Merge pull request #24 from selimfirat/develop

Release v0.3.3

Author: selimfirat

.travis.yml

@@ -2,4 +2,5 @@ v<0.1.0>, <08/15/2020> -- Initial release.
 v<0.2.0> <06/10/2023> -- Upgraded some libraries like numpy,pandas and small refactorings.
 v<0.3.0>, <05/23/2025> -- Updated dependencies to latest versions, improved documentation, added quick start example in README. Added Inqmad model. Special thanks to @TechyNilesh @Joaggi @onixlas.
 v<0.3.1>, <06/04/2025> -- Enhanced README with comprehensive community engagement section, educational content, third-party integrations, and developer community highlights. Fixed sublist formatting in reStructuredText. Improved error handling for Inqmad model imports.
-v<0.3.2>, <06/14/2025> -- Updated PyOD version and added Python 3.13 support. Fixed Flaticon link in documentation. Enhanced error handling for optional JAX dependencies. Updated rrcf version to 0.4.4 and improved quantization in RelativeEntropy model. Improved test coverage and configuration. Added deprecation warning tests and enhanced model methods array handling. Special thanks to @onixlas for multiple contributions.
+v<0.3.2>, <06/14/2025> -- Updated PyOD version and added Python 3.13 support. Fixed Flaticon link in documentation. Enhanced error handling for optional JAX dependencies. Updated rrcf version to 0.4.4 and improved quantization in RelativeEntropy model. Improved test coverage and configuration. Added deprecation warning tests and enhanced model methods array handling. Special thanks to @onixlas for multiple contributions.
+v<0.3.3>, <06/19/2025> -- Fixed critical bugs in ReferenceWindowModel: resolved data mutation issues by ensuring reference window data is properly copied, fixed window_length parameter naming to window_size, and improved reference window label management. Enhanced test coverage and fixed minor typo in bibtex reference. Special thanks to @onixlas for bug reports and fixes.

CHANGES.txt

@@ -29,6 +29,12 @@ steps:
     versionSpec: '$(python.version)'
   displayName: 'Use Python $(python.version)'
 
+- script: |
+    # Install system dependencies required for building scipy from source
+    sudo apt-get update
+    sudo apt-get install -y libopenblas-dev liblapack-dev gfortran pkg-config
+  displayName: 'Install system dependencies'
+
 - script: |
     python -m pip install --upgrade pip
     pip install -r requirements-ci.txt

azure-pipelines.yml

@@ -131,7 +131,7 @@ @article{constantinou2018pynomaly
 
 @article{pevny2016loda,
   title={Loda: Lightweight on-line detector of anomalies},
-  author={Pevn{\`y}, Tom{\'a}{\v{s}}},
+  author={Pevn{\'y}, Tom{\'a}{\v{s}}},
   journal={Machine Learning},
   volume={102},
   number={2},

docs/refs.bib

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "pysad"
-version = "0.3.2"
+version = "0.3.3"
 dependencies = [
   "numpy",
   "scikit-learn",

pyproject.toml

@@ -3,7 +3,7 @@
 
 
 class ReferenceWindowModel(PYODModel):
-    """This PyOD model wrapper wraps the batch anomaly detectors. This wrapper keeps track of the reference window of size `window_length`. For every `sliding_size` instnaces, it resets the model by training new `model_cls` instance with the reference window. This implementation is based on the reference windowing described in :cite:`xstream`.
+    """This PyOD model wrapper wraps the batch anomaly detectors. This wrapper keeps track of the reference window of size `window_size`. For every `sliding_size` instances, it resets the model by training new `model_cls` instance with the reference window. This implementation is based on the reference windowing described in :cite:`xstream`.
 
         Args:
             model_cls (class): The model class to be instantiated.
@@ -65,8 +65,8 @@ def fit_partial(self, X, y=None):
 
         if not self.initial_ref_window and len(
                 self.cur_window_X) < self.window_size:
-            self.reference_window_X = self.cur_window_X
-            self.reference_window_y = self.cur_window_y if y is not None else None
+            self.reference_window_X = self.cur_window_X.copy()
+            self.reference_window_y = self.cur_window_y.copy() if y is not None else None
             self._fit_model()
         elif len(self.cur_window_X) % self.sliding_size == 0:
             self.reference_window_X = np.concatenate(

pysad/models/integrations/reference_window_model.py

@@ -21,4 +21,4 @@
 # 'X.Y.dev0' is the canonical version of 'X.Y.dev'
 #
 
-__version__ = '0.3.2'
+__version__ = '0.3.3'

pysad/version.py

@@ -219,3 +219,55 @@ def test_reference_window_model_window_update():
 
         # Check internal state
         assert len(model.cur_window_X) <= model.window_size
+
+
+def test_reference_window_model_issue_23_fix():
+    """Test for the fix of issue #23: reference window duplication bug.
+    
+    The bug was that reference_window_X and cur_window_X pointed to the same
+    list object, causing data duplication when concatenation occurred.
+    """
+    from pysad.models.integrations.reference_window_model import ReferenceWindowModel
+    from pyod.models.iforest import IForest
+    import numpy as np
+    
+    # Setup from issue #23
+    window_size = 4
+    sliding_size = 2
+    
+    model = ReferenceWindowModel(
+        model_cls=IForest,
+        window_size=window_size,
+        sliding_size=sliding_size,
+        initial_window_X=None
+    )
+    
+    # Step 1: Add first data point
+    X1 = np.array([1.0])
+    model.fit_partial(X1)
+    
+    # Verify that cur_window_X and reference_window_X are independent objects
+    assert model.cur_window_X is not model.reference_window_X, \
+        "cur_window_X and reference_window_X should be independent objects"
+    
+    # Step 2: Add second data point
+    X2 = np.array([2.0])
+    model.fit_partial(X2)
+    
+    # Verify they are still independent
+    assert model.cur_window_X is not model.reference_window_X, \
+        "cur_window_X and reference_window_X should remain independent objects"
+    
+    # Test independence: modify cur_window_X and ensure reference_window_X is unchanged
+    original_ref_len = len(model.reference_window_X)
+    model.cur_window_X.append(np.array([999.0]))
+    
+    # reference_window_X should not be affected by changes to cur_window_X
+    assert len(model.reference_window_X) == original_ref_len, \
+        "reference_window_X should not be affected by changes to cur_window_X"
+    
+    # Extract values for comparison - should be [1.0, 2.0], NOT [1.0, 2.0, 1.0, 2.0]
+    ref_values = [x[0] for x in model.reference_window_X]
+    expected = [1.0, 2.0]
+    assert ref_values == expected, \
+        f"Expected {expected}, but got {ref_values}. Duplication detected!"