V0.7.11

CHANGELOG.md

@@ -1,5 +1,8 @@
 # Change Log
 
+## 0.7.11
+* Fixing dependency issue ([#170](https://github.com/shakedzy/dython/issues/170))
+
 ## 0.7.10
 * _Dython now officially supports only Python 3.10 or above_
 * Fix a bug in `model_utils.metric_graph`

VERSION

@@ -1 +1 @@
-0.7.10
+0.7.11

dython/__init__.py

@@ -1,13 +1,8 @@
-from . import nominal, model_utils, sampling, data_utils
+from importlib.metadata import version
 from ._private import set_is_jupyter
 
+__all__ = ["__version__", "__dist_name__"]
+__dist_name__ = "dython"
+__version__ = version(__dist_name__)
 
-def _get_version_from_setuptools():
-    from pkg_resources import get_distribution
-
-    return get_distribution("dython").version
-
-
-__all__ = ["__version__"]
-__version__ = _get_version_from_setuptools()
 set_is_jupyter()

dython/_private.py

@@ -2,15 +2,14 @@
 import numpy as np
 import pandas as pd
 import matplotlib.pyplot as plt
-from numpy.typing import NDArray
-from typing import Optional, Any, Tuple, Union, List, Literal
-from .typing import Number, OneDimArray
+from typing import Any, Literal, cast, overload, Type
+from .typing import OneDimArray, TwoDimArray
 
 
 IS_JUPYTER: bool = False
 
 
-def set_is_jupyter(force_to: Optional[bool] = None) -> None:
+def set_is_jupyter(force_to: bool | None = None) -> None:
     global IS_JUPYTER
     if force_to is not None:
         IS_JUPYTER = force_to
@@ -27,33 +26,63 @@ def plot_or_not(plot: bool) -> None:
             plt.close(fig)  
 
 
+@overload
 def convert(
-    data: Union[List[Number], NDArray, pd.DataFrame],
-    to: Literal["array", "list", "dataframe"],
+    data: OneDimArray | TwoDimArray,
+    to: Type[np.ndarray],
     copy: bool = True,
-) -> Union[List[Number], NDArray, pd.DataFrame]:
+) -> np.ndarray:
+    ...
+
+@overload
+def convert(
+    data: OneDimArray | TwoDimArray,
+    to: Type[list],
+    copy: bool = True,
+) -> list:
+    ...
+
+@overload
+def convert(
+    data: OneDimArray | TwoDimArray,
+    to: Type[pd.DataFrame],
+    copy: bool = True,
+) -> pd.DataFrame:
+    ...
+
+def convert(
+    data: OneDimArray | TwoDimArray,
+    to: Type[list | pd.DataFrame | np.ndarray],
+    copy: bool = True,
+) -> list | pd.DataFrame | np.ndarray:
+
     converted = None
-    if to == "array":
+
+    if to == np.ndarray:
         if isinstance(data, np.ndarray):
             converted = data.copy() if copy else data
         elif isinstance(data, pd.Series):
             converted = data.values
         elif isinstance(data, list):
             converted = np.array(data)
         elif isinstance(data, pd.DataFrame):
-            converted = data.values()  # type: ignore
-    elif to == "list":
+            converted = data.values             
+        converted = cast(np.ndarray, converted)
+
+    elif to == list:
         if isinstance(data, list):
             converted = data.copy() if copy else data
         elif isinstance(data, pd.Series):
             converted = data.values.tolist()
         elif isinstance(data, np.ndarray):
             converted = data.tolist()
-    elif to == "dataframe":
+
+    elif to == pd.DataFrame:
         if isinstance(data, pd.DataFrame):
             converted = data.copy(deep=True) if copy else data
         elif isinstance(data, np.ndarray):
             converted = pd.DataFrame(data)
+
     else:
         raise ValueError("Unknown data conversion: {}".format(to))
     if converted is None:
@@ -63,12 +92,14 @@ def convert(
             )
         )
     else:
-        return converted  # type: ignore
+        return converted
 
 
 def remove_incomplete_samples(
-    x: Union[List[Any], OneDimArray], y: Union[List[Any], OneDimArray]
-) -> Tuple[Union[List[Any], OneDimArray], Union[List[Any], OneDimArray]]:
+    x: OneDimArray, 
+    y: OneDimArray,
+) -> tuple[OneDimArray, OneDimArray]:
+
     x = [v if v is not None else np.nan for v in x]
     y = [v if v is not None else np.nan for v in y]
     arr = np.array([x, y]).transpose()
@@ -80,10 +111,11 @@ def remove_incomplete_samples(
 
 
 def replace_nan_with_value(
-    x: Union[List[Any], OneDimArray],
-    y: Union[List[Any], OneDimArray],
+    x: OneDimArray,
+    y: OneDimArray,
     value: Any,
-) -> Tuple[NDArray, NDArray]:
+) -> tuple[np.ndarray, np.ndarray]:
+
     x = np.array(
         [v if v == v and v is not None else value for v in x]
     )  # NaN != NaN

dython/data_utils.py

@@ -1,9 +1,9 @@
 import numpy as np
 import pandas as pd
 import matplotlib.pyplot as plt
-from typing import Optional, Tuple, List, Any, Union
-from numpy.typing import NDArray
-from .typing import Number, TwoDimArray
+from matplotlib.axes._axes import Axes
+from typing import Any, Sequence, cast
+from .typing import TwoDimArray, OneDimArray
 from ._private import convert, plot_or_not
 
 
@@ -16,9 +16,9 @@
 
 
 def one_hot_encode(
-    array: Union[List[Union[Number, str]], NDArray],
-    classes: Optional[int] = None,
-) -> NDArray:
+    array: OneDimArray,
+    classes: int | None = None,
+) -> np.ndarray:
     """
     One-hot encode a 1D array.
     Based on this StackOverflow answer: https://stackoverflow.com/a/29831596/5863503
@@ -41,15 +41,16 @@ def one_hot_encode(
            [1., 0., 0., 0., 0., 0.],
            [0., 0., 0., 0., 0., 1.]])
     """
-    arr: NDArray = convert(array, "array").astype(int)  # type: ignore
+    arr = convert(array, np.ndarray).astype(int)
     if not len(arr.shape) == 1:
         raise ValueError(
             f"array must have only one dimension, but has shape: {arr.shape}"
         )
     if arr.min() < 0:
         raise ValueError("array cannot contain negative values")
     classes = classes if classes is not None else arr.max() + 1
-    h = np.zeros((arr.size, classes))  # type: ignore
+    classes = cast(int, classes)
+    h = np.zeros((arr.size, classes))
     h[np.arange(arr.size), arr] = 1
     return h
 
@@ -58,14 +59,14 @@ def split_hist(
     dataset: pd.DataFrame,
     values: str,
     split_by: str,
-    title: Optional[str] = "",
-    xlabel: Optional[str] = "",
-    ylabel: Optional[str] = None,
-    figsize: Optional[Tuple[int, int]] = None,
-    legend: Optional[str] = "best",
+    title: str | None = "",
+    xlabel: str | None = "",
+    ylabel: str | None = None,
+    figsize: tuple[int, int] | None = None,
+    legend: str | None = "best",
     plot: bool = True,
     **hist_kwargs,
-) -> plt.Axes:
+) -> Axes:  
     """
     Plot a histogram of values from a given dataset, split by the values of a chosen column
 
@@ -125,8 +126,9 @@ def split_hist(
 
 
 def identify_columns_by_type(
-    dataset: TwoDimArray, include: List[str]
-) -> List[Any]:
+    dataset: TwoDimArray, 
+    include: Sequence[str]
+) -> list[Any]:
     """
     Given a dataset, identify columns of the types requested.
 
@@ -147,8 +149,8 @@ def identify_columns_by_type(
     ['col2', 'col3']
 
     """
-    df: pd.DataFrame = convert(dataset, "dataframe")  # type: ignore
-    columns = list(df.select_dtypes(include=include).columns)
+    df = convert(dataset, pd.DataFrame)
+    columns = list(df.select_dtypes(include=include).columns)   # pyright: ignore[reportCallIssue, reportArgumentType]
     return columns
 
 
@@ -173,7 +175,7 @@ def identify_columns_with_na(dataset: TwoDimArray) -> pd.DataFrame:
     1   col2         2
     0   col1         1
     """
-    df: pd.DataFrame = convert(dataset, "dataframe")  # type: ignore
+    df = convert(dataset, pd.DataFrame)
     na_count = [sum(df[cc].isnull()) for cc in df.columns]
     return (
         pd.DataFrame({"column": df.columns, "na_count": na_count})

dython/examples.py

@@ -22,8 +22,8 @@ def roc_graph_example():
 
     # Load data
     iris = datasets.load_iris()
-    X = iris.data
-    y = label_binarize(iris.target, classes=[0, 1, 2])
+    X = iris.data                                                                   # pyright: ignore[reportAttributeAccessIssue]
+    y = label_binarize(iris.target, classes=[0, 1, 2])                              # pyright: ignore[reportAttributeAccessIssue]   
 
     # Add noisy features
     random_state = np.random.RandomState(4)
@@ -43,7 +43,7 @@ def roc_graph_example():
 
     # Plot ROC graphs
     return metric_graph(
-        y_test, y_score, "roc", class_names_list=iris.target_names
+        y_test, y_score, "roc", class_names_list=iris.target_names                  # pyright: ignore[reportAttributeAccessIssue, reportCallIssue]
     )
 
 
@@ -55,8 +55,8 @@ def pr_graph_example():
 
     # Load data
     iris = datasets.load_iris()
-    X = iris.data
-    y = label_binarize(iris.target, classes=[0, 1, 2])
+    X = iris.data                                                                   # pyright: ignore[reportAttributeAccessIssue]   
+    y = label_binarize(iris.target, classes=[0, 1, 2])                              # pyright: ignore[reportAttributeAccessIssue]
 
     # Add noisy features
     random_state = np.random.RandomState(4)
@@ -76,7 +76,7 @@ def pr_graph_example():
 
     # Plot PR graphs
     return metric_graph(
-        y_test, y_score, "pr", class_names_list=iris.target_names
+        y_test, y_score, "pr", class_names_list=iris.target_names                  # pyright: ignore[reportAttributeAccessIssue, reportCallIssue]
     )
 
 
@@ -91,10 +91,10 @@ def associations_iris_example():
 
     # Convert int classes to strings to allow associations method
     # to automatically recognize categorical columns
-    target = ["C{}".format(i) for i in iris.target]
+    target = ["C{}".format(i) for i in iris.target]                                 # pyright: ignore[reportAttributeAccessIssue]   
 
     # Prepare data
-    X = pd.DataFrame(data=iris.data, columns=iris.feature_names)
+    X = pd.DataFrame(data=iris.data, columns=iris.feature_names)                    # pyright: ignore[reportAttributeAccessIssue]
     y = pd.DataFrame(data=target, columns=["target"])
     df = pd.concat([X, y], axis=1)
 
@@ -151,8 +151,8 @@ def split_hist_example():
 
     # Load data and convert to DataFrame
     data = datasets.load_breast_cancer()
-    df = pd.DataFrame(data=data.data, columns=data.feature_names)
-    df["malignant"] = [not bool(x) for x in data.target]
+    df = pd.DataFrame(data=data.data, columns=data.feature_names)                       # pyright: ignore[reportAttributeAccessIssue]
+    df["malignant"] = [not bool(x) for x in data.target]                                # pyright: ignore[reportAttributeAccessIssue]
 
     # Plot histogram
     return split_hist(df, "mean radius", "malignant", bins=20, figsize=(15, 7))
@@ -167,7 +167,7 @@ def ks_abc_example():
     # Load and split data
     data = datasets.load_breast_cancer()
     X_train, X_test, y_train, y_test = train_test_split(
-        data.data, data.target, test_size=0.5, random_state=0
+        data.data, data.target, test_size=0.5, random_state=0                           # pyright: ignore[reportAttributeAccessIssue]
     )
 
     # Train model and predict