{ "cells": [ { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "# Copyright 2019 IBM Corporation\n", "#\n", "# Licensed under the Apache License, Version 2.0 (the \"License\");\n", "# you may not use this file except in compliance with the License.\n", "# You may obtain a copy of the License at\n", "#\n", "# http://www.apache.org/licenses/LICENSE-2.0\n", "#\n", "# Unless required by applicable law or agreed to in writing, software\n", "# distributed under the License is distributed on an \"AS IS\" BASIS,\n", "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n", "# See the License for the specific language governing permissions and\n", "# limitations under the License.\n", "\n", "import sklearn.svm\n", "import lale.docstrings\n", "import lale.operators\n", "\n", "class SVCImpl():\n", " def __init__(self, **hyperparams):\n", " self._hyperparams = hyperparams\n", " self._wrapped_model = sklearn.svm.SVC(**self._hyperparams)\n", "\n", " def fit(self, X, y=None, sample_weight=None):\n", " self._wrapped_model.fit(X, y, sample_weight)\n", " return self\n", "\n", " def predict(self, X):\n", " return self._wrapped_model.predict(X)\n", "\n", " def predict_proba(self, X):\n", " return self._wrapped_model.predict_proba(X)\n", "\n", " def decision_function(self, X):\n", " return self._wrapped_model.decision_function(X)\n", "\n", "_hyperparams_schema = {\n", " 'allOf': [\n", " { 'type': 'object',\n", " 'additionalProperties': False,\n", " 'required': ['kernel', 'degree', 'gamma', 'shrinking', 'tol', 'cache_size', 'max_iter', 'decision_function_shape'],\n", " 'relevantToOptimizer': ['kernel', 'degree', 'gamma', 'shrinking', 'probability', 'tol'],\n", " 'properties': {\n", " 'C': {\n", " 'description': 'Penalty parameter C of the error term.',\n", " 'type': 'number',\n", " 'distribution': 'loguniform',\n", " 'minimum': 0.0,\n", " 'exclusiveMinimum': True,\n", " 'default': 1.0,\n", " 'minimumForOptimizer': 0.03125,\n", " 'maximumForOptimizer': 32768},\n", " 'kernel': {\n", " 'anyOf': [\n", " { 'enum':['precomputed'], 'forOptimizer': False}, \n", " { 'enum': ['linear', 'poly', 'rbf', 'sigmoid']},\n", " { 'laleType': 'callable', 'forOptimizer': False}],\n", " 'default': 'rbf',\n", " 'description':\n", " 'Specifies the kernel type to be used in the algorithm.'},\n", " 'degree': {\n", " 'type': 'integer',\n", " 'minimum': 0,\n", " 'minimumForOptimizer': 2,\n", " 'maximumForOptimizer': 5,\n", " 'default': 3,\n", " 'description': \"Degree of the polynomial kernel function ('poly').\"},\n", " 'gamma': {\n", " 'anyOf': [\n", " { 'type': 'number',\n", " 'minimum': 0.0,\n", " 'exclusiveMinimum': True,\n", " 'minimumForOptimizer': 3.0517578125e-05,\n", " 'maximumForOptimizer': 8,\n", " 'distribution': 'loguniform'},\n", " { 'enum': ['auto', 'auto_deprecated', 'scale']}],\n", " 'default': 'auto_deprecated', #going to change to 'scale' from sklearn 0.22.\n", " 'description': \"Kernel coefficient for 'rbf', 'poly', and 'sigmoid'.\"},\n", " 'coef0': {\n", " 'type': 'number',\n", " 'default': 0.0,\n", " 'description': 'Independent term in kernel function.'},\n", " 'shrinking': {\n", " 'type': 'boolean',\n", " 'default': True,\n", " 'description': 'Whether to use the shrinking heuristic.'},\n", " 'probability': {\n", " 'type': 'boolean',\n", " 'default': False,\n", " 'description': 'Whether to enable probability estimates.'},\n", " 'tol': {\n", " 'type': 'number',\n", " 'distribution': 'loguniform',\n", " 'minimum': 0.0,\n", " 'exclusiveMinimum': True,\n", " 'maximumForOptimizer': 0.01,\n", " 'default': 0.0001,\n", " 'description': 'Tolerance for stopping criteria.'},\n", " 'cache_size': {\n", " 'type': 'integer',\n", " 'default': 200,\n", " 'description': 'Specify the size of the kernel cache (in MB).'},\n", " 'class_weight': {\n", " 'anyOf': [\n", " { 'description': 'By default, all classes have weight 1.',\n", " 'enum': [None]},\n", " { 'description': 'Adjust weights by inverse frequency.',\n", " 'enum': ['balanced']},\n", " { 'description': 'Dictionary mapping class labels to weights.',\n", " 'type': 'object',\n", " 'additionalProperties': {'type': 'number'},\n", " 'forOptimizer': False}],\n", " 'default': None},\n", " 'verbose': {\n", " 'type': 'boolean',\n", " 'default': False,\n", " 'description': 'Enable verbose output.'},\n", " 'max_iter': {\n", " 'type': 'integer',\n", " 'default': -1,\n", " 'description': 'Hard limit on iterations within solver, or -1 for no limit.'},\n", " 'decision_function_shape': {\n", " 'enum': ['ovo', 'ovr'],\n", " 'default': 'ovr',\n", " 'description': \"Whether to return a one-vs-rest ('ovr') decision function of shape (n_samples, n_classes) as all other classifiers, or the original one-vs-one (‘ovo’) decision function of libsvm which has shape (n_samples, n_classes * (n_classes - 1) / 2).\"},\n", " 'random_state': {\n", " 'description':\n", " 'Seed of pseudo-random number generator.',\n", " 'anyOf': [\n", " { 'laleType': 'numpy.random.RandomState'},\n", " { 'description': 'RandomState used by np.random',\n", " 'enum': [None]},\n", " { 'description': 'Explicit seed.',\n", " 'type': 'integer'}],\n", " 'default': None}}},\n", " { 'description': 'coef0 only significant in kernel ‘poly’ and ‘sigmoid’.',\n", " 'anyOf': [\n", " { 'type': 'object',\n", " 'properties': {'kernel': {'enum': ['poly', 'sigmoid']}}},\n", " { 'type': 'object',\n", " 'properties': {'coef0': {'enum': [0.0]}}}]}]}\n", "\n", "_input_fit_schema = {\n", " 'type': 'object',\n", " 'required': ['X', 'y'],\n", " 'properties': {\n", " 'X': {\n", " 'type': 'array',\n", " 'description': 'The outer array is over samples aka rows.',\n", " 'items': {\n", " 'type': 'array',\n", " 'description': 'The inner array is over features aka columns.',\n", " 'items': {\n", " 'type': 'number'}}},\n", " 'y': {\n", " 'description': 'The predicted classes.',\n", " 'anyOf': [\n", " { 'type': 'array', 'items': {'type': 'number'}},\n", " { 'type': 'array', 'items': {'type': 'string'}},\n", " { 'type': 'array', 'items': {'type': 'boolean'}}]},\n", " 'sample_weight': {\n", " 'anyOf': [\n", " { 'type': 'array',\n", " 'items': {'type': 'number'}},\n", " { 'enum': [None],\n", " 'description': 'Samples are equally weighted.'}],\n", " 'description': 'Sample weights.'}}}\n", "\n", "_input_predict_schema = {\n", " 'type': 'object',\n", " 'properties': {\n", " 'X': {\n", " 'type': 'array',\n", " 'description': 'The outer array is over samples aka rows.',\n", " 'items': {\n", " 'type': 'array',\n", " 'description': 'The inner array is over features aka columns.',\n", " 'items': {\n", " 'type': 'number'}}}}}\n", "\n", "_output_predict_schema = {\n", " 'description': 'The predicted classes.',\n", " 'anyOf': [\n", " { 'type': 'array', 'items': {'type': 'number'}},\n", " { 'type': 'array', 'items': {'type': 'string'}},\n", " { 'type': 'array', 'items': {'type': 'boolean'}}]}\n", "\n", "_input_predict_proba_schema = {\n", " 'type': 'object',\n", " 'properties': {\n", " 'X': {\n", " 'type': 'array',\n", " 'description': 'The outer array is over samples aka rows.',\n", " 'items': {\n", " 'type': 'array',\n", " 'description': 'The inner array is over features aka columns.',\n", " 'items': {\n", " 'type': 'number'}}}}}\n", "\n", "_output_predict_proba_schema = {\n", " 'type': 'array',\n", " 'description': 'The outer array is over samples aka rows.',\n", " 'items': {\n", " 'type': 'array',\n", " 'description': 'The inner array has items corresponding to each class.',\n", " 'items': {\n", " 'type': 'number'}}}\n", "\n", "_input_decision_function_schema = {\n", " 'type': 'object',\n", " 'required': ['X'],\n", " 'properties': {\n", " 'X': {\n", " 'type': 'array',\n", " 'description': 'The outer array is over samples aka rows.',\n", " 'items': {\n", " 'type': 'array',\n", " 'description': 'The inner array is over features aka columns.',\n", " 'items': {\n", " 'type': 'number'}}}}}\n", "\n", "_output_decision_function_schema = {\n", " 'description': 'Confidence scores for samples for each class in the model.',\n", " 'anyOf': [\n", " { 'description': 'In the multi-way case, score per (sample, class) combination.',\n", " 'type': 'array',\n", " 'items': {'type': 'array', 'items': {'type': 'number'}}},\n", " { 'description': 'In the binary case, score for `self._classes[1]`.',\n", " 'type': 'array',\n", " 'items': {'type': 'number'}}]}\n", "\n", "_combined_schemas = {\n", " '$schema': 'http://json-schema.org/draft-04/schema#',\n", " 'description': \"\"\"`Support Vector Classification`_ from scikit-learn.\n", "\n", ".. _`Support Vector Classification`: https://scikit-learn.org/0.20/modules/generated/sklearn.svm.SVC.html#sklearn-svm-svc\n", "\"\"\",\n", " 'documentation_url': 'https://lale.readthedocs.io/en/latest/modules/lale.lib.sklearn.svc.html',\n", " 'import_from': 'sklearn.svm',\n", " 'type': 'object',\n", " 'tags': {\n", " 'pre': [],\n", " 'op': ['estimator', 'classifier'],\n", " 'post': []},\n", " 'properties': {\n", " 'hyperparams': _hyperparams_schema,\n", " 'input_fit': _input_fit_schema,\n", " 'input_predict': _input_predict_schema,\n", " 'output_predict': _output_predict_schema,\n", " 'input_predict_proba': _input_predict_proba_schema,\n", " 'output_predict_proba': _output_predict_proba_schema,\n", " 'input_decision_function': _input_decision_function_schema,\n", " 'output_decision_function': _output_decision_function_schema}}\n", "\n", "SVC : lale.operators.IndividualOp\n", "SVC = lale.operators.make_operator(SVCImpl, _combined_schemas)\n", "\n", "if sklearn.__version__ >= '0.22':\n", " # old: https://scikit-learn.org/0.20/modules/generated/sklearn.svm.SVC.html\n", " # new: https://scikit-learn.org/0.23/modules/generated/sklearn.svm.SVC.html\n", " from lale.schemas import AnyOf, Bool, Enum, Float\n", " SVC = SVC.customize_schema(\n", " gamma=AnyOf(\n", " types=[\n", " Enum(['scale', 'auto']),\n", " Float(\n", " min=0.0,\n", " exclusiveMin=True,\n", " minForOptimizer=3.0517578125e-05,\n", " maxForOptimizer=8,\n", " distribution='loguniform')],\n", " desc=\"Kernel coefficient for 'rbf', 'poly' and 'sigmoid'.\",\n", " default='scale'),\n", " break_ties=Bool(\n", " desc=\"If true, decision_function_shape='ovr', and number of classes > 2, predict will break ties according to the confidence values of decision_function; otherwise the first class among the tied classes is returned.\",\n", " default=False))\n", "\n", "lale.docstrings.set_docstrings(SVCImpl, SVC._schemas)\n", "\n" ] } ], "source": [ "read_data = \"\"\n", "\n", "with open('svc.py') as f:\n", " read_data = f.read()\n", " print(read_data)" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "ename": "SyntaxError", "evalue": "invalid syntax (, line 1)", "output_type": "error", "traceback": [ "\u001b[1;36m File \u001b[1;32m\"\"\u001b[1;36m, line \u001b[1;32m1\u001b[0m\n\u001b[1;33m import 'svc'\u001b[0m\n\u001b[1;37m ^\u001b[0m\n\u001b[1;31mSyntaxError\u001b[0m\u001b[1;31m:\u001b[0m invalid syntax\n" ] } ], "source": [ "import svc" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.8.5" } }, "nbformat": 4, "nbformat_minor": 4 }