PERF: lazify IO imports

pandas/__init__.py

@@ -1,3 +1,4 @@
+# pylint: disable=undefined-all-variable
 from __future__ import annotations
 
 __docformat__ = "restructuredtext"
@@ -113,8 +114,6 @@
 from pandas.tseries.api import infer_freq
 from pandas.tseries import offsets
 
-from pandas.core.computation.api import eval
-
 from pandas.core.reshape.api import (
     concat,
     lreshape,
@@ -133,43 +132,8 @@
 )
 
 from pandas import api, arrays, errors, io, plotting, tseries
-from pandas import testing
 from pandas.util._print_versions import show_versions
 
-from pandas.io.api import (
-    # excel
-    ExcelFile,
-    ExcelWriter,
-    read_excel,
-    # parsers
-    read_csv,
-    read_fwf,
-    read_table,
-    # pickle
-    read_pickle,
-    to_pickle,
-    # pytables
-    HDFStore,
-    read_hdf,
-    # sql
-    read_sql,
-    read_sql_query,
-    read_sql_table,
-    # misc
-    read_clipboard,
-    read_parquet,
-    read_orc,
-    read_feather,
-    read_gbq,
-    read_html,
-    read_xml,
-    read_json,
-    read_stata,
-    read_sas,
-    read_spss,
-)
-
-from pandas.io.json._normalize import json_normalize
 
 from pandas.util._tester import test
 
@@ -182,6 +146,92 @@
 del get_versions, v
 
 
+def __getattr__(key: str):
+    # lazy imports to speed up 'import pandas as pd'
+    if key == "eval":
+        from pandas.core.computation.api import eval
+
+        return eval
+    elif key == "testing":
+        import pandas.testing  # pylint: disable=redefined-outer-name
+
+        return pandas.testing
+
+    elif key in {
+        "ExcelFile",
+        "ExcelWriter",
+        "read_excel",
+        "read_csv",
+        "read_fwf",
+        "read_table",
+        "read_pickle",
+        "to_pickle",
+        "HDFStore",
+        "read_hdf",
+        "read_sql",
+        "read_sql_query",
+        "read_sql_table",
+        "read_clipboard",
+        "read_parquet",
+        "read_orc",
+        "read_feather",
+        "read_gbq",
+        "read_html",
+        "read_xml",
+        "read_json",
+        "read_stata",
+        "read_sas",
+        "read_spss",
+    }:
+        # Workaround to avoid false-positive in "inconsistent-namespace-usage"
+        #  complaining "Found both 'pandas.io' and 'io' "
+        import importlib
+
+        namespace = importlib.import_module(".io.api", package="pandas")
+        return getattr(namespace, key)
+    elif key == "json_normalize":
+        from pandas.io.json._normalize import json_normalize
+
+        return json_normalize
+    raise AttributeError(f"module 'pandas' has no attribute '{key}'")
+
+
+def __dir__() -> list[str]:
+    # include lazy imports defined in __getattr__ in dir()
+    base = list(globals().keys())
+    result = (
+        base
+        + [
+            "ExcelFile",
+            "ExcelWriter",
+            "read_excel",
+            "read_csv",
+            "read_fwf",
+            "read_table",
+            "read_pickle",
+            "to_pickle",
+            "HDFStore",
+            "read_hdf",
+            "read_sql",
+            "read_sql_query",
+            "read_sql_table",
+            "read_clipboard",
+            "read_parquet",
+            "read_orc",
+            "read_feather",
+            "read_gbq",
+            "read_html",
+            "read_xml",
+            "read_json",
+            "read_stata",
+            "read_sas",
+            "read_spss",
+        ]
+        + ["eval", "json_normalize", "testing"]
+    )
+    return result
+
+
 # module level doc-string
 __doc__ = """
 pandas - a powerful data analysis and manipulation library for Python
@@ -226,7 +276,7 @@
 # Use __all__ to let type checkers know what is part of the public API.
 # Pandas is not (yet) a py.typed library: the public API is determined
 # based on the documentation.
-__all__ = [
+__all__ = [  # pyright: ignore[reportUnsupportedDunderAll]
     "ArrowDtype",
     "BooleanDtype",
     "Categorical",
@@ -236,13 +286,13 @@
     "DateOffset",
     "DatetimeIndex",
     "DatetimeTZDtype",
-    "ExcelFile",
-    "ExcelWriter",
+    "ExcelFile",  # pyright: ignore[reportUnsupportedDunderAll]
+    "ExcelWriter",  # pyright: ignore[reportUnsupportedDunderAll]
     "Flags",
     "Float32Dtype",
     "Float64Dtype",
     "Grouper",
-    "HDFStore",
+    "HDFStore",  # pyright: ignore[reportUnsupportedDunderAll]
     "Index",
     "IndexSlice",
     "Int16Dtype",
@@ -280,7 +330,7 @@
     "date_range",
     "describe_option",
     "errors",
-    "eval",
+    "eval",  # pyright: ignore[reportUnsupportedDunderAll]
     "factorize",
     "get_dummies",
     "from_dummies",
@@ -290,7 +340,7 @@
     "io",
     "isna",
     "isnull",
-    "json_normalize",
+    "json_normalize",  # pyright: ignore[reportUnsupportedDunderAll]
     "lreshape",
     "melt",
     "merge",
@@ -306,36 +356,36 @@
     "pivot_table",
     "plotting",
     "qcut",
-    "read_clipboard",
-    "read_csv",
-    "read_excel",
-    "read_feather",
-    "read_fwf",
-    "read_gbq",
-    "read_hdf",
-    "read_html",
-    "read_json",
-    "read_orc",
-    "read_parquet",
-    "read_pickle",
-    "read_sas",
-    "read_spss",
-    "read_sql",
-    "read_sql_query",
-    "read_sql_table",
-    "read_stata",
-    "read_table",
-    "read_xml",
+    "read_clipboard",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_csv",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_excel",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_feather",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_fwf",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_gbq",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_hdf",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_html",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_json",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_orc",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_parquet",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_pickle",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_sas",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_spss",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_sql",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_sql_query",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_sql_table",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_stata",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_table",  # pyright: ignore[reportUnsupportedDunderAll]
+    "read_xml",  # pyright: ignore[reportUnsupportedDunderAll]
     "reset_option",
     "set_eng_float_format",
     "set_option",
     "show_versions",
     "test",
-    "testing",
+    "testing",  # pyright: ignore[reportUnsupportedDunderAll]
     "timedelta_range",
     "to_datetime",
     "to_numeric",
-    "to_pickle",
+    "to_pickle",  # pyright: ignore[reportUnsupportedDunderAll]
     "to_timedelta",
     "tseries",
     "unique",

pandas/api/interchange/__init__.py

@@ -1,8 +1,23 @@
+# pylint: disable=undefined-all-variable
 """
 Public API for DataFrame interchange protocol.
 """
 
-from pandas.core.interchange.dataframe_protocol import DataFrame
-from pandas.core.interchange.from_dataframe import from_dataframe
 
-__all__ = ["from_dataframe", "DataFrame"]
+def __getattr__(key: str):
+    # lazy imports to speed 'import pandas as pd'
+    if key == "DataFrame":
+        from pandas.core.interchange.dataframe_protocol import DataFrame
+
+        return DataFrame
+    elif key == "from_dataframe":
+        from pandas.core.interchange.from_dataframe import from_dataframe
+
+        return from_dataframe
+    raise AttributeError(key)
+
+
+__all__ = [
+    "from_dataframe",  # pyright: ignore[reportUnsupportedDunderAll]
+    "DataFrame",  # pyright: ignore[reportUnsupportedDunderAll]
+]

pandas/api/typing/__init__.py

@@ -1,6 +1,8 @@
+# pylint: disable=undefined-all-variable
 """
 Public API classes that store intermediate results useful for type-hinting.
 """
+from __future__ import annotations
 
 from pandas.core.groupby import (
     DataFrameGroupBy,
@@ -25,8 +27,27 @@
 
 # TODO: Can't import Styler without importing jinja2
 # from pandas.io.formats.style import Styler
-from pandas.io.json._json import JsonReader
-from pandas.io.stata import StataReader
+
+
+def __getattr__(key: str):
+    if key == "JsonReader":
+        from pandas.io.json._json import JsonReader
+
+        return JsonReader
+    elif key == "StataReader":
+        from pandas.io.stata import StataReader
+
+        return StataReader
+    else:
+        raise AttributeError(f"module 'pandas.api.typing' has no attribute '{key}'")
+
+
+def __dir__() -> list[str]:
+    # include lazy imports defined in __getattr__ in dir()
+    base = list(globals().keys())
+    result = base + ["JsonReader", "StataReader"]
+    return sorted(result)
+
 
 __all__ = [
     "DataFrameGroupBy",
@@ -35,13 +56,13 @@
     "ExpandingGroupby",
     "ExponentialMovingWindow",
     "ExponentialMovingWindowGroupby",
-    "JsonReader",
+    "JsonReader",  # pyright: ignore[reportUnsupportedDunderAll]
     "PeriodIndexResamplerGroupby",
     "Resampler",
     "Rolling",
     "RollingGroupby",
     "SeriesGroupBy",
-    "StataReader",
+    "StataReader",  # pyright: ignore[reportUnsupportedDunderAll]
     # See TODO above
     # "Styler",
     "TimedeltaIndexResamplerGroupby",