annotate model_validation.xml @ 16:86e1e2874460 draft

planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 57f4407e278a615f47a377a3328782b1d8e0b54d

<tool id="sklearn_model_validation" name="Model Validation" version="@VERSION@">
    <description>evaluates estimator performance by cross-validation</description>
    <macros>
        <import>main_macros.xml</import>
    </macros>
    <expand macro="python_requirements"/>
    <expand macro="macro_stdio"/>
    <version_command>echo "@VERSION@"</version_command>
    <command>
        <![CDATA[
        python "$sklearn_model_validation_script" '$inputs'
        ]]>
    </command>
    <configfiles>
        <inputs name="inputs" />
        <configfile name="sklearn_model_validation_script">
            <![CDATA[
import sys
import os
import json
import pandas
import numpy as np
from sklearn import preprocessing, model_selection, svm, linear_model, ensemble, naive_bayes, tree, neighbors
from sklearn.pipeline import Pipeline

exec(open('$__tool_directory__/utils.py').read(), globals())

warnings.filterwarnings('ignore')

safe_eval = SafeEval()

input_json_path = sys.argv[1]
with open(input_json_path, 'r') as param_handler:
    params = json.load(param_handler)

input_type = params['input_options']['selected_input']
if input_type == 'tabular':
    header = 'infer' if params['input_options']['header1'] else None
    column_option = params['input_options']['column_selector_options_1']['selected_column_selector_option']
    if column_option in ['by_index_number', 'all_but_by_index_number', 'by_header_name', 'all_but_by_header_name']:
        c = params['input_options']['column_selector_options_1']['col1']
    else:
        c = None
    X = read_columns(
            '$input_options.infile1',
            c = c,
            c_option = column_option,
            sep='\t',
            header=header,
            parse_dates=True
    )
else:
    X = mmread('$input_options.infile1')

header = 'infer' if params['input_options']['header2'] else None
column_option = params['input_options']['column_selector_options_2']['selected_column_selector_option2']
if column_option in ['by_index_number', 'all_but_by_index_number', 'by_header_name', 'all_but_by_header_name']:
    c = params['input_options']['column_selector_options_2']['col2']
else:
    c = None
y = read_columns(
        '$input_options.infile2',
        c = c,
        c_option = column_option,
        sep='\t',
        header=header,
        parse_dates=True
)
y=y.ravel()

options = params['model_validation_functions']['options']
splitter, groups = get_cv( options.pop('cv_selector') )
if groups is None:
    options['cv'] = splitter
elif groups == '':
    options['cv'] = list( splitter.split(X, y, groups=None) )
else:
    options['cv'] = list( splitter.split(X, y, groups=groups) )
options['n_jobs'] = N_JOBS
if 'scoring' in options:
    options['scoring'] = get_scoring(options['scoring'])
if 'pre_dispatch' in options and options['pre_dispatch'] == '':
    options['pre_dispatch'] = None

pipeline_steps = []

## Set up pre_processor and add to pipeline steps.
if params['pre_processing']['do_pre_processing'] == 'Yes':
    preprocessor = params['pre_processing']['pre_processors']['selected_pre_processor']
    pre_processor_options = params['pre_processing']['pre_processors']['options']
    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']['fs_algorithm_selector'])
    pipeline_steps.append( ('feature_selector', feature_selector) )

## Set up estimator and add to pipeline.
estimator_json = params['model_validation_functions']['estimator_selector']
estimator = get_estimator(estimator_json)

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(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':
    res = validator(pipeline, X, y, **options)
    rval = res[rval_type]
elif selected_function == 'learning_curve':
    options['train_sizes'] = eval(options['train_sizes'])
    train_sizes_abs, train_scores, test_scores = validator(pipeline, X, y, **options)
    rval = eval(rval_type)
elif selected_function == 'permutation_test_score':
    score, permutation_scores, pvalue = validator(pipeline, X, y, **options)
    rval = eval(rval_type)
    if rval_type in ['score', 'pvalue']:
        rval = [rval]
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)
else:
    rval = validator(pipeline, X, y, **options)

rval = pandas.DataFrame(rval)
rval.to_csv(path_or_buf='$outfile', sep='\t', header=False, index=False)

            ]]>
        </configfile>
    </configfiles>
    <inputs>
        <conditional name="pre_processing">
            <param name="do_pre_processing" type="select" label="Do pre_processing?">
                <option value="No" selected="true"/>
                <option value="Yes"/>
            </param>
            <when value="No"/>
            <when value="Yes">
                <conditional name="pre_processors">
                    <expand macro="sparse_preprocessors_ext" />
                    <expand macro="sparse_preprocessor_options_ext" />
                </conditional>
            </when>
        </conditional>
        <conditional name="feature_selection">
            <param name="do_feature_selection" type="select" label="Do feature selection?">
                <option value="No" selected="true"/>
                <option value="Yes"/>
            </param>
            <when value="No"/>
            <when value="Yes">
                <expand macro="feature_selection_pipeline"/>
            </when>
        </conditional>
        <conditional name="model_validation_functions">
            <param name="selected_function" type="select" label="Select a model validation function">
                <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="cross_validate">
                <expand macro="estimator_selector_all" />
                <section name="options" title="Other Options" expanded="false">
                    <!--groups-->
                    <expand macro="model_validation_common_options"/>
                    <expand macro="scoring_selection"/>
                    <!--fit_params-->
                    <expand macro="pre_dispatch"/>
                </section>
                <param name="return_type" type="select" label="Select a return type">
                    <option value="test_score" selected="true">test_score</option>
                    <option value="train_score">train_score</option>
                    <option value="fit_time">fit_time</option>
                    <option value="score_time">score_time</option>
                </param>
            </when>
            <when value="cross_val_predict">
                <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"/>
                    <param argument="method" type="select" label="Invokes the passed method name of the passed estimator">
                        <option value="predict" selected="true">predict</option>
                        <option value="predict_proba">predict_proba</option>
                    </param>
                </section>
            </when>
            <when value="cross_val_score">
                <expand macro="estimator_selector_all" />
                <section name="options" title="Other Options" expanded="false">
                    <!--groups-->
                    <expand macro="model_validation_common_options"/>
                    <expand macro="scoring_selection"/>
                    <!--fit_params-->
                    <expand macro="pre_dispatch"/>
                </section>
            </when>
            <when value="learning_curve">
                <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_selection"/>
                    <param argument="exploit_incremental_learning" type="boolean" optional="true" truevalue="booltrue" falsevalue="boolfalse" checked="false" label="exploit_incremental_learning" help="Whether to apply incremental learning to speed up fitting of the estimator if supported"/>
                    <expand macro="pre_dispatch"/>
                    <expand macro="shuffle" checked="false" label="shuffle" help="Whether to shuffle training data before taking prefixes"/>
                    <expand macro="random_state"/>
                </section>
                <param name="return_type" type="select" label="Select a return type">
                    <option value="train_sizes_abs" selected="true">train_sizes_abs</option>
                    <option value="train_scores">train_scores</option>
                    <option value="test_scores">test_scores</option>
                </param>
            </when>
            <when value="permutation_test_score">
                <expand macro="estimator_selector_all" />
                <section name="options" title="Other Options" expanded="false">
                    <!--groups-->
                    <expand macro="model_validation_common_options"/>
                    <expand macro="scoring_selection"/>
                    <param name="n_permutations" type="integer" value="100" optional="true" label="n_permutations" help="Number of times to permute y"/>
                    <expand macro="random_state"/>
                </section>
                <param name="return_type" type="select" label="Select a return type">
                    <option value="score" selected="true">score</option>
                    <option value="permutation_scores">permutation_scores</option>
                    <option value="pvalue">pvalue</option>
                </param>
            </when>
            <when value="validation_curve">
                <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"/>
                    <expand macro="scoring_selection"/>
                    <expand macro="pre_dispatch"/>
                </section>
                <param name="return_type" type="select" label="Select a return type">
                    <option value="train_scores" selected="true">train_scores</option>
                    <option value="test_scores">test_scores</option>
                </param>
            </when>
        </conditional>
        <expand macro="sl_mixed_input"/>
    </inputs>
    <outputs>
        <data format="tabular" name="outfile"/>
    </outputs>
    <tests>
        <test>
            <param name="selected_function" value="cross_validate"/>
            <param name="selected_module" value="linear_model"/>
            <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="selected_module" value="linear_model"/>
            <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="selected_module" value="linear_model"/>
            <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="selected_module" value="linear_model"/>
            <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="selected_module" value="linear_model"/>
            <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="selected_module" value="svm"/>
            <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>
    </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 model validation functions and their parameter settings please refer to `Scikit-learn model_selection`_.

.. _`Scikit-learn model_selection`: http://scikit-learn.org/stable/modules/classes.html#module-sklearn.model_selection
        ]]>
    </help>
    <expand macro="sklearn_citation">
        <expand macro="skrebate_citation"/>
        <expand macro="xgboost_citation"/>
    </expand>
</tool>