# -*- encoding: utf-8 -*-
# pylint: disable=E0203,E1101,C0111
"""
Runtime operator.
:githublink:`%|py|7`
"""
from collections import OrderedDict
import numpy
from ._op_helper import _get_typed_class_attribute
from ._op import OpRunClassifierProb, RuntimeTypeError
from ._op_classifier_string import _ClassifierCommon
from ._new_ops import OperatorSchema
from .op_tree_ensemble_classifier_ import ( # pylint: disable=E0611,E0401
RuntimeTreeEnsembleClassifierDouble,
RuntimeTreeEnsembleClassifierFloat,
)
from .op_tree_ensemble_classifier_p_ import ( # pylint: disable=E0611,E0401
RuntimeTreeEnsembleClassifierPFloat,
RuntimeTreeEnsembleClassifierPDouble,
)
[docs]class TreeEnsembleClassifierCommon(OpRunClassifierProb, _ClassifierCommon):
[docs] def __init__(self, dtype, onnx_node, desc=None,
expected_attributes=None,
runtime_version=3, **options):
OpRunClassifierProb.__init__(
self, onnx_node, desc=desc,
expected_attributes=expected_attributes, **options)
self._init(dtype=dtype, version=runtime_version)
[docs] def _get_typed_attributes(self, k):
return _get_typed_class_attribute(self, k, self.__class__.atts)
[docs] def _find_custom_operator_schema(self, op_name):
"""
Finds a custom operator defined by this runtime.
:githublink:`%|py|39`
"""
if op_name == "TreeEnsembleClassifierDouble":
return TreeEnsembleClassifierDoubleSchema()
raise RuntimeError( # pragma: no cover
"Unable to find a schema for operator '{}'.".format(op_name))
[docs] def _init(self, dtype, version):
self._post_process_label_attributes()
if dtype == numpy.float32:
if version == 0:
self.rt_ = RuntimeTreeEnsembleClassifierFloat()
elif version == 1:
self.rt_ = RuntimeTreeEnsembleClassifierPFloat(
60, 20, False, False)
elif version == 2:
self.rt_ = RuntimeTreeEnsembleClassifierPFloat(
60, 20, True, False)
elif version == 3:
self.rt_ = RuntimeTreeEnsembleClassifierPFloat(
60, 20, True, True)
else:
raise ValueError("Unknown version '{}'.".format(version))
elif dtype == numpy.float64:
if version == 0:
self.rt_ = RuntimeTreeEnsembleClassifierDouble()
elif version == 1:
self.rt_ = RuntimeTreeEnsembleClassifierPDouble(
60, 20, False, False)
elif version == 2:
self.rt_ = RuntimeTreeEnsembleClassifierPDouble(
60, 20, True, False)
elif version == 3:
self.rt_ = RuntimeTreeEnsembleClassifierPDouble(
60, 20, True, True)
else:
raise ValueError( # pragma: no cover
"Unknown version '{}'.".format(version))
else:
raise RuntimeTypeError("Unsupported dtype={}.".format(dtype))
atts = [self._get_typed_attributes(k)
for k in self.__class__.atts]
self.rt_.init(*atts)
[docs] def _run(self, x): # pylint: disable=W0221
"""
This is a C++ implementation coming from
:epkg:`onnxruntime`.
`tree_ensemble_classifier.cc
<https://github.com/microsoft/onnxruntime/blob/master/onnxruntime/core/providers/cpu/ml/tree_ensemble_classifier.cc>`_.
See class :class:`RuntimeTreeEnsembleClassifier
<mlprodict.onnxrt.ops_cpu.op_tree_ensemble_classifier_.RuntimeTreeEnsembleClassifier>`.
:githublink:`%|py|90`
"""
label, scores = self.rt_.compute(x)
if scores.shape[0] != label.shape[0]:
scores = scores.reshape(label.shape[0],
scores.shape[0] // label.shape[0])
return self._post_process_predicted_label(label, scores)
[docs]class TreeEnsembleClassifier(TreeEnsembleClassifierCommon):
atts = OrderedDict([
('base_values', numpy.empty(0, dtype=numpy.float32)),
('class_ids', numpy.empty(0, dtype=numpy.int64)),
('class_nodeids', numpy.empty(0, dtype=numpy.int64)),
('class_treeids', numpy.empty(0, dtype=numpy.int64)),
('class_weights', numpy.empty(0, dtype=numpy.float32)),
('classlabels_int64s', numpy.empty(0, dtype=numpy.int64)),
('classlabels_strings', []),
('nodes_falsenodeids', numpy.empty(0, dtype=numpy.int64)),
('nodes_featureids', numpy.empty(0, dtype=numpy.int64)),
('nodes_hitrates', numpy.empty(0, dtype=numpy.float32)),
('nodes_missing_value_tracks_true', numpy.empty(0, dtype=numpy.int64)),
('nodes_modes', []),
('nodes_nodeids', numpy.empty(0, dtype=numpy.int64)),
('nodes_treeids', numpy.empty(0, dtype=numpy.int64)),
('nodes_truenodeids', numpy.empty(0, dtype=numpy.int64)),
('nodes_values', numpy.empty(0, dtype=numpy.float32)),
('post_transform', b'NONE')
])
[docs] def __init__(self, onnx_node, desc=None, **options):
TreeEnsembleClassifierCommon.__init__(
self, numpy.float32, onnx_node, desc=desc,
expected_attributes=TreeEnsembleClassifier.atts, **options)
[docs]class TreeEnsembleClassifierDouble(TreeEnsembleClassifierCommon):
atts = OrderedDict([
('base_values', numpy.empty(0, dtype=numpy.float64)),
('class_ids', numpy.empty(0, dtype=numpy.int64)),
('class_nodeids', numpy.empty(0, dtype=numpy.int64)),
('class_treeids', numpy.empty(0, dtype=numpy.int64)),
('class_weights', numpy.empty(0, dtype=numpy.float64)),
('classlabels_int64s', numpy.empty(0, dtype=numpy.int64)),
('classlabels_strings', []),
('nodes_falsenodeids', numpy.empty(0, dtype=numpy.int64)),
('nodes_featureids', numpy.empty(0, dtype=numpy.int64)),
('nodes_hitrates', numpy.empty(0, dtype=numpy.float64)),
('nodes_missing_value_tracks_true', numpy.empty(0, dtype=numpy.int64)),
('nodes_modes', []),
('nodes_nodeids', numpy.empty(0, dtype=numpy.int64)),
('nodes_treeids', numpy.empty(0, dtype=numpy.int64)),
('nodes_truenodeids', numpy.empty(0, dtype=numpy.int64)),
('nodes_values', numpy.empty(0, dtype=numpy.float64)),
('post_transform', b'NONE')
])
[docs] def __init__(self, onnx_node, desc=None, **options):
TreeEnsembleClassifierCommon.__init__(
self, numpy.float64, onnx_node, desc=desc,
expected_attributes=TreeEnsembleClassifier.atts, **options)
[docs]class TreeEnsembleClassifierDoubleSchema(OperatorSchema):
"""
Defines a schema for operators added in this package
such as :class:`TreeEnsembleClassifierDouble <mlprodict.onnxrt.ops_cpu.op_tree_ensemble_classifier.TreeEnsembleClassifierDouble>`.
:githublink:`%|py|158`
"""
[docs] def __init__(self):
OperatorSchema.__init__(self, 'TreeEnsembleClassifierDouble')
self.attributes = TreeEnsembleClassifierDouble.atts
