onekit 1.3.0__tar.gz → 1.4.0__tar.gz

{onekit-1.3.0 → onekit-1.4.0}/PKG-INFO

@@ -1,13 +1,13 @@
 Metadata-Version: 2.1
 Name: onekit
-Version: 1.3.0
+Version: 1.4.0
 Summary: All-in-One Python Kit.
 Home-page: https://github.com/estripling/onekit
 License: BSD 3-Clause
 Keywords: onekit
 Author: Eugen Stripling
 Author-email: estripling042@gmail.com
-Requires-Python: >=3.8.1
+Requires-Python: >=3.9
 Classifier: License :: Other/Proprietary License
 Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.9
@@ -15,7 +15,6 @@ Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3 :: Only
-Classifier: Programming Language :: Python :: 3.8
 Requires-Dist: pytz (>=2024.1,<2025.0)
 Requires-Dist: toolz (>=0.12.0,<0.13.0)
 Project-URL: Documentation, https://onekit.readthedocs.io/en/stable/
@@ -46,7 +45,7 @@ All-in-One Python Kit:
 
 ## Installation
 
-`onekit` is available on [PyPI](https://pypi.org/project/onekit/) for Python 3.8+:
+`onekit` is available on [PyPI](https://pypi.org/project/onekit/) for Python 3.9+:
 
 ```console
 pip install onekit

{onekit-1.3.0 → onekit-1.4.0}/README.md

@@ -22,7 +22,7 @@ All-in-One Python Kit:
 
 ## Installation
 
-`onekit` is available on [PyPI](https://pypi.org/project/onekit/) for Python 3.8+:
+`onekit` is available on [PyPI](https://pypi.org/project/onekit/) for Python 3.9+:
 
 ```console
 pip install onekit

{onekit-1.3.0 → onekit-1.4.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "onekit"
-version = "1.3.0"
+version = "1.4.0"
 description = "All-in-One Python Kit."
 authors = ["Eugen Stripling <estripling042@gmail.com>"]
 license = "BSD 3-Clause"
@@ -10,7 +10,6 @@ documentation = "https://onekit.readthedocs.io/en/stable/"
 keywords = ["onekit"]
 classifiers = [
     "Programming Language :: Python :: 3 :: Only",
-    "Programming Language :: Python :: 3.8",
     "Programming Language :: Python :: 3.9",
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
@@ -18,7 +17,7 @@ classifiers = [
 ]
 
 [tool.poetry.dependencies]
-python = ">=3.8.1"
+python = ">=3.9"
 toolz = "^0.12.0"
 pytz = "^2024.1"
 
@@ -50,6 +49,9 @@ python-semantic-release = "^8.3.0"
 [tool.poetry.group.pandaskit.dependencies]
 pandas = ">=0.23.2"
 
+[tool.poetry.group.sklearnkit.dependencies]
+scikit-learn = ">=1.3"
+
 [tool.poetry.group.sparkkit.dependencies]
 pyspark = "3.1.1"
 

onekit-1.4.0/src/onekit/sklearnkit.py

@@ -0,0 +1,153 @@
+from typing import (
+    Optional,
+    Union,
+)
+
+import numpy as np
+import numpy.typing as npt
+import pandas as pd
+from pandas import DataFrame as PandasDF
+from sklearn import metrics
+from sklearn.utils import validation
+
+__all__ = (
+    "precision_given_recall_score",
+    "threshold_summary",
+)
+
+
+ArrayLike = npt.ArrayLike
+
+
+def precision_given_recall_score(
+    y_true: ArrayLike,
+    y_score: ArrayLike,
+    *,
+    min_recall: float,
+    pos_label: Optional[Union[int, str]] = None,
+) -> float:
+    """Compute precision given a desired minimum recall level.
+
+    Examples
+    --------
+    >>> import onekit.sklearnkit as slk
+    >>> y_true = [0, 1, 1, 1, 0, 0, 0, 1]
+    >>> y_score = [0.1, 0.4, 0.35, 0.8, 0.5, 0.2, 0.75, 0.5]
+    >>> slk.precision_given_recall_score(y_true, y_score, min_recall=0.7)
+    0.6
+    """
+    if not (0 < min_recall <= 1):
+        raise ValueError(f"{min_recall=} - must be a float in (0, 1]")
+
+    df = (
+        threshold_summary(y_true, y_score, pos_label=pos_label)
+        .filter(items=["precision", "recall"])
+        .query(f"recall >= {min_recall}")
+    )
+
+    min_empirical_recall = df["recall"].min()
+
+    return float(
+        0
+        if df.empty
+        else df.query(f"recall == {min_empirical_recall}")["precision"].max()
+    )
+
+
+def threshold_summary(
+    y_true: ArrayLike,
+    y_score: ArrayLike,
+    *,
+    pos_label: Optional[Union[int, str]] = None,
+) -> PandasDF:
+    """Threshold summary.
+
+    Notes
+    -----
+    - Support for binary classification only
+    - Assumpution: classifier returns scores
+    - First values correspond to the edge case where everything is predicted positive
+    - Last values correspond to the edge case where everything is predicted negative
+
+    Examples
+    --------
+    >>> import onekit.sklearnkit as slk
+    >>> y_true = [0, 1, 1, 1, 0, 0, 0, 1]
+    >>> y_score = [0.1, 0.4, 0.35, 0.8, 0.5, 0.2, 0.75, 0.5]
+    >>> with pd.option_context("display.float_format", "{:.2f}".format):
+    ...     slk.threshold_summary(y_true, y_score).T
+                          0    1    2    3    4    5    6    7
+    threshold          0.10 0.20 0.35 0.40 0.50 0.75 0.80  inf
+    predicted_positive 8.00 7.00 6.00 5.00 4.00 2.00 1.00 0.00
+    true_positive      4.00 4.00 4.00 3.00 2.00 1.00 1.00 0.00
+    false_positive     4.00 3.00 2.00 2.00 2.00 1.00 0.00 0.00
+    false_negative     0.00 0.00 0.00 1.00 2.00 3.00 3.00 4.00
+    true_negative      0.00 1.00 2.00 2.00 2.00 3.00 4.00 4.00
+    precision          0.50 0.57 0.67 0.60 0.50 0.50 1.00 1.00
+    recall             1.00 1.00 1.00 0.75 0.50 0.25 0.25 0.00
+    f1                 0.67 0.73 0.80 0.67 0.50 0.33 0.40 0.00
+    accuracy           0.50 0.62 0.75 0.62 0.50 0.50 0.62 0.50
+    balanced_accuracy  0.50 0.62 0.75 0.62 0.50 0.50 0.62 0.50
+    matthews_corrcoef   NaN 0.38 0.58 0.26 0.00 0.00 0.38  NaN
+    """
+    y = validation.column_or_1d(y_true)
+    s = validation.column_or_1d(y_score)
+    validation.check_consistent_length(y, s)
+    validation.assert_all_finite(y)
+    validation.assert_all_finite(s)
+    pos_label = validation._check_pos_label_consistency(pos_label, y)
+
+    precision, recall, thresholds = metrics.precision_recall_curve(
+        y_true=y,
+        y_score=s,
+        pos_label=pos_label,
+        sample_weight=None,
+        drop_intermediate=False,
+    )
+
+    is_true_pos = y == pos_label
+    is_true_neg = y != pos_label
+
+    def is_pred_pos(t: float) -> np.ndarray:
+        return s >= t
+
+    def is_pred_neg(t: float) -> np.ndarray:
+        return s < t
+
+    return (
+        pd.DataFrame(np.append(thresholds, np.inf), columns=["t"])
+        .assign(
+            pp=lambda df: df.t.map(lambda t: is_pred_pos(t).sum()),
+            tp=lambda df: df.t.map(lambda t: (is_pred_pos(t) & is_true_pos).sum()),
+            fp=lambda df: df.t.map(lambda t: (is_pred_pos(t) & is_true_neg).sum()),
+            fn=lambda df: df.t.map(lambda t: (is_pred_neg(t) & is_true_pos).sum()),
+            tn=lambda df: df.t.map(lambda t: (is_pred_neg(t) & is_true_neg).sum()),
+            precision=precision,
+            recall=recall,
+            f1=2 * (precision * recall) / (precision + recall),
+            acc=lambda df: (df.tp + df.tn) / (df.tp + df.tn + df.fp + df.fn),
+            bacc=lambda df: 0.5 * (df.tp / (df.tp + df.fn) + df.tn / (df.tn + df.fp)),
+            mcc=lambda df: np.true_divide(
+                (df.tp * df.tn - df.fp * df.fn),
+                np.sqrt(
+                    (df.tp + df.fp)
+                    * (df.tp + df.fn)
+                    * (df.tn + df.fp)
+                    * (df.tn + df.fn)
+                ),
+            ),
+        )
+        .rename(
+            columns={
+                "t": "threshold",
+                "pp": "predicted_positive",
+                "tp": "true_positive",
+                "fp": "false_positive",
+                "fn": "false_negative",
+                "tn": "true_negative",
+                "acc": "accuracy",
+                "bacc": "balanced_accuracy",
+                "mcc": "matthews_corrcoef",
+            },
+        )
+    )

{onekit-1.3.0 → onekit-1.4.0}/src/onekit/sparkkit.py

@@ -1075,9 +1075,11 @@ def select_col_types(df: SparkDF, /, *col_types: T.DataType) -> List[str]:
     >>> sk.select_col_types(df, T.IntegerType, T.LongType)
     ['int', 'long']
     """
+    valid_types = {v.typeName() for k, v in T.__dict__.items() if k.endswith("Type")}
     col_types = tuple(pk.flatten(col_types))
-    if not all(isinstance(col_type, T.DataTypeSingleton) for col_type in col_types):
-        raise TypeError(f"{col_types=} - must be a data type of pyspark.sql.types")
+    for col_type in col_types:
+        if not hasattr(col_type, "typeName") or col_type.typeName() not in valid_types:
+            raise TypeError(f"{col_type=} - must be a valid data type: {valid_types}")
     return [c for c in df.columns if isinstance(df.schema[c].dataType, col_types)]