sbss 0.0.1__tar.gz

sbss-0.0.1/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2023 tim
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

sbss-0.0.1/PKG-INFO

@@ -0,0 +1,62 @@
+Metadata-Version: 2.1
+Name: sbss
+Version: 0.0.1
+Summary: Similarity-Based Stratified Splitting Algorithm
+Home-page: https://github.com/timothyckl/similarity-stratified-split
+Author: timothyckl
+Author-email: timothy.ckl@outlook.com
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Operating System :: OS Independent
+Description-Content-Type: text/markdown
+License-File: LICENSE
+
+# Similarity Stratified Split
+
+Implementation of the Similarity-Based Stratified Splitting algorithm described in [Similarity Based Stratified Splitting: an approach to train better classifiers](https://arxiv.org/abs/2010.06099).
+
+## Overview
+
+The authors propose a Similarity-Based Stratified Splitting (SBSS) technique, which uses both the output and input space information to split a dataset. Splits are generated using similarity functions among samples to place similar samples in different splits. This approach allows for a better representation of the data in the training phase. This strategy leads to a more realistic performance estimation when used in real-world applications.
+
+## Install
+
+<!-- **PyPI**
+```bash
+pip install sbss
+``` -->
+
+**Local**
+
+```
+git clone https://github.com/timothyckl/similarity-stratified-split.git
+cd ./similarity-stratified-split
+pip install -e .
+```
+
+## Usage
+
+```python
+import numpy as np
+from scipy.spatial import distance
+from sbss import SimilarityStratifiedSplit
+
+def get_distances(x):
+    distances = distance.squareform(distance.pdist(x, metric='euclidean'))
+    return distances
+
+# inputs are recommended to be normalized
+X = np.random.rand(1000, 128)
+y = np.random.randint(0, 10, (1000,))
+
+n_splits = 3
+s = SimilarityStratifiedSplit(n_splits, get_distances)
+
+for train_index, test_index in s.split(X, y):
+  print(f"Train indices: {train_index}\nTest indices: {test_index}")
+  print("="*100)
+```
+
+## References
+
+- Farias, F., Ludermir, T. and Bastos-Filho, C. (2020) Similarity based stratified splitting: An approach to train better classifiers, arXiv.org. Available at: https://arxiv.org/abs/2010.06099 (Accessed: 27 November 2023). 

sbss-0.0.1/README.md

@@ -0,0 +1,49 @@
+# Similarity Stratified Split
+
+Implementation of the Similarity-Based Stratified Splitting algorithm described in [Similarity Based Stratified Splitting: an approach to train better classifiers](https://arxiv.org/abs/2010.06099).
+
+## Overview
+
+The authors propose a Similarity-Based Stratified Splitting (SBSS) technique, which uses both the output and input space information to split a dataset. Splits are generated using similarity functions among samples to place similar samples in different splits. This approach allows for a better representation of the data in the training phase. This strategy leads to a more realistic performance estimation when used in real-world applications.
+
+## Install
+
+<!-- **PyPI**
+```bash
+pip install sbss
+``` -->
+
+**Local**
+
+```
+git clone https://github.com/timothyckl/similarity-stratified-split.git
+cd ./similarity-stratified-split
+pip install -e .
+```
+
+## Usage
+
+```python
+import numpy as np
+from scipy.spatial import distance
+from sbss import SimilarityStratifiedSplit
+
+def get_distances(x):
+    distances = distance.squareform(distance.pdist(x, metric='euclidean'))
+    return distances
+
+# inputs are recommended to be normalized
+X = np.random.rand(1000, 128)
+y = np.random.randint(0, 10, (1000,))
+
+n_splits = 3
+s = SimilarityStratifiedSplit(n_splits, get_distances)
+
+for train_index, test_index in s.split(X, y):
+  print(f"Train indices: {train_index}\nTest indices: {test_index}")
+  print("="*100)
+```
+
+## References
+
+- Farias, F., Ludermir, T. and Bastos-Filho, C. (2020) Similarity based stratified splitting: An approach to train better classifiers, arXiv.org. Available at: https://arxiv.org/abs/2010.06099 (Accessed: 27 November 2023). 

sbss-0.0.1/sbss/sbss.egg-info/PKG-INFO

@@ -0,0 +1,62 @@
+Metadata-Version: 2.1
+Name: sbss
+Version: 0.0.1
+Summary: Similarity-Based Stratified Splitting Algorithm
+Home-page: https://github.com/timothyckl/similarity-stratified-split
+Author: timothyckl
+Author-email: timothy.ckl@outlook.com
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Operating System :: OS Independent
+Description-Content-Type: text/markdown
+License-File: LICENSE
+
+# Similarity Stratified Split
+
+Implementation of the Similarity-Based Stratified Splitting algorithm described in [Similarity Based Stratified Splitting: an approach to train better classifiers](https://arxiv.org/abs/2010.06099).
+
+## Overview
+
+The authors propose a Similarity-Based Stratified Splitting (SBSS) technique, which uses both the output and input space information to split a dataset. Splits are generated using similarity functions among samples to place similar samples in different splits. This approach allows for a better representation of the data in the training phase. This strategy leads to a more realistic performance estimation when used in real-world applications.
+
+## Install
+
+<!-- **PyPI**
+```bash
+pip install sbss
+``` -->
+
+**Local**
+
+```
+git clone https://github.com/timothyckl/similarity-stratified-split.git
+cd ./similarity-stratified-split
+pip install -e .
+```
+
+## Usage
+
+```python
+import numpy as np
+from scipy.spatial import distance
+from sbss import SimilarityStratifiedSplit
+
+def get_distances(x):
+    distances = distance.squareform(distance.pdist(x, metric='euclidean'))
+    return distances
+
+# inputs are recommended to be normalized
+X = np.random.rand(1000, 128)
+y = np.random.randint(0, 10, (1000,))
+
+n_splits = 3
+s = SimilarityStratifiedSplit(n_splits, get_distances)
+
+for train_index, test_index in s.split(X, y):
+  print(f"Train indices: {train_index}\nTest indices: {test_index}")
+  print("="*100)
+```
+
+## References
+
+- Farias, F., Ludermir, T. and Bastos-Filho, C. (2020) Similarity based stratified splitting: An approach to train better classifiers, arXiv.org. Available at: https://arxiv.org/abs/2010.06099 (Accessed: 27 November 2023). 

sbss-0.0.1/sbss/sbss.egg-info/SOURCES.txt

@@ -0,0 +1,9 @@
+LICENSE
+README.md
+setup.py
+sbss/sbss.egg-info/PKG-INFO
+sbss/sbss.egg-info/SOURCES.txt
+sbss/sbss.egg-info/dependency_links.txt
+sbss/sbss.egg-info/requires.txt
+sbss/sbss.egg-info/top_level.txt
+tests/test_sbss.py

sbss-0.0.1/sbss/sbss.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

sbss-0.0.1/sbss/sbss.egg-info/requires.txt

@@ -0,0 +1 @@
+numpy==1.23.5

sbss-0.0.1/sbss/sbss.egg-info/top_level.txt

@@ -0,0 +1 @@
+

sbss-0.0.1/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

sbss-0.0.1/setup.py

@@ -0,0 +1,28 @@
+from setuptools import find_packages, setup
+
+with open("./README.md", "r") as f:
+    long_desc = f.read()
+
+setup(
+    name="sbss",
+    version="0.0.1",
+    description="Similarity-Based Stratified Splitting Algorithm",
+    package_dir={"": "sbss"},
+    packages=find_packages(where="sbss"),
+    long_description=long_desc,
+    long_description_content_type='text/markdown',
+    url="https://github.com/timothyckl/similarity-stratified-split",
+    author="timothyckl",
+    author_email="timothy.ckl@outlook.com",
+    license="",
+    classifiers=[
+        "License :: OSI Approved :: MIT License",
+        "Programming Language :: Python :: 3.9",
+        "Operating System :: OS Independent"
+    ],
+    install_requires=["numpy==1.23.5"],
+    extra_require={
+        "dev": ["pytest==7.4.3"]
+    },
+    python_require=">=3.9"
+)

sbss-0.0.1/tests/test_sbss.py

@@ -0,0 +1,72 @@
+import pytest
+import numpy as np
+from sbss import SimilarityStratifiedSplit
+
+def sample_sim_func(X):
+    # Example similarity function returning random values for each pair of samples
+    n_samples = X.shape[0]
+    return np.random.rand(n_samples, n_samples)
+
+class TestSimilarityStratifiedSplit:
+    @staticmethod
+    def test_init():
+        splitter = SimilarityStratifiedSplit(n_splits=3, sim_func=sample_sim_func, shuffle=True)
+        assert splitter.n_splits == 3
+        assert splitter.sim_func == sample_sim_func
+        assert splitter.shuffle is True
+
+    @staticmethod
+    def test_get_n_splits():
+        splitter = SimilarityStratifiedSplit(n_splits=5, sim_func=sample_sim_func)
+        assert splitter.get_n_splits() == 5
+
+    @staticmethod
+    def test_shuffle_dataset():
+        splitter = SimilarityStratifiedSplit(n_splits=2, sim_func=sample_sim_func, shuffle=True)
+        X = np.array([[1, 2], [3, 4], [5, 6], [7, 8]])
+        y = np.array([0, 1, 0, 1])
+        shuffled_X, shuffled_y = splitter._shuffle_dataset(X, y)
+        assert len(shuffled_X) == len(X)
+        assert len(shuffled_y) == len(y)
+
+    @staticmethod
+    def test_validate():
+        splitter = SimilarityStratifiedSplit(n_splits=4, sim_func=sample_sim_func)
+        class_counts = np.array([3, 2])
+        min_samples_per_class = 2
+        with pytest.raises(ValueError):
+            splitter._validate(class_counts, min_samples_per_class)
+
+    @staticmethod
+    def test_encode_labels():
+        splitter = SimilarityStratifiedSplit(n_splits=2, sim_func=sample_sim_func)
+        y = np.array([1, 2, 3, 1, 2, 3])
+        num_classes = splitter._encode_labels(y)
+        assert num_classes == 3
+
+    @staticmethod
+    def test_split():
+        splitter = SimilarityStratifiedSplit(n_splits=2, sim_func=sample_sim_func)
+        X = np.array([[1, 2], [3, 4], [5, 6], [7, 8]])
+        y = np.array([0, 1, 0, 1])
+        splits = list(splitter.split(X, y))
+        assert len(splits) == 2
+        for train_indices, test_indices in splits:
+            assert len(train_indices) + len(test_indices) == len(X)
+            assert len(np.intersect1d(train_indices, test_indices)) == 0
+
+    @staticmethod
+    def test_validate_pass():
+        splitter = SimilarityStratifiedSplit(n_splits=2, sim_func=sample_sim_func)
+        class_counts = np.array([3, 2])
+        min_samples_per_class = 2
+        splitter._validate(class_counts, min_samples_per_class)
+        assert True 
+
+    @staticmethod
+    def test_validate_fail():
+        splitter = SimilarityStratifiedSplit(n_splits=4, sim_func=sample_sim_func)
+        class_counts = np.array([3, 2])
+        min_samples_per_class = 2
+        with pytest.raises(ValueError):
+            splitter._validate(class_counts, min_samples_per_class)