dscience-tools 2.3.0__tar.gz → 2.3.2__tar.gz

{dscience_tools-2.3.0/dscience_tools.egg-info → dscience_tools-2.3.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dscience_tools
-Version: 2.3.0
+Version: 2.3.2
 Summary: DSTools: Data Science Tools Library
 Home-page: https://github.com/s-kav/ds_tools
 Author: Sergii Kavun
@@ -421,12 +421,14 @@ A high-performance toolkit for calculating loss functions and their gradients.
 - **Classification Losses:** `hinge_loss`, `log_loss` (Binary Cross-Entropy).
 - **Embedding Losses:** `triplet_loss`.
 - **Monitoring:** `start_monitoring`, `update`, `get_history_df`, `plot_history`.
+- **`list_metrics`**: Prints a list of all available metrics.
 
 ### Distance Toolkit (`tools.distance.*`)
 A high-performance toolkit for calculating distances and similarities.
 - **Vector-to-Vector:** `euclidean`, `manhattan`, `cosine_similarity`, `minkowski`, `chebyshev`, `mahalanobis`, `haversine`, `hamming`, `jaccard`.
 - **Matrix Operations:** `pairwise_euclidean`, `kmeans_distance`.
 - **Neighbor Searches:** `knn_distances`, `radius_neighbors`.
+- **`list_distances`**: Prints a list of all available distances.
 
 
 # Authors
@@ -440,19 +442,28 @@ See [CONTRIBUTING](/CONTRIBUTING.md)
 
 # TODO
 1. Add some important kind of plots:
+
 📌 KS Plot – Measures how well your model separates positive and negative classes.
+
 📌 SHAP Plot – Explains feature impact and model interpretability.
+
 📌 QQ Plot – Checks if your data follows a theoretical distribution.
+
 📌 Cumulative Explained Variance – Helps decide the optimal number of PCA components.
+
 📌 Gini vs Entropy – Key metrics for understanding decision tree impurity.
+
 📌 Bias–Variance Tradeoff – Shows the balance between underfitting and overfitting.
+
 📌 ROC Curve – Evaluates classification performance across thresholds.
+
 📌 Precision–Recall Curve – Crucial for imbalanced datasets.
+
 📌 Elbow Curve – Helps choose the right number of clusters in K-Means.
 
-2. Implement Fast Fourier Transform (FFT) algorithm and Shannon’s interpolation formula
+2. Implement Fast Fourier Transform (FFT) algorithm and Shannon’s interpolation formula: done, 2.3.0.
 
-3. Add some fast distance metrics (expand of existed).
+3. Add some fast distance metrics (expand of existed): done, 2.3.0.
 
 
 # References

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/README.md

@@ -377,12 +377,14 @@ A high-performance toolkit for calculating loss functions and their gradients.
 - **Classification Losses:** `hinge_loss`, `log_loss` (Binary Cross-Entropy).
 - **Embedding Losses:** `triplet_loss`.
 - **Monitoring:** `start_monitoring`, `update`, `get_history_df`, `plot_history`.
+- **`list_metrics`**: Prints a list of all available metrics.
 
 ### Distance Toolkit (`tools.distance.*`)
 A high-performance toolkit for calculating distances and similarities.
 - **Vector-to-Vector:** `euclidean`, `manhattan`, `cosine_similarity`, `minkowski`, `chebyshev`, `mahalanobis`, `haversine`, `hamming`, `jaccard`.
 - **Matrix Operations:** `pairwise_euclidean`, `kmeans_distance`.
 - **Neighbor Searches:** `knn_distances`, `radius_neighbors`.
+- **`list_distances`**: Prints a list of all available distances.
 
 
 # Authors
@@ -396,19 +398,28 @@ See [CONTRIBUTING](/CONTRIBUTING.md)
 
 # TODO
 1. Add some important kind of plots:
+
 📌 KS Plot – Measures how well your model separates positive and negative classes.
+
 📌 SHAP Plot – Explains feature impact and model interpretability.
+
 📌 QQ Plot – Checks if your data follows a theoretical distribution.
+
 📌 Cumulative Explained Variance – Helps decide the optimal number of PCA components.
+
 📌 Gini vs Entropy – Key metrics for understanding decision tree impurity.
+
 📌 Bias–Variance Tradeoff – Shows the balance between underfitting and overfitting.
+
 📌 ROC Curve – Evaluates classification performance across thresholds.
+
 📌 Precision–Recall Curve – Crucial for imbalanced datasets.
+
 📌 Elbow Curve – Helps choose the right number of clusters in K-Means.
 
-2. Implement Fast Fourier Transform (FFT) algorithm and Shannon’s interpolation formula
+2. Implement Fast Fourier Transform (FFT) algorithm and Shannon’s interpolation formula: done, 2.3.0.
 
-3. Add some fast distance metrics (expand of existed).
+3. Add some fast distance metrics (expand of existed): done, 2.3.0.
 
 
 # References

{dscience_tools-2.3.0 → dscience_tools-2.3.2/dscience_tools.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dscience_tools
-Version: 2.3.0
+Version: 2.3.2
 Summary: DSTools: Data Science Tools Library
 Home-page: https://github.com/s-kav/ds_tools
 Author: Sergii Kavun
@@ -421,12 +421,14 @@ A high-performance toolkit for calculating loss functions and their gradients.
 - **Classification Losses:** `hinge_loss`, `log_loss` (Binary Cross-Entropy).
 - **Embedding Losses:** `triplet_loss`.
 - **Monitoring:** `start_monitoring`, `update`, `get_history_df`, `plot_history`.
+- **`list_metrics`**: Prints a list of all available metrics.
 
 ### Distance Toolkit (`tools.distance.*`)
 A high-performance toolkit for calculating distances and similarities.
 - **Vector-to-Vector:** `euclidean`, `manhattan`, `cosine_similarity`, `minkowski`, `chebyshev`, `mahalanobis`, `haversine`, `hamming`, `jaccard`.
 - **Matrix Operations:** `pairwise_euclidean`, `kmeans_distance`.
 - **Neighbor Searches:** `knn_distances`, `radius_neighbors`.
+- **`list_distances`**: Prints a list of all available distances.
 
 
 # Authors
@@ -440,19 +442,28 @@ See [CONTRIBUTING](/CONTRIBUTING.md)
 
 # TODO
 1. Add some important kind of plots:
+
 📌 KS Plot – Measures how well your model separates positive and negative classes.
+
 📌 SHAP Plot – Explains feature impact and model interpretability.
+
 📌 QQ Plot – Checks if your data follows a theoretical distribution.
+
 📌 Cumulative Explained Variance – Helps decide the optimal number of PCA components.
+
 📌 Gini vs Entropy – Key metrics for understanding decision tree impurity.
+
 📌 Bias–Variance Tradeoff – Shows the balance between underfitting and overfitting.
+
 📌 ROC Curve – Evaluates classification performance across thresholds.
+
 📌 Precision–Recall Curve – Crucial for imbalanced datasets.
+
 📌 Elbow Curve – Helps choose the right number of clusters in K-Means.
 
-2. Implement Fast Fourier Transform (FFT) algorithm and Shannon’s interpolation formula
+2. Implement Fast Fourier Transform (FFT) algorithm and Shannon’s interpolation formula: done, 2.3.0.
 
-3. Add some fast distance metrics (expand of existed).
+3. Add some fast distance metrics (expand of existed): done, 2.3.0.
 
 
 # References

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "dscience_tools"
-version = "2.3.0"
+version = "2.3.2"
 description = "DSTools: Data Science Tools Library"
 readme = "README.md"
 
@@ -86,5 +86,4 @@ filterwarnings = [
 testpaths = ["tests"]
 pythonpath = [
   ".",
-  "src",
 ]

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
 name = dscience_tools
-version = 2.3.0
+version = 2.3.2
 author = Sergii Kavun
 author_email = kavserg@gmail.com
 description = A library of helpful functions for various data science research stages.

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/src/__init__.py

@@ -13,10 +13,10 @@
 DSTools: A library of helpful functions for various data science research stages.
 """
 
-__version__ = "2.3.0"
+__version__ = "2.3.2"
 
-from ds_tool import DSTools
-from models import (
+from .ds_tool import DSTools
+from .models import (
     CorrelationConfig,
     DistributionConfig,
     GrubbsTestResult,

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/src/distance.py

@@ -630,20 +630,20 @@ class Distance:
             f"Distance initialized. GPU: {self.gpu_available}, Numba: {self.numba_available}"
         )
 
-    def list_metrics(self) -> pd.DataFrame:
+    def list_distances(self) -> pd.DataFrame:
         """
-        Returns a DataFrame listing all available metric functions, their
+        Returns a DataFrame listing all available distances functions, their
         descriptions, and usage signatures.
 
         Returns:
-            pd.DataFrame: A table with columns 'Metric', 'Description', and 'Usage'.
+            pd.DataFrame: A table with columns 'Distance', 'Description', and 'Usage'.
         """
         methods_data = []
 
         # Iterate over all members of the instance
         for name, func in inspect.getmembers(self, predicate=inspect.ismethod):
-            # Skip private methods (starting with _) and the list_metrics method itself
-            if name.startswith("_") or name == "list_metrics":
+            # Skip private methods (starting with _) and the list_distances method itself
+            if name.startswith("_") or name == "list_distances":
                 continue
 
             # Get the first line of the docstring
@@ -658,16 +658,16 @@ class Distance:
 
             methods_data.append(
                 {
-                    "Metric": name,
+                    "Distance": name,
                     "Description": description,
                     "Usage": f"{name}{signature}",
                 }
             )
 
-        # Create DataFrame and sort by Metric name
+        # Create DataFrame and sort by Distance name
         df = pd.DataFrame(methods_data)
         if not df.empty:
-            df = df.sort_values(by="Metric").reset_index(drop=True)
+            df = df.sort_values(by="Distance").reset_index(drop=True)
 
         return df
 

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/src/ds_tool.py

@@ -49,9 +49,9 @@ from sklearn.metrics import (
 from sklearn.preprocessing import OrdinalEncoder
 from statsmodels.tsa.stattools import adfuller
 
-from distance import Distance
-from metrics import Metrics
-from models import (
+from .distance import Distance
+from .metrics import Metrics
+from .models import (
     CorrelationConfig,
     DistributionConfig,
     GrubbsTestResult,

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/tests/test_compute_metrics.py

@@ -12,8 +12,7 @@
 
 import numpy as np
 import pytest
-
-from ds_tool import MetricsConfig
+from ds_tools.models import MetricsConfig
 
 N_SAMPLES = 200
 THRESHOLD = 0.5

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/tests/test_corr_matrix.py

@@ -13,8 +13,7 @@
 import numpy as np
 import pandas as pd
 import pytest
-
-from ds_tool import CorrelationConfig
+from ds_tools.models import CorrelationConfig
 
 N_SAMPLES = 100
 

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/tests/test_distance.py

@@ -22,13 +22,12 @@ Error and edge case handling (incorrect sizes, empty arrays, incorrect parameter
 import numpy as np
 import pandas as pd
 import pytest
+from ds_tools.distance import CUPY_AVAILABLE, NUMBA_AVAILABLE
 from scipy.spatial.distance import cdist
 from scipy.stats import entropy
 
-from distance import CUPY_AVAILABLE, NUMBA_AVAILABLE
-
 if NUMBA_AVAILABLE:
-    from distance import (
+    from ds_tools.distance import (
         _canberra_numba,
         _chebyshev_numba,
         _cosine_similarity_numba,
@@ -236,14 +235,16 @@ def test_public_interface_calls_correct_backend(tools, mocker, small_sample_vect
     if NUMBA_AVAILABLE:
         mocker.patch.object(tools.distance, "gpu_available", False)
         mocker.patch.object(tools.distance, "numba_available", True)
-        mock_numba = mocker.patch("distance._manhattan_numba", return_value=1.0)
+        mock_numba = mocker.patch(
+            "ds_tools.distance._manhattan_numba", return_value=1.0
+        )
         tools.distance.manhattan(u, v)
         mock_numba.assert_called_once()
 
     # Case 2: Only NumPy is available
     mocker.patch.object(tools.distance, "gpu_available", False)
     mocker.patch.object(tools.distance, "numba_available", False)
-    mock_numpy = mocker.patch("distance._manhattan_numpy", return_value=2.0)
+    mock_numpy = mocker.patch("ds_tools.distance._manhattan_numpy", return_value=2.0)
     tools.distance.manhattan(u, v)
     mock_numpy.assert_called_once()
 
@@ -264,9 +265,11 @@ def test_backend_dispatching_logic(
 
     dist = tools.distance
     mocker.patch.object(dist, "gpu_available", True)
-    mocker.patch("distance.cp.asarray", side_effect=lambda x: x)
-    mock_numba = mocker.patch("distance._euclidean_numba", return_value=1.0)
-    mock_cupy = mocker.patch("distance._euclidean_cupy", return_value=2.0, create=True)
+    mocker.patch("ds_tools.distance.cp.asarray", side_effect=lambda x: x)
+    mock_numba = mocker.patch("ds_tools.distance._euclidean_numba", return_value=1.0)
+    mock_cupy = mocker.patch(
+        "ds_tools.distance._euclidean_cupy", return_value=2.0, create=True
+    )
 
     # 1. Small data -> Numba should be used
     u_small, v_small = small_sample_vectors
@@ -393,7 +396,9 @@ def test_fallback_without_numba(tools, mocker, small_sample_vectors):
     mocker.patch.object(dist, "numba_available", False)
 
     # --- Spy on the NumPy backend to ensure it's called ---
-    mock_numpy_fallback = mocker.patch("distance._euclidean_numpy", return_value=99.0)
+    mock_numpy_fallback = mocker.patch(
+        "ds_tools.distance._euclidean_numpy", return_value=99.0
+    )
 
     # Call the method on the original 'tools' instance
     result = dist.euclidean(u, v)
@@ -922,19 +927,19 @@ def test_cupy_backend_correctness(
     assert np.isclose(result, expected, rtol=1e-5)
 
 
-def test_list_metrics_returns_correct_dataframe(tools):
-    """Tests that list_metrics returns a non-empty DataFrame with correct columns."""
-    df = tools.distance.list_metrics()
+def test_list_distances_returns_correct_dataframe(tools):
+    """Tests that list_distances returns a non-empty DataFrame with correct columns."""
+    df = tools.distance.list_distances()
 
     assert isinstance(df, pd.DataFrame)
     assert not df.empty
 
     # Check expected columns
-    expected_cols = ["Metric", "Description", "Usage"]
+    expected_cols = ["Distance", "Description", "Usage"]
     assert list(df.columns) == expected_cols
 
     # Check that known metrics are present
-    metrics = df["Metric"].tolist()
+    metrics = df["Distance"].tolist()
     assert "euclidean" in metrics
     assert "manhattan" in metrics
     assert "cosine_similarity" in metrics
@@ -944,6 +949,6 @@ def test_list_metrics_returns_correct_dataframe(tools):
     assert "_dispatch_v2v" not in metrics
 
     # Check content of a specific row (e.g., euclidean)
-    euclidean_row = df[df["Metric"] == "euclidean"].iloc[0]
+    euclidean_row = df[df["Distance"] == "euclidean"].iloc[0]
     assert "Euclidean" in euclidean_row["Description"]
     assert "force_cpu" in euclidean_row["Usage"]

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/tests/test_distance_additional.py

@@ -14,9 +14,8 @@ import types
 
 import numpy as np
 import pytest
-
-import distance
-from distance import CUPY_AVAILABLE, NUMBA_AVAILABLE
+from ds_tools import distance
+from ds_tools.distance import CUPY_AVAILABLE, NUMBA_AVAILABLE
 
 # --- Helpers ---------------------------------------------------------------
 
@@ -76,7 +75,8 @@ def test_pairwise_euclidean_fallback_when_cupy_missing(monkeypatch, mocker):
     monkeypatch.delitem(distance.__dict__, "_pairwise_euclidean_cupy", raising=False)
 
     mock_numpy = mocker.patch(
-        "distance._pairwise_euclidean_numpy", return_value=np.full((10, 12), 777.0)
+        "ds_tools.distance._pairwise_euclidean_numpy",
+        return_value=np.full((10, 12), 777.0),
     )
 
     res = dist.pairwise_euclidean(X, Y, force_cpu=False)

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/tests/test_distance_numba_cupy.py

@@ -1,8 +1,7 @@
 # tests/test_distance_numba_cupy.py
 import numpy as np
 import pytest
-
-import distance
+from ds_tools import distance
 
 NUMBA_AVAILABLE = distance.NUMBA_AVAILABLE
 CUPY_AVAILABLE = distance.CUPY_AVAILABLE

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/tests/test_generate_dist.py

@@ -12,10 +12,9 @@
 
 import numpy as np
 import pytest
+from ds_tools.models import DistributionConfig
 from scipy import stats
 
-from ds_tool import DistributionConfig
-
 
 def describe_generated_data(data: np.ndarray):
     return {
@@ -88,7 +87,7 @@ def test_generate_distribution_zero_std_case(tools, mocker):
         mean=50, median=50, std=1, min_val=0, max_val=100, skewness=0, kurtosis=3, n=10
     )
     # artificially create a situation with zero std by mocking np.std
-    mocker.patch("ds_tool.np.std", return_value=0)
+    mocker.patch("ds_tools.ds_tool.np.std", return_value=0)
     data = tools.generate_distribution(config)
 
     unique_values = np.unique(data)

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/tests/test_generate_from_metrics.py

@@ -14,10 +14,9 @@ import numpy as np
 import pandas as pd
 import polars as pl
 import pytest
+from ds_tools.models import DistributionConfig
 from scipy import stats
 
-from ds_tool import DistributionConfig
-
 
 def describe_metrics(arr: np.ndarray):
     return {
@@ -146,7 +145,7 @@ def test_generate_distribution_zero_std_case(tools, mocker):
         mean=50, median=50, std=1, min_val=0, max_val=100, skewness=0, kurtosis=3, n=10
     )
     # artificially create a situation with zero std by mocking np.std
-    mocker.patch("ds_tool.np.std", return_value=0)
+    mocker.patch("ds_tools.ds_tool.np.std", return_value=0)
     data = tools.generate_distribution(config)
     unique_values = np.unique(data)
     expected_values = {config.min_val, config.max_val, config.mean}

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/tests/test_metrics.py

@@ -25,9 +25,7 @@ Error and edge case handling (empty arrays, size mismatch, invalid parameters).
 import numpy as np
 import pandas as pd
 import pytest
-
-from ds_tool import Metrics
-from metrics import CUPY_AVAILABLE, NUMBA_AVAILABLE
+from ds_tools.metrics import CUPY_AVAILABLE, NUMBA_AVAILABLE, Metrics
 
 pytestmark_cupy = pytest.mark.skipif(
     not CUPY_AVAILABLE, reason="CuPy or compatible GPU is not available"
@@ -247,10 +245,12 @@ def test_gpu_threshold_logic(tools, mocker, small_sample_data, large_sample_data
         pytest.skip("This test requires both CuPy and Numba to be installed.")
 
     mocker.patch.object(tools.metrics, "gpu_available", True)
-    mocker.patch("metrics.cp.asarray", side_effect=lambda x: x, create=True)
+    mocker.patch("ds_tools.metrics.cp.asarray", side_effect=lambda x: x, create=True)
     # Mock the backends to see which one is called
-    mock_cupy = mocker.patch("metrics._mae_cupy", return_value=1.0, create=True)
-    mock_numba = mocker.patch("metrics._mae_numba", return_value=2.0)
+    mock_cupy = mocker.patch(
+        "ds_tools.metrics._mae_cupy", return_value=1.0, create=True
+    )
+    mock_numba = mocker.patch("ds_tools.metrics._mae_numba", return_value=2.0)
 
     # 1. Test with small data -> Numba should be called
     y_true_small, y_pred_small = small_sample_data
@@ -271,7 +271,7 @@ def test_gpu_threshold_logic(tools, mocker, small_sample_data, large_sample_data
 )
 def test_dispatch_fallback_to_numpy(tools, mocker, small_sample_data):
     """Tests that dispatcher falls back to NumPy if Numba is missing."""
-    mock_numpy = mocker.patch("metrics._mae_numpy", return_value=99.0)
+    mock_numpy = mocker.patch("ds_tools.metrics._mae_numpy", return_value=99.0)
     y_true, y_pred = small_sample_data
     result = tools.metrics.mae(y_true, y_pred, force_cpu=True)
     mock_numpy.assert_called_once()

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/tests/test_models.py

@@ -13,8 +13,7 @@ This file validates default values, constraints, and custom validation logic.
  *
 """
 import pytest
-
-from models import (
+from ds_tools.models import (
     CorrelationConfig,
     DistributionConfig,
     GrubbsTestResult,

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/tests/test_outliers.py

@@ -13,8 +13,7 @@
 import numpy as np
 import pandas as pd
 import pytest
-
-from ds_tool import OutlierConfig
+from ds_tools.models import OutlierConfig
 
 
 @pytest.fixture(scope="module")

{dscience_tools-2.3.0 → dscience_tools-2.3.2}/tests/test_zip_io.py

@@ -135,9 +135,12 @@ def test_read_dataframes_from_zip_pandas_fallback_to_csv(
     # mock pd.read_parquet so it throws an error,
     # and check that after this pd.read_csv will be called.
     mocker.patch(
-        "ds_tool.pd.read_parquet", side_effect=ValueError("Simulated read error")
+        "ds_tools.ds_tool.pd.read_parquet",
+        side_effect=ValueError("Simulated read error"),
+    )
+    mock_read_csv = mocker.patch(
+        "ds_tools.ds_tool.pd.read_csv", return_value=sample_pandas_df
     )
-    mock_read_csv = mocker.patch("ds_tool.pd.read_csv", return_value=sample_pandas_df)
 
     with tempfile.TemporaryDirectory() as temp_dir:
         # Create a dummy zip archive with a "non-parquet" file, but with the .parquet extension