MDL-Density-Histogram 1.1.4__tar.gz

mdl_density_histogram-1.1.4/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

mdl_density_histogram-1.1.4/PKG-INFO

@@ -0,0 +1,108 @@
+Metadata-Version: 2.4
+Name: MDL-Density-Histogram
+Version: 1.1.4
+Summary: Cython-accelerated MDL histogram density estimation with dynamic programming, implementing Kontkanen & Myllymaki's algorithm (JMLR 2007).
+Author-email: Götz Grimmer <goetz-dev@web.de>
+Maintainer-email: Götz Grimmer <goetz-dev@web.de>
+License: Apache-2.0
+Project-URL: Repository, https://github.com/MrTarantoga/MDL-Density-Histogram
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Provides-Extra: tests
+Requires-Dist: pytest; extra == "tests"
+Requires-Dist: pytest-xdist; extra == "tests"
+Requires-Dist: pandas; extra == "tests"
+Requires-Dist: pyarrow; extra == "tests"
+Requires-Dist: brotli; extra == "tests"
+Dynamic: license-file
+
+[![Upload Python Package](https://github.com/MrTarantoga/MDL-Density-Histogram/actions/workflows/python-publish.yml/badge.svg?event=release)](https://github.com/MrTarantoga/MDL-Density-Histogram/actions/workflows/python-publish.yml)
+[![Python application test](https://github.com/MrTarantoga/MDL-Density-Histogram/actions/workflows/python-app.yml/badge.svg)](https://github.com/MrTarantoga/MDL-Density-Histogram/actions/workflows/python-app.yml)
+
+# MDL Optimal Histogram Density Estimation
+
+This package provides a Cython-accelerated implementation of the **Minimum Description Length (MDL) optimal histogram density estimation** algorithm from Kontkanen & Myllymaki (2007). It uses information-theoretic principles to automatically determine optimal variable-width bins for density estimation.
+
+![Freedman-Diaconis vs. MDL-Optimization](https://raw.githubusercontent.com/MrTarantoga/MDL-Density-Histogram/main/gmm5_idx_3.png)
+
+## Features
+- **MDL Principle**: Uses stochastic complexity for model selection
+- **Dynamic Programming**: Efficient O(E²·K_max) optimization (cache parametric complexity computation, speed up)
+- **Score of each *K*th bin**: The score of each bin is returned to understand the performance of different properties of the same dataset.
+- **Variable-Width Bins**: Adapts to data density variations
+- **Automatic Bin Count**: No manual parameter tuning required (except maximum bin count to consider $K_{max}$ and data resolution $\epsilon$)
+- **Cython Acceleration**: Critical operations compiled to C
+
+## Installation
+You can install the package using pip:
+```bash
+pip install MDL-Density-Histogram
+```
+Alternatively, you can install it from source by cloning the repository and running:
+```bash
+# From project root directory
+pip install .
+```
+
+Requires:
+- Python 3.11+
+- NumPy
+- Cython
+- C compiler (GCC/Clang/MSVC)
+
+## Usage Example
+```python
+import numpy as np
+from mdl_density_hist import mdl_optimal_histogram
+
+# Generate sample data
+data = np.random.normal(0, 1, 1000)
+
+# Compute optimal histogram
+cut_points, K_scores = mdl_optimal_histogram(data, epsilon=0.1)
+
+# Print score of each bin
+print(f"K_scores: {K_scores}")
+
+# Visualize result
+import matplotlib.pyplot as plt
+plt.hist(data, bins=cut_points, density=True)
+plt.title('MDL Optimal Histogram')
+plt.show()
+```
+
+## Parameters
+- `data`: Input array (1D numpy array)
+- `epsilon`: Quantization precision (default: 0.1)
+- `K_max`: Maximum number of bins (default: 10)
+
+## Algorithm Highlights
+- Uses **Ramanujan's factorial approximation** for efficient parametric complexity
+- Cache parameteric complexity to speed up computation
+
+## Paper Citation
+Kontkanen, P., & Myllymäki, P. (2007).  
+*MDL Histogram Density Estimation*  
+Journal of Machine Learning Research 8 (2007)
+[PDF](https://proceedings.mlr.press/v2/kontkanen07a/kontkanen07a.pdf)
+
+## License
+Apache 2.0 License - See LICENSE file
+
+## Project Structure
+```
+src/
+├── mdl_density_hist/
+│   ├── __init__.py
+│   └── mdl_hist.pyx  # Core Cython implementation
+└── pyproject.toml
+```
+
+## Performance Notes
+- Precomputed parametric complexity using dynamic programming
+- Memory-optimized array operations via NumPy
+- Candidate cut point pruning for reduced search space
+
+
+For implementation details, see the [paper](https://proceedings.mlr.press/v2/kontkanen07a/kontkanen07a.pdf) and inline code comments.

mdl_density_histogram-1.1.4/README.md

@@ -0,0 +1,89 @@
+[![Upload Python Package](https://github.com/MrTarantoga/MDL-Density-Histogram/actions/workflows/python-publish.yml/badge.svg?event=release)](https://github.com/MrTarantoga/MDL-Density-Histogram/actions/workflows/python-publish.yml)
+[![Python application test](https://github.com/MrTarantoga/MDL-Density-Histogram/actions/workflows/python-app.yml/badge.svg)](https://github.com/MrTarantoga/MDL-Density-Histogram/actions/workflows/python-app.yml)
+
+# MDL Optimal Histogram Density Estimation
+
+This package provides a Cython-accelerated implementation of the **Minimum Description Length (MDL) optimal histogram density estimation** algorithm from Kontkanen & Myllymaki (2007). It uses information-theoretic principles to automatically determine optimal variable-width bins for density estimation.
+
+![Freedman-Diaconis vs. MDL-Optimization](https://raw.githubusercontent.com/MrTarantoga/MDL-Density-Histogram/main/gmm5_idx_3.png)
+
+## Features
+- **MDL Principle**: Uses stochastic complexity for model selection
+- **Dynamic Programming**: Efficient O(E²·K_max) optimization (cache parametric complexity computation, speed up)
+- **Score of each *K*th bin**: The score of each bin is returned to understand the performance of different properties of the same dataset.
+- **Variable-Width Bins**: Adapts to data density variations
+- **Automatic Bin Count**: No manual parameter tuning required (except maximum bin count to consider $K_{max}$ and data resolution $\epsilon$)
+- **Cython Acceleration**: Critical operations compiled to C
+
+## Installation
+You can install the package using pip:
+```bash
+pip install MDL-Density-Histogram
+```
+Alternatively, you can install it from source by cloning the repository and running:
+```bash
+# From project root directory
+pip install .
+```
+
+Requires:
+- Python 3.11+
+- NumPy
+- Cython
+- C compiler (GCC/Clang/MSVC)
+
+## Usage Example
+```python
+import numpy as np
+from mdl_density_hist import mdl_optimal_histogram
+
+# Generate sample data
+data = np.random.normal(0, 1, 1000)
+
+# Compute optimal histogram
+cut_points, K_scores = mdl_optimal_histogram(data, epsilon=0.1)
+
+# Print score of each bin
+print(f"K_scores: {K_scores}")
+
+# Visualize result
+import matplotlib.pyplot as plt
+plt.hist(data, bins=cut_points, density=True)
+plt.title('MDL Optimal Histogram')
+plt.show()
+```
+
+## Parameters
+- `data`: Input array (1D numpy array)
+- `epsilon`: Quantization precision (default: 0.1)
+- `K_max`: Maximum number of bins (default: 10)
+
+## Algorithm Highlights
+- Uses **Ramanujan's factorial approximation** for efficient parametric complexity
+- Cache parameteric complexity to speed up computation
+
+## Paper Citation
+Kontkanen, P., & Myllymäki, P. (2007).  
+*MDL Histogram Density Estimation*  
+Journal of Machine Learning Research 8 (2007)
+[PDF](https://proceedings.mlr.press/v2/kontkanen07a/kontkanen07a.pdf)
+
+## License
+Apache 2.0 License - See LICENSE file
+
+## Project Structure
+```
+src/
+├── mdl_density_hist/
+│   ├── __init__.py
+│   └── mdl_hist.pyx  # Core Cython implementation
+└── pyproject.toml
+```
+
+## Performance Notes
+- Precomputed parametric complexity using dynamic programming
+- Memory-optimized array operations via NumPy
+- Candidate cut point pruning for reduced search space
+
+
+For implementation details, see the [paper](https://proceedings.mlr.press/v2/kontkanen07a/kontkanen07a.pdf) and inline code comments.

mdl_density_histogram-1.1.4/pyproject.toml

@@ -0,0 +1,45 @@
+[build-system]
+requires = ["setuptools", "cython", "wheel", "numpy"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "MDL-Density-Histogram"
+version = "1.1.4"
+requires-python = ">= 3.11"
+description = "Cython-accelerated MDL histogram density estimation with dynamic programming, implementing Kontkanen & Myllymaki's algorithm (JMLR 2007)."
+authors = [
+    {name = "Götz Grimmer", email = "goetz-dev@web.de"}
+]
+maintainers = [
+    {name = "Götz Grimmer", email = "goetz-dev@web.de"}
+]
+license = {text = "Apache-2.0"}
+readme = "README.md"
+
+[project.urls]
+Repository = "https://github.com/MrTarantoga/MDL-Density-Histogram"
+
+[project.optional-dependencies]
+tests = [
+    "pytest",
+    "pytest-xdist",
+    "pandas",
+    "pyarrow",
+    "brotli"
+]
+
+[tool.setuptools.packages.find]
+where = ["src"]
+
+[tool.setuptools]
+package-dir = {"" = "src"}
+
+[tool.cibuildwheel.linux.environment]
+CFLAGS = "-O3 -funroll-loops -flto -ftree-vectorize"
+
+[tool.cibuildwheel.macos.environment]
+CFLAGS = "-O3 -funroll-loops -flto -ftree-vectorize"
+
+[tool.cibuildwheel.windows.environment]
+# Use /O2 instead of -O3 for MSVC; /GL enables link-time optimization
+CFLAGS = "/O2 /GL"

mdl_density_histogram-1.1.4/setup.cfg

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

mdl_density_histogram-1.1.4/setup.py

@@ -0,0 +1,25 @@
+from setuptools import setup, Extension
+# Use Cython.Build instead of Cython.Distutils
+from Cython.Build import build_ext
+import numpy as np
+
+# Define the source directory relative to setup.py
+SRC_DIR = "src/mdl_density_hist"
+
+# Define the Cython extension
+ext_1 = Extension(
+    name="mdl_density_hist.mdl_hist",
+    sources=[SRC_DIR + "/mdl_hist.pyx"],
+    libraries=[],
+    include_dirs=[np.get_include()]
+)
+
+EXTENSIONS = [ext_1]
+
+# Setup configuration focused only on building the extension
+# Remove the __name__ == "__main__" guard
+setup(
+    # Removed packages and package_dir
+    cmdclass={"build_ext": build_ext},
+    ext_modules=EXTENSIONS
+)

mdl_density_histogram-1.1.4/src/MDL_Density_Histogram.egg-info/PKG-INFO

@@ -0,0 +1,108 @@
+Metadata-Version: 2.4
+Name: MDL-Density-Histogram
+Version: 1.1.4
+Summary: Cython-accelerated MDL histogram density estimation with dynamic programming, implementing Kontkanen & Myllymaki's algorithm (JMLR 2007).
+Author-email: Götz Grimmer <goetz-dev@web.de>
+Maintainer-email: Götz Grimmer <goetz-dev@web.de>
+License: Apache-2.0
+Project-URL: Repository, https://github.com/MrTarantoga/MDL-Density-Histogram
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Provides-Extra: tests
+Requires-Dist: pytest; extra == "tests"
+Requires-Dist: pytest-xdist; extra == "tests"
+Requires-Dist: pandas; extra == "tests"
+Requires-Dist: pyarrow; extra == "tests"
+Requires-Dist: brotli; extra == "tests"
+Dynamic: license-file
+
+[![Upload Python Package](https://github.com/MrTarantoga/MDL-Density-Histogram/actions/workflows/python-publish.yml/badge.svg?event=release)](https://github.com/MrTarantoga/MDL-Density-Histogram/actions/workflows/python-publish.yml)
+[![Python application test](https://github.com/MrTarantoga/MDL-Density-Histogram/actions/workflows/python-app.yml/badge.svg)](https://github.com/MrTarantoga/MDL-Density-Histogram/actions/workflows/python-app.yml)
+
+# MDL Optimal Histogram Density Estimation
+
+This package provides a Cython-accelerated implementation of the **Minimum Description Length (MDL) optimal histogram density estimation** algorithm from Kontkanen & Myllymaki (2007). It uses information-theoretic principles to automatically determine optimal variable-width bins for density estimation.
+
+![Freedman-Diaconis vs. MDL-Optimization](https://raw.githubusercontent.com/MrTarantoga/MDL-Density-Histogram/main/gmm5_idx_3.png)
+
+## Features
+- **MDL Principle**: Uses stochastic complexity for model selection
+- **Dynamic Programming**: Efficient O(E²·K_max) optimization (cache parametric complexity computation, speed up)
+- **Score of each *K*th bin**: The score of each bin is returned to understand the performance of different properties of the same dataset.
+- **Variable-Width Bins**: Adapts to data density variations
+- **Automatic Bin Count**: No manual parameter tuning required (except maximum bin count to consider $K_{max}$ and data resolution $\epsilon$)
+- **Cython Acceleration**: Critical operations compiled to C
+
+## Installation
+You can install the package using pip:
+```bash
+pip install MDL-Density-Histogram
+```
+Alternatively, you can install it from source by cloning the repository and running:
+```bash
+# From project root directory
+pip install .
+```
+
+Requires:
+- Python 3.11+
+- NumPy
+- Cython
+- C compiler (GCC/Clang/MSVC)
+
+## Usage Example
+```python
+import numpy as np
+from mdl_density_hist import mdl_optimal_histogram
+
+# Generate sample data
+data = np.random.normal(0, 1, 1000)
+
+# Compute optimal histogram
+cut_points, K_scores = mdl_optimal_histogram(data, epsilon=0.1)
+
+# Print score of each bin
+print(f"K_scores: {K_scores}")
+
+# Visualize result
+import matplotlib.pyplot as plt
+plt.hist(data, bins=cut_points, density=True)
+plt.title('MDL Optimal Histogram')
+plt.show()
+```
+
+## Parameters
+- `data`: Input array (1D numpy array)
+- `epsilon`: Quantization precision (default: 0.1)
+- `K_max`: Maximum number of bins (default: 10)
+
+## Algorithm Highlights
+- Uses **Ramanujan's factorial approximation** for efficient parametric complexity
+- Cache parameteric complexity to speed up computation
+
+## Paper Citation
+Kontkanen, P., & Myllymäki, P. (2007).  
+*MDL Histogram Density Estimation*  
+Journal of Machine Learning Research 8 (2007)
+[PDF](https://proceedings.mlr.press/v2/kontkanen07a/kontkanen07a.pdf)
+
+## License
+Apache 2.0 License - See LICENSE file
+
+## Project Structure
+```
+src/
+├── mdl_density_hist/
+│   ├── __init__.py
+│   └── mdl_hist.pyx  # Core Cython implementation
+└── pyproject.toml
+```
+
+## Performance Notes
+- Precomputed parametric complexity using dynamic programming
+- Memory-optimized array operations via NumPy
+- Candidate cut point pruning for reduced search space
+
+
+For implementation details, see the [paper](https://proceedings.mlr.press/v2/kontkanen07a/kontkanen07a.pdf) and inline code comments.

mdl_density_histogram-1.1.4/src/MDL_Density_Histogram.egg-info/SOURCES.txt

@@ -0,0 +1,12 @@
+LICENSE
+README.md
+pyproject.toml
+setup.py
+src/MDL_Density_Histogram.egg-info/PKG-INFO
+src/MDL_Density_Histogram.egg-info/SOURCES.txt
+src/MDL_Density_Histogram.egg-info/dependency_links.txt
+src/MDL_Density_Histogram.egg-info/requires.txt
+src/MDL_Density_Histogram.egg-info/top_level.txt
+src/mdl_density_hist/__init__.py
+src/mdl_density_hist/mdl_hist.pyx
+tests/test_main.py

mdl_density_histogram-1.1.4/src/MDL_Density_Histogram.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

mdl_density_histogram-1.1.4/src/MDL_Density_Histogram.egg-info/requires.txt

@@ -0,0 +1,7 @@
+
+[tests]
+pytest
+pytest-xdist
+pandas
+pyarrow
+brotli

mdl_density_histogram-1.1.4/src/MDL_Density_Histogram.egg-info/top_level.txt

@@ -0,0 +1 @@
+mdl_density_hist

mdl_density_histogram-1.1.4/src/mdl_density_hist/__init__.py

@@ -0,0 +1 @@
+from .mdl_hist import mdl_optimal_histogram

mdl_density_histogram-1.1.4/src/mdl_density_hist/mdl_hist.pyx

@@ -0,0 +1,255 @@
+# cython: language_level=3
+# cython: boundscheck=False
+# cython: wraparound=False
+# cython: nonecheck=False
+# cython: binding=False
+# cython: cdivision=True
+
+import numpy as np
+cimport cython
+from libc.math cimport floor, log, INFINITY, exp
+
+cdef const double PI = 3.1415926535897932384626433832795028841971693993751058209749445923078164062
+
+cdef double log_fac_by_Ramanujan(const unsigned long long n):  
+    assert n >= 0, "N is only allowed to be [0, inf)"
+    cdef double term_1, term_2, term_3, result
+
+    if n <= 1:
+        return log(1.0)
+    else:
+        term_1 = <double>n * log(n) - <double>n
+        term_2 = log(<double>n * <double>(1 + 4 * n * (1 + 2 * n))) / 6.0
+        term_3 = log(PI) / 2.0
+        result = term_1 + term_2 + term_3
+        return result
+
+def quantize_data(const double [:] x, const double epsilon):
+    assert len(x) > 2, "The x must contain [2, inf)"
+    assert epsilon > 0, "Epsilon must be positive"
+
+    n_bins = int(np.ceil((np.max(x) - np.min(x)) / epsilon)) + 1
+    bins = np.arange(
+        np.min(x), np.max(x) + n_bins * epsilon + 1e-10, epsilon
+    )
+    data_disc = []
+    for lower_bin, upper_bin in zip(bins, bins[1:]):
+       number_bins = np.argwhere(np.logical_and(x < upper_bin, x >= lower_bin) == True).shape[0]
+       data_disc += [lower_bin]*number_bins
+    return np.asarray(data_disc)
+
+def generate_candidate_cut_points(const double [:] x, const double epsilon):
+    """Generate candidate cut points between data points (Equation 22)"""
+    assert len(x) > 2, "The x must contain [2, inf)"
+    assert epsilon > 0, "Epsilon must be positive"
+
+    x = np.sort(x)
+    candidates = []
+    
+    for _x in x:
+        candidates.extend([_x - epsilon/2, _x + epsilon/2])
+    
+    # Remove duplicates and sort
+    candidates = sorted(list(set(candidates)))
+    
+    # Remove implicit boundaries
+    implicit_lower = min(x) - epsilon/2 
+    candidates = np.delete(candidates, np.argwhere(candidates == implicit_lower).flatten())
+
+    return candidates
+
+cpdef unsigned long long[:] precompute_n_e(const double [:] data, const double [:] candidates):
+    """Precompute n_e: number of data points in [x_min, c_e] (Section 4)"""
+    n_e = np.zeros(len(candidates), dtype=np.uint64)  # Include E+1
+    cdef Py_ssize_t i, d
+    cdef double c
+
+    for i, c in enumerate(candidates):
+        # Count data 
+        for d in range(len(data)):
+            if data[d] < c:
+                n_e[i] += 1
+    return n_e
+
+cdef double compute_parametric_complexity(const unsigned long long n, const unsigned long long K):
+    cdef double R_prev2 = 1.0  # R_n_h(K-2)
+    cdef double R_current = 0.0  # R_n_h(K)
+    cdef double R_prev1 = 0.0  # R_n_h(K-1)
+    cdef unsigned long long h2, h1, k
+    cdef double term_1, term_2, term_3
+    cdef double log_n = log(n)
+    
+    if K == 0:
+        return 1.0
+
+    if K >= 1:
+        # Precompute log(n) once for all h2 iterations
+        for h2 in range(n + 1):
+            h1 = n - h2
+            term_1 = log_fac_by_Ramanujan(n) - (log_fac_by_Ramanujan(h1) + log_fac_by_Ramanujan(h2))
+            
+            if h1 != 0:
+                term_2 = (log(<double>h1) - log_n) * <double>h1
+            else:
+                term_2 = 0.0
+                
+            if h2 != 0:
+                term_3 = (log(<double>h2) - log_n) * <double>h2
+            else:
+                term_3 = 0.0
+                
+            R_current += exp(term_1 + term_2 + term_3)
+        
+    if K >= 2:
+        R_prev1 = R_current  # R_n_h(1)
+        for k in range(2, K + 1):
+            R_current = R_prev1 + (<double>n / <double>(k-1)) * R_prev2
+            R_prev2 = R_prev1
+            R_prev1 = R_current
+
+    return R_current
+
+cdef double dp_func_init(const unsigned long long [:] n_e, 
+                         const double [:] x, 
+                         const unsigned long long n, 
+                         const double [:] candidates, 
+                         const unsigned long long e, 
+                         const double epsilon):
+    cdef double term1, term2, best_score, candidates_val, x_min
+    cdef unsigned long long n_e_val = <unsigned long long>n_e[e]
+
+    if n_e_val != 0:
+        candidates_val = <double>candidates[e]
+        x_min = <double>np.min(x)
+        
+        term1 = log(epsilon * <double>n_e_val)
+        term2 = log((candidates_val - (x_min - epsilon / 2.0)) * <double>n)
+        best_score = -<double>n_e_val * (term1 - term2)
+        return best_score
+    else:
+        return 0.0
+
+cdef tuple dp_func(const unsigned long long [:] n_e, 
+                   const double [:] x, 
+                   const unsigned long long n, 
+                   const double [:] candidates, 
+                   const unsigned long long K, 
+                   const unsigned long long E, 
+                   const unsigned long long e, 
+                   const double epsilon, 
+                   double[:, :] DP_table,
+                   dict[tuple[unsigned long long, unsigned long long], double] cache):
+    cdef double e_prime_best_score = INFINITY
+    cdef double best_score = INFINITY
+    cdef unsigned long long e_prime
+    cdef unsigned long long n_k
+    cdef double bin_width
+    cdef double term1, term2, term3, score
+    cdef double R, R_prime
+
+    for e_prime in range(K-1, e):
+        n_k = n_e[e] - n_e[e_prime]
+        bin_width = candidates[e] - candidates[e_prime]
+
+        if n_k != 0:
+            term1 = <double>n_k * (log(epsilon * <double>n_k) - log(bin_width * <double>n))
+        else:
+            term1 = 0.0
+
+        if (n_e[e], K) in cache:
+            R = cache[n_e[e], K]
+        else:
+            R = compute_parametric_complexity(n_e[e], K)
+            cache[n_e[e], K] = R
+
+        if (n_e[e_prime], K-1) in cache:
+            R_prime = cache[n_e[e_prime], K-1]
+        else:
+            R_prime = compute_parametric_complexity(n_e[e_prime], K - 1)
+            cache[n_e[e_prime], K-1] = R_prime
+
+        term2 = log(R / R_prime)
+        term3 = log(<double>(E - K + 4) / <double>K)
+
+        assert DP_table[K-1, e_prime] != INFINITY, "DP_table contain a infinity error, that is not allowed"
+
+        score = DP_table[K-1, e_prime] - term1 + term2 + term3
+        if score < best_score:
+            e_prime_best_score = <double>e_prime
+            best_score = score
+
+    return (e_prime_best_score, best_score)
+
+
+
+def mdl_optimal_histogram(const double [:] data, 
+                          const double epsilon=0.1, 
+                          const unsigned long long K_max=10):
+    cdef double[:] K_scores
+    cdef unsigned long long n = data.shape[0]
+    cdef unsigned long long i
+    cdef unsigned long long K, e, K_best, e_pos,
+    cdef double[:] candidates
+    cdef unsigned long long[:] n_e
+    cdef unsigned long long E
+    cdef double[:, :] dp_table_score
+    cdef int[:, :] dp_table_e_prime
+    cdef double best_score
+    cdef list optimal_cut_points
+
+    # Check for valid input data
+    assert len(data) > 2, "The data must contain more than 2 datapoints"
+    assert epsilon > 0, "The epsilon must be positive"
+
+    # Step 0: Create lookup dictionary for parametric compute_parametric_complexity
+    cdef dict[tuple[unsigned long long, unsigned long long], double] cpc_cache = {}
+
+    # Step 1: Quantize data
+    data = quantize_data(data, epsilon)
+
+    # Step 3: Generate candidate cut points
+    candidates = generate_candidate_cut_points(data, epsilon)
+    E = candidates.shape[0] - 1  # Exclude implicit outer boundary
+
+    # Step 4: Precompute n_e
+    n_e = precompute_n_e(data, candidates)
+
+    assert K_max > 1, "K_max must be [2, inf)"
+
+    # Step 5: Initialize DP tables
+    dp_table_score = np.full((K_max + 1, E + 1), INFINITY, dtype=np.float64)
+    dp_table_e_prime = np.full((K_max + 1, E + 1), -1, dtype=np.int32)
+
+    # Initialize DP table for K=0
+    for e in range(E + 1):
+        dp_table_score[0, e] = dp_func_init(n_e, data, n, candidates, e, epsilon)
+
+    # Fill DP table for K >= 1
+    for K in range(1, K_max + 1):
+        for e in range(K, E + 1):
+            dp_table_e_prime[K, e], dp_table_score[K, e] = dp_func(
+                n_e, data, n, candidates, K, E, e, epsilon, dp_table_score, cpc_cache
+            )
+
+    # Find best K
+    K_best = 0
+    K_scores = dp_table_score[:, E]
+
+    best_score = dp_table_score[0, E]
+    for K in range(1, K_max + 1):
+        if dp_table_score[K, E] < best_score:
+            best_score = dp_table_score[K, E]
+            K_best = K
+
+    # Backtrack to find optimal cut points
+    optimal_cut_points = [candidates[E]]
+    e_pos = E
+    for K in range(K_best - 1, 0, -1):
+        e_pos = dp_table_e_prime[K, e_pos]
+        optimal_cut_points.append(candidates[e_pos])
+
+    # Add min and sort
+    optimal_cut_points.append(np.min(data))
+    optimal_cut_points.sort()
+
+    return np.array(optimal_cut_points, dtype=np.float64), np.array(K_scores, dtype=np.float64)

mdl_density_histogram-1.1.4/tests/test_main.py

@@ -0,0 +1,197 @@
+import numpy as np
+import pandas as pd
+import pytest
+from pathlib import Path
+from mdl_density_hist import mdl_optimal_histogram
+
+import sys
+
+sys.dont_write_bytecode = True
+
+K_max = 100
+epsilon = 0.1
+
+@pytest.fixture
+def rootdir_path(request):
+    return request.config.rootdir
+
+def test_gmm3_id_0(rootdir_path):
+    gmmX_ds = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm3_samples.parquet.brotli")
+    gmmX_mdl_lower_bins = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm3_samples_MDL_lower_bins.parquet.brotli")
+    id = 0
+
+    assert all(np.unique(gmmX_ds["dataset_id"]) == np.unique(gmmX_mdl_lower_bins["dataset_id"])), "The number of dataset ids must equal"
+    dataset = gmmX_ds[gmmX_ds["dataset_id"] == id]["value"].to_numpy()
+    lower_bins_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["lower_bin"].dropna().to_numpy()
+    K_scores_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["K_score"].to_numpy()
+
+    lower_bins_pred, K_scores_pred = mdl_optimal_histogram(dataset, K_max=K_max, epsilon=epsilon)
+
+    np.testing.assert_array_almost_equal(lower_bins_pred, lower_bins_true, decimal=5)
+    np.testing.assert_allclose(K_scores_pred, K_scores_true, rtol=1e-5)
+
+def test_gmm3_id_1(rootdir_path):
+    gmmX_ds = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm3_samples.parquet.brotli")
+    gmmX_mdl_lower_bins = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm3_samples_MDL_lower_bins.parquet.brotli")
+    id = 1
+
+    assert all(np.unique(gmmX_ds["dataset_id"]) == np.unique(gmmX_mdl_lower_bins["dataset_id"])), "The number of dataset ids must equal"
+    dataset = gmmX_ds[gmmX_ds["dataset_id"] == id]["value"].to_numpy()
+    lower_bins_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["lower_bin"].dropna().to_numpy()
+    K_scores_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["K_score"].to_numpy()
+
+    lower_bins_pred, K_scores_pred = mdl_optimal_histogram(dataset, K_max=K_max, epsilon=epsilon)
+
+    np.testing.assert_array_almost_equal(lower_bins_pred, lower_bins_true, decimal=5)
+    np.testing.assert_allclose(K_scores_pred, K_scores_true, rtol=1e-5)
+
+def test_gmm3_id_2(rootdir_path):
+    gmmX_ds = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm3_samples.parquet.brotli")
+    gmmX_mdl_lower_bins = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm3_samples_MDL_lower_bins.parquet.brotli")
+    id = 2
+
+    assert all(np.unique(gmmX_ds["dataset_id"]) == np.unique(gmmX_mdl_lower_bins["dataset_id"])), "The number of dataset ids must equal"
+    dataset = gmmX_ds[gmmX_ds["dataset_id"] == id]["value"].to_numpy()
+    lower_bins_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["lower_bin"].dropna().to_numpy()
+    K_scores_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["K_score"].to_numpy()
+
+    lower_bins_pred, K_scores_pred = mdl_optimal_histogram(dataset, K_max=K_max, epsilon=epsilon)
+
+    np.testing.assert_array_almost_equal(lower_bins_pred, lower_bins_true, decimal=5)
+    np.testing.assert_allclose(K_scores_pred, K_scores_true, rtol=1e-5)
+
+
+def test_gmm4_id_0(rootdir_path):
+    gmmX_ds = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm4_samples.parquet.brotli")
+    gmmX_mdl_lower_bins = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm4_samples_MDL_lower_bins.parquet.brotli")
+    id = 0
+
+    assert all(np.unique(gmmX_ds["dataset_id"]) == np.unique(gmmX_mdl_lower_bins["dataset_id"])), "The number of dataset ids must equal"
+    dataset = gmmX_ds[gmmX_ds["dataset_id"] == id]["value"].to_numpy()
+    lower_bins_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["lower_bin"].dropna().to_numpy()
+    K_scores_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["K_score"].to_numpy()
+
+    lower_bins_pred, K_scores_pred = mdl_optimal_histogram(dataset, K_max=K_max, epsilon=epsilon)
+
+    np.testing.assert_array_almost_equal(lower_bins_pred, lower_bins_true, decimal=5)
+    np.testing.assert_allclose(K_scores_pred, K_scores_true, rtol=1e-5)
+
+def test_gmm4_id_1(rootdir_path):
+    gmmX_ds = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm4_samples.parquet.brotli")
+    gmmX_mdl_lower_bins = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm4_samples_MDL_lower_bins.parquet.brotli")
+    id = 1
+
+    assert all(np.unique(gmmX_ds["dataset_id"]) == np.unique(gmmX_mdl_lower_bins["dataset_id"])), "The number of dataset ids must equal"
+    dataset = gmmX_ds[gmmX_ds["dataset_id"] == id]["value"].to_numpy()
+    lower_bins_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["lower_bin"].dropna().to_numpy()
+    K_scores_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["K_score"].to_numpy()
+
+    lower_bins_pred, K_scores_pred = mdl_optimal_histogram(dataset, K_max=K_max, epsilon=epsilon)
+
+    np.testing.assert_array_almost_equal(lower_bins_pred, lower_bins_true, decimal=5)
+    np.testing.assert_allclose(K_scores_pred, K_scores_true, rtol=1e-5)
+
+def test_gmm4_id_2(rootdir_path):
+    gmmX_ds = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm4_samples.parquet.brotli")
+    gmmX_mdl_lower_bins = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm4_samples_MDL_lower_bins.parquet.brotli")
+    id = 2
+
+    assert all(np.unique(gmmX_ds["dataset_id"]) == np.unique(gmmX_mdl_lower_bins["dataset_id"])), "The number of dataset ids must equal"
+    dataset = gmmX_ds[gmmX_ds["dataset_id"] == id]["value"].to_numpy()
+    lower_bins_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["lower_bin"].dropna().to_numpy()
+    K_scores_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["K_score"].to_numpy()
+
+    lower_bins_pred, K_scores_pred = mdl_optimal_histogram(dataset, K_max=K_max, epsilon=epsilon)
+
+    np.testing.assert_array_almost_equal(lower_bins_pred, lower_bins_true, decimal=5)
+    np.testing.assert_allclose(K_scores_pred, K_scores_true, rtol=1e-5)
+
+def test_gmm4_id_3(rootdir_path):
+    gmmX_ds = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm4_samples.parquet.brotli")
+    gmmX_mdl_lower_bins = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm4_samples_MDL_lower_bins.parquet.brotli")
+    id = 3
+
+    assert all(np.unique(gmmX_ds["dataset_id"]) == np.unique(gmmX_mdl_lower_bins["dataset_id"])), "The number of dataset ids must equal"
+    dataset = gmmX_ds[gmmX_ds["dataset_id"] == id]["value"].to_numpy()
+    lower_bins_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["lower_bin"].dropna().to_numpy()
+    K_scores_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["K_score"].to_numpy()
+
+    lower_bins_pred, K_scores_pred = mdl_optimal_histogram(dataset, K_max=K_max, epsilon=epsilon)
+
+    np.testing.assert_array_almost_equal(lower_bins_pred, lower_bins_true, decimal=5)
+    np.testing.assert_allclose(K_scores_pred, K_scores_true, rtol=1e-5)
+
+def test_gmm5_id_0(rootdir_path):
+    gmmX_ds = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm5_samples.parquet.brotli")
+    gmmX_mdl_lower_bins = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm5_samples_MDL_lower_bins.parquet.brotli")
+    id = 0
+
+    assert all(np.unique(gmmX_ds["dataset_id"]) == np.unique(gmmX_mdl_lower_bins["dataset_id"])), "The number of dataset ids must equal"
+    dataset = gmmX_ds[gmmX_ds["dataset_id"] == id]["value"].to_numpy()
+    lower_bins_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["lower_bin"].dropna().to_numpy()
+    K_scores_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["K_score"].to_numpy()
+
+    lower_bins_pred, K_scores_pred = mdl_optimal_histogram(dataset, K_max=K_max, epsilon=epsilon)
+
+    np.testing.assert_array_almost_equal(lower_bins_pred, lower_bins_true, decimal=5)
+    np.testing.assert_allclose(K_scores_pred, K_scores_true, rtol=1e-5)
+
+def test_gmm5_id_1(rootdir_path):
+    gmmX_ds = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm5_samples.parquet.brotli")
+    gmmX_mdl_lower_bins = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm5_samples_MDL_lower_bins.parquet.brotli")
+    id = 1
+
+    assert all(np.unique(gmmX_ds["dataset_id"]) == np.unique(gmmX_mdl_lower_bins["dataset_id"])), "The number of dataset ids must equal"
+    dataset = gmmX_ds[gmmX_ds["dataset_id"] == id]["value"].to_numpy()
+    lower_bins_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["lower_bin"].dropna().to_numpy()
+    K_scores_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["K_score"].to_numpy()
+
+    lower_bins_pred, K_scores_pred = mdl_optimal_histogram(dataset, K_max=K_max, epsilon=epsilon)
+
+    np.testing.assert_array_almost_equal(lower_bins_pred, lower_bins_true, decimal=5)
+    np.testing.assert_allclose(K_scores_pred, K_scores_true, rtol=1e-5)
+
+def test_gmm5_id_2(rootdir_path):
+    gmmX_ds = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm5_samples.parquet.brotli")
+    gmmX_mdl_lower_bins = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm5_samples_MDL_lower_bins.parquet.brotli")
+    id = 2
+
+    assert all(np.unique(gmmX_ds["dataset_id"]) == np.unique(gmmX_mdl_lower_bins["dataset_id"])), "The number of dataset ids must equal"
+    dataset = gmmX_ds[gmmX_ds["dataset_id"] == id]["value"].to_numpy()
+    lower_bins_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["lower_bin"].dropna().to_numpy()
+    K_scores_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["K_score"].to_numpy()
+
+    lower_bins_pred, K_scores_pred = mdl_optimal_histogram(dataset, K_max=K_max, epsilon=epsilon)
+
+    np.testing.assert_array_almost_equal(lower_bins_pred, lower_bins_true, decimal=5)
+    np.testing.assert_allclose(K_scores_pred, K_scores_true, rtol=1e-5)
+
+def test_gmm5_id_3(rootdir_path):
+    gmmX_ds = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm5_samples.parquet.brotli")
+    gmmX_mdl_lower_bins = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm5_samples_MDL_lower_bins.parquet.brotli")
+    id = 3
+
+    assert all(np.unique(gmmX_ds["dataset_id"]) == np.unique(gmmX_mdl_lower_bins["dataset_id"])), "The number of dataset ids must equal"
+    dataset = gmmX_ds[gmmX_ds["dataset_id"] == id]["value"].to_numpy()
+    lower_bins_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["lower_bin"].dropna().to_numpy()
+    K_scores_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["K_score"].to_numpy()
+
+    lower_bins_pred, K_scores_pred = mdl_optimal_histogram(dataset, K_max=K_max, epsilon=epsilon)
+
+    np.testing.assert_array_almost_equal(lower_bins_pred, lower_bins_true, decimal=5)
+    np.testing.assert_allclose(K_scores_pred, K_scores_true, rtol=1e-5)
+
+def test_gmm5_id_4(rootdir_path):
+    gmmX_ds = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm5_samples.parquet.brotli")
+    gmmX_mdl_lower_bins = pd.read_parquet(Path(rootdir_path) / Path("tests") / "gmm5_samples_MDL_lower_bins.parquet.brotli")
+    id = 4
+
+    assert all(np.unique(gmmX_ds["dataset_id"]) == np.unique(gmmX_mdl_lower_bins["dataset_id"])), "The number of dataset ids must equal"
+    dataset = gmmX_ds[gmmX_ds["dataset_id"] == id]["value"].to_numpy()
+    lower_bins_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["lower_bin"].dropna().to_numpy()
+    K_scores_true = gmmX_mdl_lower_bins[gmmX_mdl_lower_bins["dataset_id"] == id]["K_score"].to_numpy()
+
+    lower_bins_pred, K_scores_pred = mdl_optimal_histogram(dataset, K_max=K_max, epsilon=epsilon)
+
+    np.testing.assert_array_almost_equal(lower_bins_pred, lower_bins_true, decimal=5)
+    np.testing.assert_allclose(K_scores_pred, K_scores_true, rtol=1e-5)