sklearn_model_validation: model_validation.xml comparison

comparison model_validation.xml @ 9:2593ba492276 draft

planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 76583c1fcd9d06a4679cc46ffaee44117b9e22cd

author	bgruening
date	Sat, 04 Aug 2018 12:18:01 -0400
parents	5f78d3786bfc
children	61844bce4115

comparison

equal deleted inserted replaced

-:5f78d3786bfc
+:2593ba492276
 import json
 import pandas
 import ast
 import pickle
 import numpy as np
-import sklearn.model_selection
+import sklearn.feature_selection
-from sklearn import svm, linear_model, ensemble, preprocessing
+from sklearn import preprocessing, model_selection, svm, linear_model, ensemble, naive_bayes, tree, neighbors
 from sklearn.pipeline import Pipeline
 @COLUMNS_FUNCTION@
+@GET_ESTIMATOR_FUNCTION@
 @FEATURE_SELECTOR_FUNCTION@
 input_json_path = sys.argv[1]
 with open(input_json_path, "r") as param_handler:
 params = json.load(param_handler)
 my_class = getattr(preprocessing, preprocessor)
 pipeline_steps.append( ('pre_processor', my_class(**pre_processor_options)) )
 ## Set up feature selector and add to pipeline steps.
 if params['feature_selection']['do_feature_selection'] == 'Yes':
-feature_selector = feature_selector(params['feature_selection']['feature_selection_algorithms'])
+feature_selector = feature_selector(params['feature_selection']['fs_algorithm_selector'])
 pipeline_steps.append( ('feature_selector', feature_selector) )
 ## Set up estimator and add to pipeline.
-estimator=params["model_validation_functions"]["estimator"]
+estimator_json = params["model_validation_functions"]['estimator_selector']
-if params["model_validation_functions"]["extra_estimator"]["has_estimator"] == 'no':
+estimator = get_estimator(estimator_json)
-estimator = params["model_validation_functions"]["extra_estimator"]["new_estimator"]
-estimator = eval(estimator.replace('__dq__', '"').replace("__sq__","'"))
 pipeline_steps.append( ('estimator', estimator) )
 pipeline = Pipeline(pipeline_steps)
 ## Set up validator, run pipeline through validator and return results.
 validator = params["model_validation_functions"]["selected_function"]
-validator = getattr(sklearn.model_selection, validator)
+validator = getattr(model_selection, validator)
 selected_function = params["model_validation_functions"]["selected_function"]
 rval_type = params["model_validation_functions"].get("return_type", None)
 if selected_function == 'cross_validate':
 elif selected_function == 'validation_curve':
 options['param_name'] = 'estimator__' + options['param_name']
 options['param_range'] = eval(options['param_range'])
 train_scores, test_scores = validator(pipeline, X, y, **options)
 rval = eval(rval_type)
-elif selected_function == 'GridSearchCV':
-param_grid = params["model_validation_functions"]["param_grid"].replace("__sq__","'")\
-.replace('__dq__','"').replace("__oc__", "{").replace("__cc__", "}")\
-.replace("__ob__", "[").replace("__cb__", "]")
-param_grid = ast.literal_eval(param_grid)
-grid = validator(pipeline, param_grid, **options)
-grid.fit(X, y)
-rval = getattr(grid, rval_type)
-if rval_type in ["best_estimator_", "best_score_", "best_index_"]:
-rval = [rval]
 else:
 rval = validator(pipeline, X, y, **options)
 rval = pandas.DataFrame(rval)
-if rval_type and rval_type == "cv_results_":
+rval.to_csv(path_or_buf="$outfile", sep='\t', header=False, index=False)
-rval.to_csv(path_or_buf="$outfile", sep='\t', header=True, index=False)
-else:
-rval.to_csv(path_or_buf="$outfile", sep='\t', header=False, index=False)
 ]]>
 </configfile>
 </configfiles>
 <inputs>
 <option value="No" selected="true"/>
 <option value="Yes"/>
 </param>
 <when value="No"/>
 <when value="Yes">
-<expand macro="feature_selection_all"/>
+<expand macro="feature_selection_all">
+<expand macro="fs_selectfrommodel_no_prefitted"/>
+</expand>
 </when>
 </conditional>
 <conditional name="model_validation_functions">
 <param name="selected_function" type="select" label="Select a model validation function">
-<option value="GridSearchCV">GridSearchCV - Exhaustive search over specified parameter values for an estimator </option>
 <option value="cross_validate">cross_validate - Evaluate metric(s) by cross-validation and also record fit/score times</option>
 <option value="cross_val_predict">cross_val_predict - Generate cross-validated estimates for each input data point</option>
 <option value="cross_val_score">cross_val_score - Evaluate a score by cross-validation</option>
 <option value="learning_curve">learning_curve - Learning curve</option>
 <option value="permutation_test_score">permutation_test_score - Evaluate the significance of a cross-validated score with permutations</option>
 <option value="validation_curve">validation_curve - Validation curve</option>
 </param>
-<when value="GridSearchCV">
-<expand macro="estimator_input_no_fit" />
-<param argument="param_grid" type="text" value="[{'feature_selector__k': [3, 5, 7, 9], 'estimator__C': [1, 10, 100, 1000]}]" label="param_grid" help="Dictionary with parameters names (string) as keys and lists of parameter settings to try as values, or a list of such dictionaries, in which case the grids spanned by each dictionary in the list are explored"/>
-<section name="options" title="Other Options" expanded="false">
-<expand macro="scoring"/>
-<expand macro="model_validation_common_options"/>
-<expand macro="pre_dispatch" value="2*n_jobs" help="Controls the number of jobs that get dispatched during parallel execution"/>
-<param argument="iid" type="boolean" truevalue="booltrue" falsevalue="boolfalse" checked="true" label="iid" help="Data is identically distributed?"/>
-<param argument="refit" type="boolean" truevalue="booltrue" falsevalue="boolfalse" checked="true" label="refit" help="Refit an estimator using the best found parameters on the whole dataset."/>
-<!--error_score-->
-<param argument="return_train_score" type="boolean" truevalue="booltrue" falsevalue="boolfalse" checked="false" label="return_train_score" help=""/>
-</section>
-<param name="return_type" type="select" label="Select a return type">
-<option value="cv_results_" selected="true">cv_results_</option>
-<option value="best_estimator_">best_estimator_</option>
-<option value="best_score_">best_score_</option>
-<option value="best_params_">best_params_</option>
-<option value="best_index_">best_index_</option>
-</param>
-</when>
 <when value="cross_validate">
-<expand macro="estimator_input_no_fit" />
+<expand macro="estimator_selector_all" />
 <section name="options" title="Other Options" expanded="false">
 <!--groups-->
 <expand macro="model_validation_common_options"/>
 <expand macro="scoring"/>
 <!--fit_params-->
 <option value="fit_time">fit_time</option>
 <option value="score_time">score_time</option>
 </param>
 </when>
 <when value="cross_val_predict">
-<expand macro="estimator_input_no_fit" />
+<expand macro="estimator_selector_all" />
 <section name="options" title="Other Options" expanded="false">
 <!--groups-->
 <expand macro="model_validation_common_options" />
 <!--fit_params-->
 <expand macro="pre_dispatch" value="2*n_jobs’" help="Controls the number of jobs that get dispatched during parallel execution"/>
 <option value="predict_proba">predict_proba</option>
 </param>
 </section>
 </when>
 <when value="cross_val_score">
-<expand macro="estimator_input_no_fit" />
+<expand macro="estimator_selector_all" />
 <section name="options" title="Other Options" expanded="false">
 <!--groups-->
 <expand macro="model_validation_common_options"/>
 <expand macro="scoring"/>
 <!--fit_params-->
 <expand macro="pre_dispatch"/>
 </section>
 </when>
 <when value="learning_curve">
-<expand macro="estimator_input_no_fit" />
+<expand macro="estimator_selector_all" />
 <section name="options" title="Other Options" expanded="false">
 <!--groups-->
 <expand macro="model_validation_common_options"/>
 <param argument="train_sizes" type="text" value="np.linspace(0.1, 1.0, 5)" label="train_sizes" help="Relative or absolute numbers of training examples that will be used to generate the learning curve"/>
 <expand macro="scoring"/>
 <option value="train_scores">train_scores</option>
 <option value="test_scores">test_scores</option>
 </param>
 </when>
 <when value="permutation_test_score">
-<expand macro="estimator_input_no_fit" />
+<expand macro="estimator_selector_all" />
 <section name="options" title="Other Options" expanded="false">
 <!--groups-->
 <expand macro="model_validation_common_options"/>
 <expand macro="scoring"/>
 <param name="n_permutations" type="integer" value="100" optional="true" label="n_permutations" help="Number of times to permute y"/>
 <option value="permutation_scores">permutation_scores</option>
 <option value="pvalue">pvalue</option>
 </param>
 </when>
 <when value="validation_curve">
-<expand macro="estimator_input_no_fit" />
+<expand macro="estimator_selector_all" />
 <section name="options" title="Other Options" expanded="false">
 <param name="param_name" type="text" value="gamma" label="param_name" help="Name of the parameter that will be varied"/>
 <param name="param_range" type="text" value="np.logspace(-6, -1, 5)" label="param_range" help="The values of the parameter that will be evaluated."/>
 <!--groups-->
 <expand macro="model_validation_common_options"/>
 <data format="tabular" name="outfile"/>
 </outputs>
 <tests>
 <test>
 <param name="selected_function" value="cross_validate"/>
-<param name="estimator" value="linear_model.LassoCV()"/>
+<param name="selected_module" value="linear_model"/>
-<param name="has_estimator" value="yes"/>
+<param name="selected_estimator" value="LassoCV"/>
 <param name="infile1" value="regression_train.tabular" ftype="tabular"/>
 <param name="col1" value="1,2,3,4,5"/>
 <param name="infile2" value="regression_train.tabular" ftype="tabular"/>
 <param name="col2" value="6"/>
 <output name="outfile" file="mv_result01.tabular"/>
 </test>
 <test>
 <param name="selected_function" value="cross_val_predict"/>
-<param name="estimator" value="linear_model.LassoCV()"/>
+<param name="selected_module" value="linear_model"/>
-<param name="has_estimator" value="yes"/>
+<param name="selected_estimator" value="LassoCV"/>
 <param name="infile1" value="regression_train.tabular" ftype="tabular"/>
 <param name="col1" value="1,2,3,4,5"/>
 <param name="infile2" value="regression_train.tabular" ftype="tabular"/>
 <param name="col2" value="6"/>
 <output name="outfile" file="mv_result02.tabular"/>
 </test>
 <test>
 <param name="selected_function" value="cross_val_score"/>
-<param name="estimator" value="linear_model.LassoCV()"/>
+<param name="selected_module" value="linear_model"/>
-<param name="has_estimator" value="yes"/>
+<param name="selected_estimator" value="LassoCV"/>
 <param name="infile1" value="regression_train.tabular" ftype="tabular"/>
 <param name="col1" value="1,2,3,4,5"/>
 <param name="infile2" value="regression_train.tabular" ftype="tabular"/>
 <param name="col2" value="6"/>
 <output name="outfile" file="mv_result03.tabular"/>
 </test>
 <test>
 <param name="selected_function" value="learning_curve"/>
-<param name="estimator" value="linear_model.LassoCV()"/>
+<param name="selected_module" value="linear_model"/>
-<param name="has_estimator" value="yes"/>
+<param name="selected_estimator" value="LassoCV"/>
 <param name="infile1" value="regression_X.tabular" ftype="tabular"/>
 <param name="header1" value="true" />
 <param name="col1" value="1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17"/>
 <param name="infile2" value="regression_y.tabular" ftype="tabular"/>
 <param name="header2" value="true" />
 <param name="col2" value="1"/>
 <output name="outfile" file="mv_result04.tabular"/>
 </test>
 <test>
 <param name="selected_function" value="permutation_test_score"/>
-<param name="estimator" value="linear_model.LassoCV()"/>
+<param name="selected_module" value="linear_model"/>
-<param name="has_estimator" value="yes"/>
+<param name="selected_estimator" value="LassoCV"/>
 <param name="infile1" value="regression_train.tabular" ftype="tabular"/>
 <param name="col1" value="1,2,3,4,5"/>
 <param name="infile2" value="regression_train.tabular" ftype="tabular"/>
 <param name="col2" value="6"/>
 <output name="outfile" file="mv_result05.tabular"/>
 </test>
 <test>
 <param name="selected_function" value="validation_curve"/>
-<param name="estimator" value="svm.SVC(kernel=&quot;linear&quot;)"/>
+<param name="selected_module" value="svm"/>
-<param name="has_estimator" value="yes"/>
+<param name="selected_estimator" value="SVC"/>
+<param name="text_params" value="'kernel': 'linear'"/>
 <param name="infile1" value="regression_X.tabular" ftype="tabular"/>
 <param name="header1" value="true" />
 <param name="selected_column_selector_option" value="all_columns"/>
 <param name="infile2" value="regression_y.tabular" ftype="tabular"/>
 <param name="header2" value="true" />
 <param name="col2" value="1"/>
 <param name="return_type" value="test_scores"/>
 <output name="outfile" file="mv_result06.tabular"/>
 </test>
-<test>
-<param name="do_feature_selection" value="Yes"/>
-<param name="selected_algorithm" value="SelectKBest"/>
-<param name="score_func" value="chi2"/>
-<param name="selected_function" value="GridSearchCV"/>
-<param name="estimator" value="svm.SVR(kernel=&quot;linear&quot;)"/>
-<param name="has_estimator" value="yes"/>
-<param name="param_grid" value="[{'feature_selector__k': [3, 7], 'estimator__C': [1, 100]}]"/>
-<param name="return_type" value="best_score_"/>
-<param name="infile1" value="regression_X.tabular" ftype="tabular"/>
-<param name="header1" value="true" />
-<param name="selected_column_selector_option" value="all_columns"/>
-<param name="infile2" value="regression_y.tabular" ftype="tabular"/>
-<param name="header2" value="true" />
-<param name="selected_column_selector_option2" value="all_columns"/>
-<output name="outfile" >
-<assert_contents>
-<has_line line="0.7824428015300172" />
-</assert_contents>
-</output>
-</test>
-<test>
-<param name="do_pre_processing" value="Yes"/>
-<param name="selected_pre_processor" value="RobustScaler"/>
-<param name="do_feature_selection" value="Yes"/>
-<param name="selected_algorithm" value="SelectKBest"/>
-<param name="score_func" value="f_classif"/>
-<param name="selected_function" value="GridSearchCV"/>
-<param name="estimator" value="svm.SVR(kernel=&quot;linear&quot;)"/>
-<param name="has_estimator" value="yes"/>
-<param name="param_grid" value="[{'feature_selector__k': [3, 5, 7, 9], 'estimator__C': [1, 10, 100, 1000]}]"/>
-<param name="return_type" value="best_score_"/>
-<param name="infile1" value="regression_X.tabular" ftype="tabular"/>
-<param name="header1" value="true" />
-<param name="selected_column_selector_option" value="all_columns"/>
-<param name="infile2" value="regression_y.tabular" ftype="tabular"/>
-<param name="header2" value="true" />
-<param name="selected_column_selector_option2" value="all_columns"/>
-<output name="outfile" >
-<assert_contents>
-<has_line line="0.7938837807353147" />
-</assert_contents>
-</output>
-</test>
-<test>
-<param name="do_pre_processing" value="Yes"/>
-<param name="selected_pre_processor" value="RobustScaler"/>
-<param name="selected_function" value="GridSearchCV"/>
-<param name="estimator" value="svm.SVR(kernel=&quot;linear&quot;)"/>
-<param name="has_estimator" value="yes"/>
-<param name="param_grid" value="[{'estimator__C': [1, 10, 100, 1000]}]"/>
-<param name="return_type" value="best_score_"/>
-<param name="infile1" value="regression_X.tabular" ftype="tabular"/>
-<param name="header1" value="true" />
-<param name="selected_column_selector_option" value="all_columns"/>
-<param name="infile2" value="regression_y.tabular" ftype="tabular"/>
-<param name="header2" value="true" />
-<param name="selected_column_selector_option2" value="all_columns"/>
-<output name="outfile" >
-<assert_contents>
-<has_line line="0.7904476204861263" />
-</assert_contents>
-</output>
-</test>
 </tests>
 <help>
 <![CDATA[
 **What it does**
 This tool includes model validation functions to evaluate estimator performance in the cross-validation approach. This tool is based on
 sklearn.model_selection package.
-For information about classification metric functions and their parameter settings please refer to `Scikit-learn classification metrics`_.
+For information about model validation functions and their parameter settings please refer to `Scikit-learn model_selection`_.
-.. _`Scikit-learn classification metrics`: http://scikit-learn.org/stable/modules/model_evaluation.html#classification-metrics
+.. _`Scikit-learn model_selection`: http://scikit-learn.org/stable/modules/classes.html#module-sklearn.model_selection
 ]]>
 </help>
 <expand macro="sklearn_citation"/>
 </tool>

Mercurial > repos > bgruening > sklearn_model_validation

comparison model_validation.xml @ 9:2593ba492276 draft