feat: add support for Pandas 1.5 (#1076)

requirements.txt

@@ -1,6 +1,6 @@
 joblib~=1.2.0 # 1.1.0
 scipy>=1.4.1, <1.10 
-pandas>1.1, <1.5, !=1.4.0
+pandas>1.1, <1.6, !=1.4.0
 matplotlib>=3.2, <3.6
 pydantic>=1.8.1, <1.11
 PyYAML>=5.0.0, <6.1
@@ -23,4 +23,4 @@ tqdm>=4.48.2, <4.65
 seaborn>=0.10.1, <0.13
 multimethod>=1.4, <1.10
 # metrics
-statsmodels>=0.13.2, <0.14
+statsmodels>=0.13.2, <0.14

src/pandas_profiling/model/correlations.py

@@ -5,9 +5,14 @@
 import numpy as np
 import pandas as pd
 from multimethod import multimethod
-from pandas.core.base import DataError
 
 from pandas_profiling.config import Settings
+from pandas_profiling.utils.compat import pandas_version_info
+
+if pandas_version_info() >= (1, 5):
+    from pandas.errors import DataError
+else:
+    from pandas.core.base import DataError
 
 
 class Correlation:

src/pandas_profiling/model/pandas/correlations_pandas.py

@@ -73,12 +73,16 @@ def pandas_cramers_compute(
 ) -> Optional[pd.DataFrame]:
     threshold = config.categorical_maximum_correlation_distinct
 
-    categoricals = {
-        key
-        for key, value in summary.items()
-        if value["type"] in {"Categorical", "Boolean"}
-        and value["n_distinct"] <= threshold
-    }
+    # `index` and `columns` must not be a set since Pandas 1.5,
+    # so convert it to a list. The order of the list is arbitrary.
+    categoricals = list(
+        {
+            key
+            for key, value in summary.items()
+            if value["type"] in {"Categorical", "Boolean"}
+            and value["n_distinct"] <= threshold
+        }
+    )
 
     if len(categoricals) <= 1:
         return None

src/pandas_profiling/model/pandas/describe_numeric_pandas.py

@@ -2,7 +2,13 @@
 
 import numpy as np
 import pandas as pd
-from pandas.core.arrays.integer import _IntegerDtype
+
+from pandas_profiling.utils.compat import pandas_version_info
+
+if pandas_version_info() >= (1, 5):
+    from pandas.core.arrays.integer import IntegerDtype
+else:
+    from pandas.core.arrays.integer import _IntegerDtype as IntegerDtype
 
 from pandas_profiling.config import Settings
 from pandas_profiling.model.summary_algorithms import (
@@ -95,7 +101,7 @@ def pandas_describe_numeric_1d(
 
     stats = summary
 
-    if isinstance(series.dtype, _IntegerDtype):
+    if isinstance(series.dtype, IntegerDtype):
         stats.update(numeric_stats_pandas(series))
         present_values = series.astype(str(series.dtype).lower())
         finite_values = present_values

src/pandas_profiling/model/typeset_relations.py

@@ -52,9 +52,7 @@ def to_category(series: pd.Series, state: dict) -> pd.Series:
     if hasnans:
         val = val.replace("nan", np.nan)
 
-    if int(pd.__version__.split(".")[0]) >= 1:
-        val = val.astype("string")
-    return val
+    return val.astype("string")
 
 
 @series_handle_nulls
@@ -87,7 +85,7 @@ def category_to_numeric(series: pd.Series, state: dict) -> pd.Series:
     return pd.to_numeric(series, errors="coerce")
 
 
-hasnan_bool_name = "boolean" if int(pd.__version__.split(".")[0]) >= 1 else "Bool"
+hasnan_bool_name = "boolean"
 
 
 def to_bool(series: pd.Series) -> pd.Series:

src/pandas_profiling/utils/compat.py

@@ -0,0 +1,14 @@
+"""Utility functions for (version) compatibility"""
+from functools import lru_cache
+from typing import Tuple
+
+import pandas as pd
+
+
+@lru_cache(maxsize=1)
+def pandas_version_info() -> Tuple[int, ...]:
+    """
+    Get the Pandas version as a tuple of integers,
+    akin to `sys.version_info` for the Python version.
+    """
+    return tuple(int(s) for s in pd.__version__.split("."))

tests/issues/test_issue523.py

@@ -3,14 +3,10 @@
 https://github.com/pandas-profiling/pandas-profiling/issues/XXX
 """
 import pandas as pd
-import pytest
 
 from pandas_profiling import ProfileReport
 
 
-@pytest.mark.skipif(
-    int(pd.__version__.split(".")[0]) < 1, reason="requires pandas 1 or higher"
-)
 def test_issue523():
     # https://github.com/pandas-dev/pandas/issues/33803
 

tests/issues/test_issue545.py

@@ -7,14 +7,11 @@
 import pytest
 
 from pandas_profiling import ProfileReport
-
-
-def pandas_version():
-    return tuple(int(s) for s in pd.__version__.split("."))
+from pandas_profiling.utils.compat import pandas_version_info
 
 
 @pytest.mark.skipif(
-    pandas_version() <= (1, 1, 0), reason="requires pandas 1.1.1 or higher"
+    pandas_version_info() <= (1, 1, 0), reason="requires pandas 1.1.1 or higher"
 )
 def test_issue545(get_data_file):
     file_name = get_data_file(

tests/unit/test_typeset_default.py

@@ -1,6 +1,5 @@
 import os
 
-import pandas as pd
 import pytest
 from visions.test.series import get_series
 from visions.test.utils import (
@@ -16,13 +15,6 @@
 from pandas_profiling.model.typeset import ProfilingTypeSet
 from tests.unit.test_utils import patch_arg
 
-if int(pd.__version__.split(".")[0]) < 1:
-    from visions.dtypes.boolean import BoolDtype  # noqa: F401
-
-    btype = "Bool"
-else:
-    btype = "boolean"
-
 base_path = os.path.abspath(os.path.dirname(__file__))
 
 series = get_series()
@@ -100,6 +92,7 @@
         "str_complex_nan",
         "all_null_empty_str",
         "py_datetime_str",
+        "string_dtype_series",
     },
     Boolean: {
         "bool_series",
@@ -116,53 +109,49 @@
         "timestamp_series_nat",
         "date_series_nat",
     },
-}
-
-if int(pd.__version__[0]) >= 1:
-    contains_map[Categorical].add("string_dtype_series")
-
-contains_map[Unsupported] = {
-    "module",
-    "nan_series",
-    "nan_series_2",
-    "timedelta_series",
-    "timedelta_series_nat",
-    "timedelta_negative",
-    "path_series_linux",
-    "path_series_linux_missing",
-    "path_series_windows",
-    "url_series",
-    "url_nan_series",
-    "url_none_series",
-    "file_test_py",
-    "file_mixed_ext",
-    "file_test_py_missing",
-    "image_png",
-    "image_png_missing",
-    "image_png",
-    "image_png_missing",
-    "uuid_series",
-    "uuid_series_missing",
-    "mixed_list[str,int]",
-    "mixed_dict",
-    "callable",
-    "mixed_integer",
-    "mixed_list",
-    "date",
-    "time",
-    "empty",
-    "empty_bool",
-    "empty_float",
-    "empty_object",
-    "empty_int64",
-    "ip",
-    "ip_missing",
-    "ip_mixed_v4andv6",
-    "email_address_missing",
-    "email_address",
-    "all_null_none",
-    "all_null_nan",
-    "all_null_nat",
+    Unsupported: {
+        "module",
+        "nan_series",
+        "nan_series_2",
+        "timedelta_series",
+        "timedelta_series_nat",
+        "timedelta_negative",
+        "path_series_linux",
+        "path_series_linux_missing",
+        "path_series_windows",
+        "url_series",
+        "url_nan_series",
+        "url_none_series",
+        "file_test_py",
+        "file_mixed_ext",
+        "file_test_py_missing",
+        "image_png",
+        "image_png_missing",
+        "image_png",
+        "image_png_missing",
+        "uuid_series",
+        "uuid_series_missing",
+        "mixed_list[str,int]",
+        "mixed_dict",
+        "callable",
+        "mixed_integer",
+        "mixed_list",
+        "date",
+        "time",
+        "empty",
+        "empty_bool",
+        "empty_float",
+        "empty_object",
+        "empty_int64",
+        "ip",
+        "ip_missing",
+        "ip_mixed_v4andv6",
+        "email_address_missing",
+        "email_address",
+        "all_null_none",
+        "all_null_nan",
+        "all_null_nat",
+    },
 }
 
 
@@ -293,9 +282,8 @@ def test_contains(name, series, contains_type, member):
     "all_null_none": Unsupported,
     "complex_series_py_float": Numeric,
     "all_null_nan": Unsupported,
+    "string_dtype_series": Categorical,
 }
-if int(pd.__version__[0]) >= 1:
-    inference_map["string_dtype_series"] = Categorical
 
 
 @pytest.mark.parametrize(

tests/unit/test_utils.py

@@ -3,6 +3,7 @@
 import pandas as pd
 import pytest
 
+from pandas_profiling.utils.compat import pandas_version_info
 from pandas_profiling.utils.dataframe import (
     expand_mixed,
     read_pandas,
@@ -37,10 +38,17 @@ def test_read_pandas_csv():
 
 def test_read_pandas_json():
     p = Path("dataframe.json")
-    with pytest.raises(ValueError) as e:
+
+    expected_error, expected_message = (
+        (FileNotFoundError, "File dataframe.json does not exist")
+        if pandas_version_info() >= (1, 5)
+        else (ValueError, "Expected object or value")
+    )
+
+    with pytest.raises(expected_error) as e:
         read_pandas(p)
 
-    assert str(e.value) == "Expected object or value"
+    assert str(e.value) == expected_message
 
 
 def test_warning():