Merge pull request #399 from KhiopsML/dev-v10

Release 10.3.1.0

Author: popescu-v

.github/workflows/api-docs.yml

@@ -1,5 +1,7 @@
 ---
 name: API Docs
+env:
+  DEFAULT_KHIOPS_PYTHON_TUTORIAL_REVISION: 10.3.1.0
 on:
   workflow_dispatch:
     inputs:
@@ -8,6 +10,9 @@ on:
         required: true
         type: boolean
         default: false
+      khiops-python-tutorial-revision:
+        default: 10.3.1.0
+        description: khiops-python-tutorial repo revision
       image-tag:
         default: latest
         description: Development Docker Image Tag
@@ -44,6 +49,10 @@ jobs:
       # https://github.com/actions/runner/issues/2033#issuecomment-1598547465
       options: --user 1001
     steps:
+      - name: Set parameters as env
+        run: |
+          KHIOPS_PYTHON_TUTORIAL_REVISION=${{ inputs.khiops-python-tutorial-revision || env.DEFAULT_KHIOPS_PYTHON_TUTORIAL_REVISION }}
+          echo "KHIOPS_PYTHON_TUTORIAL_REVISION=$KHIOPS_PYTHON_TUTORIAL_REVISION" >> "$GITHUB_ENV"
       - name: Checkout khiops-python
         uses: actions/checkout@v4
         with:
@@ -52,6 +61,7 @@ jobs:
         uses: actions/checkout@v4
         with:
           repository: khiopsml/khiops-python-tutorial
+          ref: ${{ env.KHIOPS_PYTHON_TUTORIAL_REVISION }}
           path: doc/khiops-python-tutorial
       - name: Add pip scripts directory to path
         run: echo PATH="$PATH:/github/home/.local/bin" >> "$GITHUB_ENV"

.github/workflows/conda.yml

@@ -3,13 +3,13 @@ name: Conda Package
 env:
   # Note: The default Khiops version must never be an alpha release as they are
   #       ephemeral. To test alpha versions run the workflow manually.
-  DEFAULT_KHIOPS_CORE_VERSION: 10.3.0
+  DEFAULT_KHIOPS_CORE_VERSION: 10.3.1
   DEFAULT_SAMPLES_VERSION: 10.2.4
 on:
   workflow_dispatch:
     inputs:
       khiops-core-version:
-        default: 10.3.0
+        default: 10.3.1
         description: khiops-core version for testing
       khiops-samples-version:
         default: 10.2.4
@@ -75,8 +75,9 @@ jobs:
       matrix:
         python-version: ['3.8', '3.9', '3.10', '3.11', '3.12', '3.13']
         env:
-          - {os: ubuntu-20.04, json-image: '{"image": null}'}
+          - {os: ubuntu-22.04, json-image: '{"image": "ubuntu:20.04"}'}
           - {os: ubuntu-22.04, json-image: '{"image": null}'}
+          - {os: ubuntu-24.04, json-image: '{"image": null}'}
           - {os: ubuntu-22.04, json-image: '{"image": "rockylinux:8"}'}
           - {os: ubuntu-22.04, json-image: '{"image": "rockylinux:9"}'}
           - {os: windows-2019, json-image: '{"image": null}'}
@@ -117,7 +118,7 @@ jobs:
           PATH=$(echo $PATH | sed "s#$CONDA_PREFIX/bin:##g")
 
           # Unset *CONDA* environment variables
-          # As a corrolary, CONDA_PREFIX is unset
+          # As a corollary, CONDA_PREFIX is unset
           # Note: There is no way to remove these variables from GITHUB_ENV
           # (see https://github.com/actions/runner/issues/1126)
           for CONDA_VAR in $(env | grep CONDA)
@@ -131,17 +132,17 @@ jobs:
         run: |
           kh-download-datasets \
             --version ${{ inputs.khiops-samples-version || env.DEFAULT_SAMPLES_VERSION }}
-      - name: Test Conda Package on Samples
+      - name: Run Samples
         env:
           # Force > 2 CPU cores to launch mpiexec
           KHIOPS_PROC_NUMBER: 4
         run: |
-          kh-samples core -i train_predictor -e
+          kh-samples core -i deploy_model -e
+          kh-samples core -i deploy_coclustering -e
           kh-samples core -i train_predictor_error_handling -e
-          kh-samples core -i train_coclustering -e
           kh-samples sklearn -i khiops_classifier -e
           kh-samples sklearn -i khiops_coclustering -e
-      - name: Test Conda Package on Some Samples (Conda-Based Environments)
+      - name: Run Samples (Conda-Based Environments)
         env:
           # Force > 2 CPU cores to launch mpiexec
           KHIOPS_PROC_NUMBER: 4
@@ -153,18 +154,57 @@ jobs:
           PATH=$(echo $PATH | sed "s#$CONDA_PREFIX/bin:##g")
 
           # Unset *CONDA* environment variables
-          # As a corrolary, CONDA_PREFIX is unset
+          # As a corollary, CONDA_PREFIX is unset
           # Note: There is no way to remove these variables from GITHUB_ENV
           # (see https://github.com/actions/runner/issues/1126)
           for CONDA_VAR in $(env | grep CONDA)
           do
             unset $(echo $CONDA_VAR | cut -d '=' -f 1)
           done
 
-          # Execute some Khiops and Khiops Coclustering samples
+          # Run samples
           # Note: kh-samples is not reachable as it is not in PATH
           $PYTHON -m khiops.samples.samples -i deploy_model -e
           $PYTHON -m khiops.samples.samples -i deploy_coclustering -e
+          $PYTHON -m khiops.samples.samples -i train_predictor_error_handling -e
+          $PYTHON -m khiops.samples.samples_sklearn -i khiops_classifier -e
+          $PYTHON -m khiops.samples.samples_sklearn -i khiops_coclustering -e
+      # Checkout the sources to obtain the tests
+      # Note: The `sparse-checkout` option of this action is bugged so we checkout all the sources
+      #       See https://github.com/actions/checkout/issues/1628
+      - name: Checkout Sources
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 1
+      # This is needed so python does not retrieve the khiops module from PWD
+      - name: Remove the khiops Module from the Sources
+        run: rm -rf khiops
+      - name: Install the Test Requirements
+        run: conda install -y --file test-requirements.txt
+      - name: Run the Runner Initialization Integration Test
+        run: |
+          python -m unittest -v \
+            tests.test_khiops_integrations.KhiopsRunnerEnvironmentTests.test_runner_environment_initialization
+      - name: Run the Runner Initialization Integration Test (Conda-Based Environments)
+        run: |
+          # Set `python` to the current Conda Python executable
+          PYTHON="$(type -P python)"
+
+          # Remove $CONDA_PREFIX/bin from PATH
+          PATH=$(echo $PATH | sed "s#$CONDA_PREFIX/bin:##g")
+
+          # Unset *CONDA* environment variables
+          # As a corolary, CONDA_PREFIX is unset
+          # Note: There is no way to remove these variables from GITHUB_ENV
+          # (see https://github.com/actions/runner/issues/1126)
+          for CONDA_VAR in $(env | grep CONDA)
+          do
+            unset $(echo $CONDA_VAR | cut -d '=' -f 1)
+          done
+
+          # Execute the runner initialization integration test
+          $PYTHON -m unittest -v \
+            tests.test_khiops_integrations.KhiopsRunnerEnvironmentTests.test_runner_environment_initialization
 
   # Release is only executed on tags
   # Note: For this job to work the secrets variables KHIOPS_ANACONDA_CHANNEL_TOKEN and

.github/workflows/dev-docker.yml

@@ -1,7 +1,7 @@
 ---
 name: Dev Docker
 env:
-  DEFAULT_KHIOPS_REVISION: 10.3.0
+  DEFAULT_KHIOPS_REVISION: 10.3.1
   DEFAULT_IMAGE_INCREMENT: 0
   DEFAULT_SERVER_REVISION: main
   DEFAULT_PYTHON_VERSIONS: 3.8 3.9 3.10 3.11 3.12 3.13
@@ -14,7 +14,7 @@ on:
     inputs:
       khiops-revision:
         type: string
-        default: 10.3.0
+        default: 10.3.1
         description: Khiops Revision
       image-increment:
         type: number

.github/workflows/tests.yml

@@ -2,7 +2,7 @@
 name: Tests
 env:
   DEFAULT_SAMPLES_REVISION: 10.2.4
-  DEFAULT_KHIOPS_DESKTOP_REVISION: 10.3.0
+  DEFAULT_KHIOPS_DESKTOP_REVISION: 10.3.1
 on:
   workflow_dispatch:
     inputs:
@@ -13,7 +13,7 @@ on:
         default: latest
         description: Development Docker Image Tag
       khiops-desktop-revision:
-        default: 10.3.0
+        default: 10.3.1
         description: Khiops Windows Desktop Application Version
       run-expensive-tests:
         type: boolean
@@ -109,8 +109,9 @@ jobs:
             $CONDA install -y -n "$CONDA_ENV" `grep -v "^\[" khiops.egg-info/requires.txt`
             rm -rf khiops.egg-info
           done
-      - name: Prepare Tests Environment
-        if: github.ref != 'dev' && github.rev != 'dev-v10' && github.ref != 'main' && ! inputs.run-expensive-tests
+      - name: Configure Expensive Tests Setting
+        # Skip expensive tests by default, unless on the `dev`, `dev-v10` or `main` branches
+        if: github.ref != 'dev' && github.ref != 'dev-v10' && github.ref != 'main' && ! inputs.run-expensive-tests
         run: echo "SKIP_EXPENSIVE_TESTS=true" >> "$GITHUB_ENV"
       - name: Prepare Integration Tests on remote files
         env:

.gitignore

@@ -48,6 +48,7 @@ tests/resources/scenario_generation/api/*/output
 tests/resources/scenario_generation/general_options/output
 tests/resources/scenario_generation/runner_version/output
 tests/resources/*/copy_output_kdic
+tests/resources/scenario_generation/system_settings/output
 
 # File generated by documentation scripts/notebooks
 kh_samples

CHANGELOG.md

@@ -6,6 +6,17 @@
   - Example: 10.2.1.4 is the 5th version that supports khiops 10.2.1.
 - Internals: Changes in *Internals* sections are unlikely to be of interest for data scientists.
 
+## 10.3.1.0 - 2025-04-16
+
+### Added
+- (`sklearn`) Support for boolean and float targets in `KhiopsClassifier`.
+
+### Fixed
+- (`sklearn`) Crash when there were no informative trees in predictors.
+
+### Deprecated
+- (`core`) The `build_multi_table_dictionary_domain` helper function.
+
 ## 10.3.0.0 - 2025-02-10
 
 ### Fixed

khiops/core/api.py

@@ -21,7 +21,7 @@
 import khiops.core.internals.filesystems as fs
 from khiops.core.dictionary import DictionaryDomain, read_dictionary_file
 from khiops.core.exceptions import KhiopsRuntimeError
-from khiops.core.helpers import build_multi_table_dictionary_domain
+from khiops.core.helpers import _build_multi_table_dictionary_domain
 from khiops.core.internals.common import (
     CommandLineOptions,
     SystemSettings,
@@ -1943,8 +1943,7 @@ def build_multi_table_dictionary(
 
     .. warning::
         This method is *deprecated* since Khiops 10.1.3 and will be removed in Khiops
-        11. Use the `.build_multi_table_dictionary_domain` helper function to
-        the same effect.
+        11.
 
     Parameters
     ----------
@@ -1979,7 +1978,7 @@ def build_multi_table_dictionary(
     # Generate multi-table domain by using the eponymous helper function
     # Honor exception API:
     try:
-        multi_table_domain = build_multi_table_dictionary_domain(
+        multi_table_domain = _build_multi_table_dictionary_domain(
             dictionary_domain, root_dictionary_name, secondary_table_variable_name
         )
     except TypeError as error:

khiops/core/dictionary.py

@@ -186,7 +186,7 @@ def add_dictionary(self, dictionary):
 
         Parameters
         ----------
-        dictionary : `DictionaryDomain`
+        dictionary : `Dictionary`
             The dictionary to be added.
 
         Raises

khiops/core/helpers.py

@@ -8,6 +8,7 @@
 import os
 import platform
 import subprocess
+import warnings
 
 import khiops.core.internals.filesystems as fs
 from khiops.core import api
@@ -19,6 +20,7 @@
 )
 from khiops.core.internals.common import (
     create_unambiguous_khiops_path,
+    deprecation_message,
     is_list_like,
     type_error_message,
 )
@@ -29,6 +31,22 @@ def build_multi_table_dictionary_domain(
 ):
     """Builds a multi-table dictionary domain from a dictionary with a key
 
+    .. note::
+
+        This is a special-purpose function whose goal is to assist in preparing the
+        coclustering deployment.
+
+        This function builds a new root dictionary and adds it to an existing dictionary
+        domain.
+        The new root dictionary only contains one field, which references a preexisting
+        dictionary from the input dictionary domain as a new (secondary) Table variable.
+        The preexisting dictionary must have a key set on it, as this is the join key
+        with the new root table.
+
+    .. warning::
+        This method is *deprecated* since Khiops 10.3.1.0 and will be removed in Khiops
+        11.
+
     Parameters
     ----------
     dictionary_domain : `.DictionaryDomain`
@@ -38,6 +56,11 @@ def build_multi_table_dictionary_domain(
     secondary_table_variable_name : str
         Name, in the root dictionary, for the "table" variable of the secondary table.
 
+    Returns
+    -------
+    `.DictionaryDomain`
+        The new dictionary domain
+
     Raises
     ------
     `TypeError`
@@ -47,6 +70,16 @@ def build_multi_table_dictionary_domain(
         - the dictionary domain doesn't contain at least a dictionary
         - the dictionary domain's root dictionary doesn't have a key set
     """
+    # Warn the user that this helper function is deprecated and will be removed
+    warnings.warn(deprecation_message("build_multi_table_dictionary_domain", "11.0.0"))
+    return _build_multi_table_dictionary_domain(
+        dictionary_domain, root_dictionary_name, secondary_table_variable_name
+    )
+
+
+def _build_multi_table_dictionary_domain(
+    dictionary_domain, root_dictionary_name, secondary_table_variable_name
+):
     # Check that `dictionary_domain` is a `DictionaryDomain`
     if not isinstance(dictionary_domain, DictionaryDomain):
         raise TypeError(
@@ -266,7 +299,7 @@ def deploy_coclustering(
     # Create a root dictionary containing the keys
     root_dictionary_name = "CC_" + dictionary_name
     table_variable_name = "Table_" + dictionary_name
-    domain = build_multi_table_dictionary_domain(
+    domain = _build_multi_table_dictionary_domain(
         tmp_domain, root_dictionary_name, table_variable_name
     )
 

khiops/sklearn/dataset.py

@@ -14,6 +14,7 @@
 
 import numpy as np
 import pandas as pd
+import sklearn
 from scipy import sparse as sp
 from sklearn.utils import check_array
 from sklearn.utils.validation import column_or_1d
@@ -430,6 +431,19 @@ def write_internal_data_table(dataframe, file_path_or_stream):
     )
 
 
+def _column_or_1d_with_dtype(y, dtype=None):
+    # 'dtype' has been introduced on `column_or_1d' since Scikit-learn 1.2;
+    if sklearn.__version__ < "1.2":
+        if pd.api.types.is_string_dtype(dtype) and y.isin(["True", "False"]).all():
+            warnings.warn(
+                "'y' stores strings restricted to 'True'/'False' values: "
+                "The predict method may return a bool vector."
+            )
+        return column_or_1d(y, warn=True)
+    else:
+        return column_or_1d(y, warn=True, dtype=dtype)
+
+
 class Dataset:
     """A representation of a dataset
 
@@ -738,8 +752,22 @@ def _init_target_column(self, y):
         if isinstance(y, str):
             y_checked = y
         else:
-            y_checked = column_or_1d(y, warn=True)
-
+            if hasattr(y, "dtype"):
+                if isinstance(y.dtype, pd.CategoricalDtype):
+                    y_checked = _column_or_1d_with_dtype(
+                        y, dtype=y.dtype.categories.dtype
+                    )
+                else:
+                    y_checked = _column_or_1d_with_dtype(y, dtype=y.dtype)
+            elif hasattr(y, "dtypes"):
+                if isinstance(y.dtypes[0], pd.CategoricalDtype):
+                    y_checked = _column_or_1d_with_dtype(
+                        y, dtype=y.dtypes[0].categories.dtype
+                    )
+                else:
+                    y_checked = _column_or_1d_with_dtype(y)
+            else:
+                y_checked = _column_or_1d_with_dtype(y)
         # Check the target type coherence with those of X's tables
         if isinstance(
             self.main_table, (PandasTable, SparseTable, NumpyTable)