns):
             projection = list(columns)
-            columns = None
         elif not is_str_sequence(columns):
             raise ValueError(
                 "'columns' arg should contain a list of all integers or all strings"
                 " values."
             )
+        else:
+            new_columns = columns
         if columns and len(set(columns)) != len(columns):
             raise ValueError(
                 f"'columns' arg should only have unique values. Got '{columns}'."
@@ -223,7 +227,7 @@ def handle_projection_columns(
             raise ValueError(
                 f"'columns' arg should only have unique values. Got '{projection}'."
             )
-    return projection, columns  # type: ignore[return-value]
+    return projection, new_columns
 
 
 def _to_python_time(value: int) -> time:
@@ -317,7 +321,7 @@ def _to_python_datetime(
 # cache here as we have a single tz per column
 # and this function will be called on every conversion
 @functools.lru_cache(16)
-def _parse_fixed_tz_offset(offset: str) -> tzinfo:
+def _parse_fixed_tz_offset(offset: str) -> tzinfo_t:
     try:
         # use fromisoformat to parse the offset
         dt_offset = datetime.fromisoformat("2000-01-01T00:00:00" + offset)
@@ -328,16 +332,16 @@ def _parse_fixed_tz_offset(offset: str) -> tzinfo:
     except ValueError:
         raise ValueError(f"Offset: {offset} not understood.") from None
 
-    return dt_offset.tzinfo  # type: ignore[return-value]
+    return cast(tzinfo_t, dt_offset.tzinfo)
 
 
 def _localize(dt: datetime, tz: str) -> datetime:
     # zone info installation should already be checked
     try:
-        tzinfo = ZoneInfo(tz)
+        tzinfo: ZoneInfo | tzinfo_t = ZoneInfo(tz)
     except zoneinfo.ZoneInfoNotFoundError:
         # try fixed offset, which is not supported by ZoneInfo
-        tzinfo = _parse_fixed_tz_offset(tz)  # type: ignore[assignment]
+        tzinfo = _parse_fixed_tz_offset(tz)
 
     return dt.astimezone(tzinfo)