sklearn-migrator 0.1.0__tar.gz

sklearn_migrator-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025 Alberto Valdés
+
+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.

sklearn_migrator-0.1.0/PKG-INFO

@@ -0,0 +1,46 @@
+Metadata-Version: 2.4
+Name: sklearn-migrator
+Version: 0.1.0
+Summary: A utility to migrate scikit-learn models between versions.
+Home-page: https://github.com/anvaldes/sklearn_migrator
+Author: Alberto Valdés
+Author-email: alberto.valdes.gonzalez.96.2@gmail.com
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: scikit-learn>=0.21.3
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: license-file
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
+
+# sklearn_migrator
+
+**sklearn_migrator** is a lightweight Python library that helps you serialize and migrate scikit-learn models across different versions.  
+It is especially useful when models are trained using older versions of scikit-learn and need to be loaded or interpreted in newer environments.
+
+---
+
+## 🔧 Features
+
+- ✅ Serialize `DecisionTreeRegressor` models safely
+- 🔄 Add compatibility adjustments for internal node structures
+- 🧪 Facilitate model version migration without retraining
+- 📦 Easy to install and use
+
+---
+
+## 📦 Installation
+
+```bash
+pip install git+https://github.com/tu_usuario/sklearn_migrator.git
+```

sklearn_migrator-0.1.0/README.md

@@ -0,0 +1,21 @@
+# sklearn_migrator
+
+**sklearn_migrator** is a lightweight Python library that helps you serialize and migrate scikit-learn models across different versions.  
+It is especially useful when models are trained using older versions of scikit-learn and need to be loaded or interpreted in newer environments.
+
+---
+
+## 🔧 Features
+
+- ✅ Serialize `DecisionTreeRegressor` models safely
+- 🔄 Add compatibility adjustments for internal node structures
+- 🧪 Facilitate model version migration without retraining
+- 📦 Easy to install and use
+
+---
+
+## 📦 Installation
+
+```bash
+pip install git+https://github.com/tu_usuario/sklearn_migrator.git
+```

sklearn_migrator-0.1.0/setup.cfg

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

sklearn_migrator-0.1.0/setup.py

@@ -0,0 +1,25 @@
+import setuptools
+
+with open("README.md", "r", encoding="utf-8") as fh:
+    long_description = fh.read()
+
+setuptools.setup(
+    name='sklearn-migrator',  # Este será el nombre visible en PyPI
+    version='0.1.0',
+    author="Alberto Valdés",
+    author_email="alberto.valdes.gonzalez.96.2@gmail.com",
+    description="A utility to migrate scikit-learn models between versions.",
+    long_description=long_description,
+    long_description_content_type="text/markdown",
+    url="https://github.com/anvaldes/sklearn_migrator",
+    packages=setuptools.find_packages(),
+    install_requires=[
+        'scikit-learn>=0.21.3',
+    ],
+    classifiers=[
+        "Programming Language :: Python :: 3",
+        "License :: OSI Approved :: MIT License",
+        "Operating System :: OS Independent",
+    ],
+    python_requires='>=3.8',
+)

sklearn_migrator-0.1.0/sklearn_migrator/regression/decision_tree_reg.py

@@ -0,0 +1,144 @@
+import numpy as np
+from sklearn.tree._tree import Tree
+from sklearn.tree import DecisionTreeRegressor
+
+import warnings
+warnings.filterwarnings("ignore")
+
+
+def _version_range_check(version, lower, upper):
+    return lower <= version <= upper
+
+
+def _get_extended_nodes(nodes, version_in):
+    if '0.21.3' <= version_in < '1.3.0':
+        return [node + (0,) for node in nodes]
+    return nodes
+
+
+def _build_dtype_dict(dtypes, version_in):
+    field_names = dtypes.names
+    formats = [dtypes.fields[name][0] for name in field_names]
+    offsets = [dtypes.fields[name][1] for name in field_names]
+    itemsize = dtypes.itemsize
+
+    if '0.21.3' <= version_in < '1.3.0':
+        return {
+            'field_names': list(field_names + ('missing_go_to_left',)),
+            'formats': [str(fmt) for fmt in formats + [np.dtype('uint8')]],
+            'offsets': [int(off) for off in offsets + [56]],
+            'itemsize': 64
+        }
+
+    return {
+        'field_names': list(field_names),
+        'formats': [str(fmt) for fmt in formats],
+        'offsets': [int(off) for off in offsets],
+        'itemsize': int(itemsize)
+    }
+
+
+def _get_metadata(model, version_in):
+    if _version_range_check(version_in, '0.21.3', '0.22.1'):
+        return {
+            'n_features_in': None,
+            'n_features': model.n_features_,
+            'n_outputs': model.n_outputs_,
+            'n_classes': model.n_classes_
+        }
+    elif _version_range_check(version_in, '0.23.0', '0.23.2'):
+        return {
+            'n_features_in': model.n_features_in_,
+            'n_features': model.n_features_,
+            'n_outputs': model.n_outputs_,
+            'n_classes': model.n_classes_
+        }
+    elif _version_range_check(version_in, '0.24.0', '1.1.3'):
+        return {
+            'n_features_in': model.n_features_in_,
+            'n_features': model.n_features_,
+            'n_outputs': model.n_outputs_,
+            'n_classes': None
+        }
+    elif version_in >= '1.2.0':
+        return {
+            'n_features_in': model.n_features_in_,
+            'n_features': None,
+            'n_outputs': model.n_outputs_,
+            'n_classes': None
+        }
+    return {}
+
+
+def serialize_decision_tree_reg(model, version_in):
+    tree = model.tree_
+    state = tree.__getstate__()
+
+    serialized_tree = {
+        'max_depth': int(state['max_depth']),
+        'node_count': int(state['node_count']),
+        'values': state['values'].tolist(),
+        'nodes': [list(n) for n in _get_extended_nodes(state['nodes'].tolist(), version_in)],
+        'dtypes': _build_dtype_dict(state['nodes'].dtype, version_in)
+    }
+
+    metadata = _get_metadata(model, version_in)
+    metadata['serialized_tree'] = serialized_tree
+    metadata['version_sklearn_in'] = version_in
+
+    return metadata
+
+
+def _build_tree_dtype(dtypes_dict, version_out):
+    version_lt_1_3 = version_out < '1.3.0'
+    num_elements = 7 if version_lt_1_3 else 8
+
+    field_names = dtypes_dict['field_names'][:num_elements]
+    formats = [np.dtype(fmt) for fmt in dtypes_dict['formats'][:num_elements]]
+    offsets = dtypes_dict['offsets'][:num_elements]
+    itemsize = 56 if version_lt_1_3 else 64
+
+    return np.dtype({
+        'names': field_names,
+        'formats': formats,
+        'offsets': offsets,
+        'itemsize': itemsize
+    }), num_elements
+
+
+def deserialize_decision_tree_reg(data, version_out):
+    version_in = data['version_sklearn_in']
+    serialized = data['serialized_tree']
+    dtype_dict = serialized['dtypes']
+
+    tree_dtype, num_elements = _build_tree_dtype(dtype_dict, version_out)
+
+    serialized['nodes'] = [tuple(n[:num_elements]) for n in serialized['nodes']]
+    nodes_array = np.array(serialized['nodes'], dtype=tree_dtype)
+    values_array = np.array(serialized['values'])
+
+    n_classes = np.array([1], dtype=np.intp)  # regression
+    n_outputs = data['n_outputs']
+    n_features = (data['n_features'] or data['n_features_in'])
+
+    tree_obj = Tree(n_features, n_classes, n_outputs)
+    tree_obj.__setstate__({
+        'max_depth': serialized['max_depth'],
+        'node_count': serialized['node_count'],
+        'nodes': nodes_array,
+        'values': values_array
+    })
+
+    new_tree = DecisionTreeRegressor(max_depth=serialized['max_depth'], random_state=42)
+    new_tree.tree_ = tree_obj
+    new_tree.n_outputs_ = n_outputs
+
+    if _version_range_check(version_out, '0.21.3', '0.22.1'):
+        new_tree.n_features_ = n_features
+    elif _version_range_check(version_out, '0.23.0', '0.24.2'):
+        new_tree.n_features_ = n_features
+        new_tree.n_features_in_ = n_features
+    elif version_out >= '1.0.0':
+        new_tree.n_features_in_ = n_features
+
+    return new_tree

sklearn_migrator-0.1.0/sklearn_migrator.egg-info/PKG-INFO

@@ -0,0 +1,46 @@
+Metadata-Version: 2.4
+Name: sklearn-migrator
+Version: 0.1.0
+Summary: A utility to migrate scikit-learn models between versions.
+Home-page: https://github.com/anvaldes/sklearn_migrator
+Author: Alberto Valdés
+Author-email: alberto.valdes.gonzalez.96.2@gmail.com
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: scikit-learn>=0.21.3
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: license-file
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
+
+# sklearn_migrator
+
+**sklearn_migrator** is a lightweight Python library that helps you serialize and migrate scikit-learn models across different versions.  
+It is especially useful when models are trained using older versions of scikit-learn and need to be loaded or interpreted in newer environments.
+
+---
+
+## 🔧 Features
+
+- ✅ Serialize `DecisionTreeRegressor` models safely
+- 🔄 Add compatibility adjustments for internal node structures
+- 🧪 Facilitate model version migration without retraining
+- 📦 Easy to install and use
+
+---
+
+## 📦 Installation
+
+```bash
+pip install git+https://github.com/tu_usuario/sklearn_migrator.git
+```

sklearn_migrator-0.1.0/sklearn_migrator.egg-info/SOURCES.txt

@@ -0,0 +1,12 @@
+LICENSE
+README.md
+setup.py
+sklearn_migrator/__init__.py
+sklearn_migrator.egg-info/PKG-INFO
+sklearn_migrator.egg-info/SOURCES.txt
+sklearn_migrator.egg-info/dependency_links.txt
+sklearn_migrator.egg-info/requires.txt
+sklearn_migrator.egg-info/top_level.txt
+sklearn_migrator/regression/__init__.py
+sklearn_migrator/regression/decision_tree_reg.py
+tests/test_decision_tree_reg.py

sklearn_migrator-0.1.0/sklearn_migrator.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

sklearn_migrator-0.1.0/sklearn_migrator.egg-info/requires.txt

@@ -0,0 +1 @@
+scikit-learn>=0.21.3

sklearn_migrator-0.1.0/sklearn_migrator.egg-info/top_level.txt

@@ -0,0 +1 @@
+sklearn_migrator

sklearn_migrator-0.1.0/tests/test_decision_tree_reg.py

@@ -0,0 +1,29 @@
+from sklearn.tree import DecisionTreeRegressor
+from sklearn_migrator.regression.decision_tree_reg import serialize_decision_tree_reg
+from sklearn_migrator.regression.decision_tree_reg import deserialize_decision_tree_reg
+import sklearn
+
+def test_decision_tree_reg():
+    X = [[0], [1], [2], [3]]
+    y = [0, 1, 2, 3]
+    
+    model = DecisionTreeRegressor()
+    model.fit(X, y)
+
+    version = sklearn.__version__
+    result = serialize_decision_tree_reg(model, version_in=version)
+    new_model = deserialize_decision_tree_reg(result, version_out=version)
+
+    assert isinstance(result, dict)
+
+    assert 'n_features_in' in result
+    assert 'n_features' in result
+    assert 'n_outputs' in result
+    assert 'n_classes' in result
+    assert 'serialized_tree' in result
+    assert 'version_sklearn_in' in result
+
+    assert result['serialized_tree']['max_depth'] > 0
+    assert isinstance(result['serialized_tree']['nodes'], list)
+
+    assert isinstance(new_model, DecisionTreeRegressor)