fix many tiny bugs

Signed-off-by: Xavier Dupre <[email protected]>

.azure-pipelines/linux-conda-CI.yml

@@ -54,7 +54,7 @@ jobs:
         pandas.version: ''
         lgbm.version: ''
         onnxcc.version: '>=1.8.1'
-        run.example: '1'
+        run.example: '0'
 
       Py311-Onnx140-Rt151-Skl130:
         do.bench: '0'

docs/conf.py

@@ -35,14 +35,12 @@
     "sphinx.ext.graphviz",
     "sphinx_skl2onnx_extension",
     "matplotlib.sphinxext.plot_directive",
-    "pyquickhelper.sphinxext.sphinx_cmdref_extension",
-    "pyquickhelper.sphinxext.sphinx_collapse_extension",
-    "pyquickhelper.sphinxext.sphinx_docassert_extension",
-    "pyquickhelper.sphinxext.sphinx_epkg_extension",
-    "pyquickhelper.sphinxext.sphinx_exref_extension",
-    "pyquickhelper.sphinxext.sphinx_faqref_extension",
-    "pyquickhelper.sphinxext.sphinx_gdot_extension",
-    "pyquickhelper.sphinxext.sphinx_runpython_extension",
+    "sphinx_runpython.blocdefs.sphinx_exref_extension",
+    "sphinx_runpython.blocdefs.sphinx_faqref_extension",
+    "sphinx_runpython.blocdefs.sphinx_mathdef_extension",
+    "sphinx_runpython.epkg",
+    "sphinx_runpython.gdot",
+    "sphinx_runpython.runpython",
     "sphinxcontrib.blockdiag",
 ]
 

docs/requirements.txt

@@ -18,14 +18,14 @@ pandas
 pydot
 pyinstrument
 pyod
-pyquickhelper>=1.11.3762
 pytest
 pytest-cov
 scikit-learn>=1.1
 skl2onnx
 sphinx
 sphinxcontrib-blockdiag
 sphinx-gallery
+sphinx-runpython
 tabulate
 tqdm
 wheel

docs/tests/test_documentation_examples.py

@@ -73,6 +73,8 @@ def add_test_methods(cls):
                 if not name.endswith(".py") or not name.startswith("plot_"):
                     continue
                 reason = None
+                if name in {"plot_woe_transformer.py"}:
+                    reason = "dot not available"
 
                 if reason:
 

docs/tutorial/plot_gbegin_dataframe.py

@@ -15,17 +15,14 @@
 +++++++++++++++++++++++++
 
 """
-from mlinsights.plotting import pipeline2dot
 import numpy
 import pprint
-from onnx.reference import ReferenceEvaluator
 from onnxruntime import InferenceSession
 from pandas import DataFrame
 from sklearn.pipeline import Pipeline
 from sklearn.compose import ColumnTransformer
 from sklearn.preprocessing import OneHotEncoder
 from sklearn.ensemble import RandomForestClassifier
-from pyquickhelper.helpgen.graphviz_helper import plot_graphviz
 from skl2onnx import to_onnx
 from skl2onnx.algebra.type_helper import guess_initial_types
 
@@ -54,14 +51,6 @@
 pipe = Pipeline([("preprocess", preprocessor), ("rf", RandomForestClassifier())])
 pipe.fit(train_data, data["y"])
 
-#####################################
-# Display.
-
-dot = pipeline2dot(pipe, train_data)
-ax = plot_graphviz(dot)
-ax.get_xaxis().set_visible(False)
-ax.get_yaxis().set_visible(False)
-
 #######################################
 # Conversion to ONNX
 # ++++++++++++++++++
@@ -84,24 +73,6 @@
 except Exception as e:
     print(e)
 
-###########################
-# Let's use a shortcut
-
-oinf = ReferenceEvaluator(onx)
-got = oinf.run(None, train_data)
-print(pipe.predict(train_data))
-print(got["label"])
-
-#################################
-# And probilities.
-
-print(pipe.predict_proba(train_data))
-print(got["probabilities"])
-
-######################################
-# It looks ok. Let's dig into the details to
-# directly use *onnxruntime*.
-#
 # Unhide conversion logic with a dataframe
 # ++++++++++++++++++++++++++++++++++++++++
 #

skl2onnx/algebra/type_helper.py

@@ -90,7 +90,7 @@ def _guess_dim(value):
     )
 
 
-def guess_initial_types(X, initial_types):
+def guess_initial_types(X, initial_types=None):
     if X is None and initial_types is None:
         raise NotImplementedError("Initial types must be specified.")
     if initial_types is None:

tests/test_onnx_helper.py

@@ -31,6 +31,12 @@ def one_hot_encoder_supports_string():
     return pv.Version(vers) >= pv.Version("0.20.0")
 
 
+def skl12():
+    # pv.Version does not work with development versions
+    vers = ".".join(sklearn_version.split(".")[:2])
+    return pv.Version(vers) >= pv.Version("1.2")
+
+
 class TestOnnxHelper(unittest.TestCase):
     def get_model(self, model):
         try:
@@ -73,6 +79,7 @@ def test_onnx_helper_load_save(self):
         not one_hot_encoder_supports_string(),
         reason="OneHotEncoder did not have categories_ before 0.20",
     )
+    @unittest.skipIf(not skl12(), reason="sparse_output")
     def test_onnx_helper_load_save_init(self):
         model = make_pipeline(
             Binarizer(),
@@ -105,6 +112,7 @@ def test_onnx_helper_load_save_init(self):
         not one_hot_encoder_supports_string(),
         reason="OneHotEncoder did not have categories_ before 0.20",
     )
+    @unittest.skipIf(not skl12(), reason="sparse_output")
     def test_onnx_helper_load_save_init_meta(self):
         model = make_pipeline(
             Binarizer(), OneHotEncoder(sparse_output=False), StandardScaler()

tests/test_sklearn_array_feature_extractor.py

@@ -6,6 +6,7 @@
 import pandas as pd
 import numpy as np
 from onnxruntime import __version__ as ort_version
+from sklearn import __version__ as sklearn_version
 from sklearn.mixture import GaussianMixture
 from sklearn.preprocessing import OneHotEncoder
 
@@ -20,11 +21,18 @@
 from test_utils import dump_data_and_model, TARGET_OPSET
 
 
+def skl12():
+    # pv.Version does not work with development versions
+    vers = ".".join(sklearn_version.split(".")[:2])
+    return pv.Version(vers) >= pv.Version("1.2")
+
+
 class TestSklearnArrayFeatureExtractor(unittest.TestCase):
     @unittest.skipIf(
         ColumnTransformer is None or pv.Version(ort_version) <= pv.Version("0.4.0"),
         reason="onnxruntime too old",
     )
+    @unittest.skipIf(not skl12(), reason="sparse_output")
     def test_array_feature_extractor(self):
         data_to_cluster = pd.DataFrame(
             [[1, 2, 3.5, 4.5], [1, 2, 1.7, 4.0], [2, 4, 2.4, 4.3], [2, 4, 2.5, 4.0]],

tests/test_sklearn_calibrated_classifier_cv_converter.py

@@ -9,6 +9,7 @@
 import numpy as np
 from numpy.testing import assert_almost_equal
 from onnxruntime import __version__ as ort_version
+from sklearn import __version__ as sklearn_version
 from sklearn.calibration import CalibratedClassifierCV
 from sklearn.datasets import load_digits, load_iris
 from sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifier
@@ -49,6 +50,12 @@
 ort_version = ort_version.split("+")[0]
 
 
+def skl12():
+    # pv.Version does not work with development versions
+    vers = ".".join(sklearn_version.split(".")[:2])
+    return pv.Version(vers) >= pv.Version("1.2")
+
+
 class TestSklearnCalibratedClassifierCVConverters(unittest.TestCase):
     @ignore_warnings(category=(FutureWarning, ConvergenceWarning, DeprecationWarning))
     def test_model_calibrated_classifier_cv_float(self):
@@ -186,6 +193,7 @@ def test_model_calibrated_classifier_cv_isotonic_binary_knn(self):
         pv.Version(ort_version) < pv.Version("0.5.0"), reason="not available"
     )
     @ignore_warnings(category=(FutureWarning, ConvergenceWarning, DeprecationWarning))
+    @unittest.skipIf(not skl12(), reason="base_estimator")
     def test_model_calibrated_classifier_cv_logistic_regression(self):
         data = load_iris()
         X, y = data.data, data.target
@@ -210,6 +218,7 @@ def test_model_calibrated_classifier_cv_logistic_regression(self):
         pv.Version(ort_version) < pv.Version("0.5.0"), reason="not available"
     )
     @ignore_warnings(category=(FutureWarning, ConvergenceWarning, DeprecationWarning))
+    @unittest.skipIf(not skl12(), reason="base_estimator")
     def test_model_calibrated_classifier_cv_rf(self):
         data = load_iris()
         X, y = data.data, data.target
@@ -234,6 +243,7 @@ def test_model_calibrated_classifier_cv_rf(self):
         pv.Version(ort_version) < pv.Version("0.5.0"), reason="not available"
     )
     @ignore_warnings(category=(FutureWarning, ConvergenceWarning, DeprecationWarning))
+    @unittest.skipIf(not skl12(), reason="base_estimator")
     def test_model_calibrated_classifier_cv_gbt(self):
         data = load_iris()
         X, y = data.data, data.target
@@ -259,6 +269,7 @@ def test_model_calibrated_classifier_cv_gbt(self):
         pv.Version(ort_version) < pv.Version("0.5.0"), reason="not available"
     )
     @ignore_warnings(category=(FutureWarning, ConvergenceWarning, DeprecationWarning))
+    @unittest.skipIf(not skl12(), reason="base_estimator")
     def test_model_calibrated_classifier_cv_hgbt(self):
         data = load_iris()
         X, y = data.data, data.target
@@ -308,6 +319,7 @@ def test_model_calibrated_classifier_cv_tree(self):
     )
     @unittest.skipIf(apply_less is None, reason="onnxconverter-common old")
     @ignore_warnings(category=(FutureWarning, ConvergenceWarning, DeprecationWarning))
+    @unittest.skipIf(not skl12(), reason="base_estimator")
     def test_model_calibrated_classifier_cv_svc(self):
         data = load_iris()
         X, y = data.data, data.target
@@ -330,6 +342,7 @@ def test_model_calibrated_classifier_cv_svc(self):
     )
     @unittest.skipIf(apply_less is None, reason="onnxconverter-common old")
     @ignore_warnings(category=(FutureWarning, ConvergenceWarning, DeprecationWarning))
+    @unittest.skipIf(not skl12(), reason="base_estimator")
     def test_model_calibrated_classifier_cv_linearsvc(self):
         data = load_iris()
         X, y = data.data, data.target
@@ -354,6 +367,7 @@ def test_model_calibrated_classifier_cv_linearsvc(self):
     )
     @unittest.skipIf(apply_less is None, reason="onnxconverter-common old")
     @ignore_warnings(category=(FutureWarning, ConvergenceWarning, DeprecationWarning))
+    @unittest.skipIf(not skl12(), reason="base_estimator")
     def test_model_calibrated_classifier_cv_linearsvc2(self):
         data = load_iris()
         X, y = data.data, data.target