Mercurial > repos > bgruening > sklearn_discriminant_classifier

comparison discriminant.xml @ 41:d769d83ec796 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
author bgruening
date Wed, 09 Aug 2023 13:14:12 +0000
parents eeaf989f1024
children
comparison
equal deleted inserted replaced
40:4335cec181ba 41:d769d83ec796
1 <tool id="sklearn_discriminant_classifier" name="Discriminant Analysis" version="@VERSION@" profile="20.05"> 1 <tool id="sklearn_discriminant_classifier" name="Discriminant Analysis" version="@VERSION@" profile="@PROFILE@">
2 <description></description> 2 <description></description>
3 <macros> 3 <macros>
4 <import>main_macros.xml</import> 4 <import>main_macros.xml</import>
5 <!--macro name="priors"--> 5 <!--macro name="priors"-->
6 </macros> 6 </macros>
20 import pandas 20 import pandas
21 import pickle 21 import pickle
22 import sklearn.discriminant_analysis 22 import sklearn.discriminant_analysis
23 import sys 23 import sys
24 24
25 from galaxy_ml.utils import load_model, get_X_y 25 from galaxy_ml.model_persist import dump_model_to_h5, load_model_from_h5
26 from galaxy_ml.utils import clean_params, get_X_y
26 27
27 28
28 input_json_path = sys.argv[1] 29 input_json_path = sys.argv[1]
29 with open(input_json_path, "r") as param_handler: 30 with open(input_json_path, "r") as param_handler:
30 params = json.load(param_handler) 31 params = json.load(param_handler)
31 32
32 #if $selected_tasks.selected_task == "load": 33 #if $selected_tasks.selected_task == "load":
33 34
34 with open("$infile_model", 'rb') as model_handler: 35 classifier_object = load_model_from_h5('$infile_model')
35 classifier_object = load_model(model_handler) 36 classifier_object = clean_params(classifier_object)
36 37
37 header = 'infer' if params["selected_tasks"]["header"] else None 38 header = 'infer' if params["selected_tasks"]["header"] else None
38 data = pandas.read_csv("$selected_tasks.infile_data", sep='\t', header=header, index_col=None, parse_dates=True, encoding=None) 39 data = pandas.read_csv("$selected_tasks.infile_data", sep='\t', header=header, index_col=None, parse_dates=True, encoding=None)
39 prediction = classifier_object.predict(data) 40 prediction = classifier_object.predict(data)
40 prediction_df = pandas.DataFrame(prediction) 41 prediction_df = pandas.DataFrame(prediction)
49 selected_algorithm = params["selected_tasks"]["selected_algorithms"]["selected_algorithm"] 50 selected_algorithm = params["selected_tasks"]["selected_algorithms"]["selected_algorithm"]
50 51
51 my_class = getattr(sklearn.discriminant_analysis, selected_algorithm) 52 my_class = getattr(sklearn.discriminant_analysis, selected_algorithm)
52 classifier_object = my_class(**options) 53 classifier_object = my_class(**options)
53 classifier_object.fit(X, y) 54 classifier_object.fit(X, y)
54 with open("$outfile_fit", 'wb') as out_handler: 55 dump_model_to_h5(classifier_object, '$outfile_fit')
55 pickle.dump(classifier_object, out_handler, pickle.HIGHEST_PROTOCOL)
56 56
57 #end if 57 #end if
58 ]]> 58 ]]>
59 </configfile> 59 </configfile>
60 </configfiles> 60 </configfiles>
61 <inputs> 61 <inputs>
62 <expand macro="sl_Conditional" model="zip"> 62 <expand macro="sl_Conditional" model="h5mlm">
63 <param name="selected_algorithm" type="select" label="Classifier type"> 63 <param name="selected_algorithm" type="select" label="Classifier type">
64 <option value="LinearDiscriminantAnalysis" selected="true">Linear Discriminant Classifier</option> 64 <option value="LinearDiscriminantAnalysis" selected="true">Linear Discriminant Classifier</option>
65 <option value="QuadraticDiscriminantAnalysis">Quadratic Discriminant Classifier</option> 65 <option value="QuadraticDiscriminantAnalysis">Quadratic Discriminant Classifier</option>
66 </param> 66 </param>
67 <when value="LinearDiscriminantAnalysis"> 67 <when value="LinearDiscriminantAnalysis">
93 <expand macro="output" /> 93 <expand macro="output" />
94 <tests> 94 <tests>
95 <test> 95 <test>
96 <param name="infile1" value="train.tabular" ftype="tabular" /> 96 <param name="infile1" value="train.tabular" ftype="tabular" />
97 <param name="infile2" value="train.tabular" ftype="tabular" /> 97 <param name="infile2" value="train.tabular" ftype="tabular" />
98 <param name="header1" value="True" />
99 <param name="header2" value="True" />
100 <param name="col1" value="1,2,3,4" /> 98 <param name="col1" value="1,2,3,4" />
101 <param name="col2" value="5" /> 99 <param name="col2" value="5" />
102 <param name="selected_task" value="train" /> 100 <param name="selected_task" value="train" />
103 <param name="selected_algorithm" value="LinearDiscriminantAnalysis" /> 101 <param name="selected_algorithm" value="LinearDiscriminantAnalysis" />
104 <param name="solver" value="svd" /> 102 <param name="solver" value="svd" />
106 <output name="outfile_fit" file="lda_model01" compare="sim_size" delta="1" /> 104 <output name="outfile_fit" file="lda_model01" compare="sim_size" delta="1" />
107 </test> 105 </test>
108 <test> 106 <test>
109 <param name="infile1" value="train.tabular" ftype="tabular" /> 107 <param name="infile1" value="train.tabular" ftype="tabular" />
110 <param name="infile2" value="train.tabular" ftype="tabular" /> 108 <param name="infile2" value="train.tabular" ftype="tabular" />
111 <param name="header1" value="True" />
112 <param name="header2" value="True" />
113 <param name="col1" value="1,2,3,4" /> 109 <param name="col1" value="1,2,3,4" />
114 <param name="col2" value="5" /> 110 <param name="col2" value="5" />
115 <param name="selected_task" value="train" /> 111 <param name="selected_task" value="train" />
116 <param name="selected_algorithm" value="LinearDiscriminantAnalysis" /> 112 <param name="selected_algorithm" value="LinearDiscriminantAnalysis" />
117 <param name="solver" value="lsqr" /> 113 <param name="solver" value="lsqr" />
118 <output name="outfile_fit" file="lda_model02" compare="sim_size" delta="1" /> 114 <output name="outfile_fit" file="lda_model02" compare="sim_size" delta="1" />
119 </test> 115 </test>
120 <test> 116 <test>
121 <param name="infile1" value="train.tabular" ftype="tabular" /> 117 <param name="infile1" value="train.tabular" ftype="tabular" />
122 <param name="infile2" value="train.tabular" ftype="tabular" /> 118 <param name="infile2" value="train.tabular" ftype="tabular" />
123 <param name="header1" value="True" />
124 <param name="header2" value="True" />
125 <param name="col1" value="1,2,3,4" /> 119 <param name="col1" value="1,2,3,4" />
126 <param name="col2" value="5" /> 120 <param name="col2" value="5" />
127 <param name="selected_task" value="train" /> 121 <param name="selected_task" value="train" />
128 <param name="selected_algorithm" value="QuadraticDiscriminantAnalysis" /> 122 <param name="selected_algorithm" value="QuadraticDiscriminantAnalysis" />
129 <output name="outfile_fit" file="qda_model01" compare="sim_size" delta="1" /> 123 <output name="outfile_fit" file="qda_model01" compare="sim_size" delta="1" />
130 </test> 124 </test>
131 <test> 125 <test>
132 <param name="infile_model" value="lda_model01" ftype="zip" /> 126 <param name="infile_model" value="lda_model01" ftype="h5mlm" />
133 <param name="infile_data" value="test.tabular" ftype="tabular" /> 127 <param name="infile_data" value="test.tabular" ftype="tabular" />
134 <param name="header" value="True" />
135 <param name="selected_task" value="load" /> 128 <param name="selected_task" value="load" />
136 <output name="outfile_predict" file="lda_prediction_result01.tabular" /> 129 <output name="outfile_predict" file="lda_prediction_result01.tabular" />
137 </test> 130 </test>
138 <test> 131 <test>
139 <param name="infile_model" value="lda_model02" ftype="zip" /> 132 <param name="infile_model" value="lda_model02" ftype="h5mlm" />
140 <param name="infile_data" value="test.tabular" ftype="tabular" /> 133 <param name="infile_data" value="test.tabular" ftype="tabular" />
141 <param name="header" value="True" />
142 <param name="selected_task" value="load" /> 134 <param name="selected_task" value="load" />
143 <output name="outfile_predict" file="lda_prediction_result02.tabular" /> 135 <output name="outfile_predict" file="lda_prediction_result02.tabular" />
144 </test> 136 </test>
145 <test> 137 <test>
146 <param name="infile_model" value="qda_model01" ftype="zip" /> 138 <param name="infile_model" value="qda_model01" ftype="h5mlm" />
147 <param name="infile_data" value="test.tabular" ftype="tabular" /> 139 <param name="infile_data" value="test.tabular" ftype="tabular" />
148 <param name="header" value="True" />
149 <param name="selected_task" value="load" /> 140 <param name="selected_task" value="load" />
150 <output name="outfile_predict" file="qda_prediction_result01.tabular" /> 141 <output name="outfile_predict" file="qda_prediction_result01.tabular" />
151 </test> 142 </test>
152 </tests> 143 </tests>
153 <help><![CDATA[ 144 <help><![CDATA[