Review

Author: jonashaag

pandas/core/describe.py

@@ -35,6 +35,8 @@
 
 from pandas.core.reshape.concat import concat
 
+from pandas.io.formats.format import format_percentiles
+
 if TYPE_CHECKING:
     from pandas import (
         DataFrame,
@@ -228,8 +230,6 @@ def describe_numeric_1d(series: Series, percentiles: Sequence[float]) -> Series:
     """
     from pandas import Series
 
-    from pandas.io.formats.format import format_percentiles
-
     # error: Argument 1 to "format_percentiles" has incompatible type "Sequence[float]";
     # expected "Union[ndarray, List[Union[int, float]], List[float], List[Union[str,
     # float]]]"
@@ -337,8 +337,6 @@ def describe_timestamp_1d(data: Series, percentiles: Sequence[float]) -> Series:
     # GH-30164
     from pandas import Series
 
-    from pandas.io.formats.format import format_percentiles
-
     # error: Argument 1 to "format_percentiles" has incompatible type "Sequence[float]";
     # expected "Union[ndarray, List[Union[int, float]], List[float], List[Union[str,
     # float]]]"

pandas/io/common.py

@@ -54,7 +54,7 @@
 
 from pandas.core.dtypes.common import is_file_like
 
-from pandas.core import generic
+from pandas.core.shared_docs import _shared_docs
 
 _VALID_URLS = set(uses_relative + uses_netloc + uses_params)
 _VALID_URLS.discard("")
@@ -248,10 +248,7 @@ def is_fsspec_url(url: FilePath | BaseBuffer) -> bool:
     )
 
 
-@doc(
-    compression_options=generic._shared_docs["compression_options"]
-    % "filepath_or_buffer"
-)
+@doc(compression_options=_shared_docs["compression_options"] % "filepath_or_buffer")
 def _get_filepath_or_buffer(
     filepath_or_buffer: FilePath | BaseBuffer,
     encoding: str = "utf-8",
@@ -497,10 +494,7 @@ def get_compression_method(
     return compression_method, compression_args
 
 
-@doc(
-    compression_options=generic._shared_docs["compression_options"]
-    % "filepath_or_buffer"
-)
+@doc(compression_options=_shared_docs["compression_options"] % "filepath_or_buffer")
 def infer_compression(
     filepath_or_buffer: FilePath | BaseBuffer, compression: str | None
 ) -> str | None:
@@ -604,7 +598,7 @@ def get_handle(
     ...
 
 
-@doc(compression_options=generic._shared_docs["compression_options"] % "path_or_buf")
+@doc(compression_options=_shared_docs["compression_options"] % "path_or_buf")
 def get_handle(
     path_or_buf: FilePath | BaseBuffer,
     mode: str,
@@ -1136,18 +1130,20 @@ def _is_binary_mode(handle: FilePath | BaseBuffer, mode: str) -> bool:
     )
 
 
-def _get_binary_io_classes() -> tuple[type, ...]:
+def _get_binary_io_classes(_cache: list[type] = []) -> tuple[type, ...]:
     """IO classes that that expect bytes"""
-    binary_classes: tuple[type, ...] = (BufferedIOBase, RawIOBase)
-
-    # python-zstandard doesn't use any of the builtin base classes; instead we
-    # have to use the `zstd.ZstdDecompressionReader` class for isinstance checks.
-    # Unfortunately `zstd.ZstdDecompressionReader` isn't exposed by python-zstandard
-    # so we have to get it from a `zstd.ZstdDecompressor` instance.
-    # See also https://github.com/indygreg/python-zstandard/pull/165.
-    zstd = import_optional_dependency("zstandard", errors="ignore")
-    if zstd is not None:
-        with zstd.ZstdDecompressor().stream_reader(b"") as reader:
-            binary_classes += (type(reader),)
-
-    return binary_classes
+    if not _cache:
+        binary_classes: tuple[type, ...] = (BufferedIOBase, RawIOBase)
+
+        # python-zstandard doesn't use any of the builtin base classes; instead we
+        # have to use the `zstd.ZstdDecompressionReader` class for isinstance checks.
+        # Unfortunately `zstd.ZstdDecompressionReader` isn't exposed by python-zstandard
+        # so we have to get it from a `zstd.ZstdDecompressor` instance.
+        # See also https://github.com/indygreg/python-zstandard/pull/165.
+        zstd = import_optional_dependency("zstandard", errors="ignore")
+        if zstd is not None:
+            with zstd.ZstdDecompressor().stream_reader(b"") as reader:
+                binary_classes += (type(reader),)
+
+        _cache[:] = binary_classes
+    return tuple(_cache)

pandas/io/formats/xml.py

@@ -19,7 +19,7 @@
 
 from pandas.core.dtypes.common import is_list_like
 
-from pandas.core import generic
+from pandas.core import _shared_docs
 from pandas.core.frame import DataFrame
 
 from pandas.io.common import get_handle
@@ -29,7 +29,7 @@
 )
 
 
-@doc(compression_options=generic._shared_docs["compression_options"] % "path_or_buffer")
+@doc(compression_options=_shared_docs["compression_options"] % "path_or_buffer")
 class BaseXMLFormatter:
     """
     Subclass for formatting data in XML.

pandas/tests/io/conftest.py

@@ -10,6 +10,7 @@
     is_platform_mac,
     is_platform_windows,
 )
+import pandas.util._test_decorators as td
 
 import pandas._testing as tm
 
@@ -184,3 +185,29 @@ def add_tips_files(bucket_name):
     while cli.list_buckets()["Buckets"] and timeout > 0:
         time.sleep(0.1)
         timeout -= 0.1
+
+
+_compression_formats_params = [
+    (".no_compress", None),
+    ("", None),
+    (".gz", "gzip"),
+    (".GZ", "gzip"),
+    (".bz2", "bz2"),
+    (".BZ2", "bz2"),
+    (".zip", "zip"),
+    (".ZIP", "zip"),
+    (".xz", "xz"),
+    (".XZ", "xz"),
+    pytest.param((".zst", "zstd"), marks=td.skip_if_no("zstandard")),
+    pytest.param((".ZST", "zstd"), marks=td.skip_if_no("zstandard")),
+]
+
+
+@pytest.fixture(params=_compression_formats_params[1:])
+def compression_format(request):
+    return request.param
+
+
+@pytest.fixture(params=_compression_formats_params)
+def compression_ext(request):
+    return request.param[0]

pandas/tests/io/test_common.py

@@ -48,25 +48,6 @@ def __fspath__(self):
 HERE = os.path.abspath(os.path.dirname(__file__))
 
 
-@pytest.fixture(
-    params=[
-        ("", None),
-        (".gz", "gzip"),
-        (".GZ", "gzip"),
-        (".bz2", "bz2"),
-        (".BZ2", "bz2"),
-        (".zip", "zip"),
-        (".ZIP", "zip"),
-        (".xz", "xz"),
-        (".XZ", "xz"),
-        pytest.param((".zst", "zstd"), marks=td.skip_if_no("zstandard")),
-        pytest.param((".ZST", "zstd"), marks=td.skip_if_no("zstandard")),
-    ]
-)
-def compression_format(request):
-    return request.param
-
-
 # https://github.com/cython/cython/issues/1720
 @pytest.mark.filterwarnings("ignore:can't resolve package:ImportWarning")
 class TestCommonIOCapabilities:

pandas/tests/io/test_pickle.py

@@ -286,26 +286,6 @@ def get_random_path():
     return f"__{tm.rands(10)}__.pickle"
 
 
-@pytest.fixture(
-    params=[
-        "",
-        ".no_compress",
-        ".gz",
-        ".GZ",
-        ".bz2",
-        ".BZ2",
-        ".zip",
-        ".ZIP",
-        ".xz",
-        ".XZ",
-        pytest.param(".zst", marks=td.skip_if_no("zstandard")),
-        pytest.param(".ZST", marks=td.skip_if_no("zstandard")),
-    ]
-)
-def compression_ext(request):
-    return request.param
-
-
 class TestCompression:
 
     _extension_to_compression = {