Mercurial > repos > bgruening > sklearn_data_preprocess

annotate label_encoder.py @ 37:8ed14b9a1777 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 ca87db9c038a6fcf96aa39da50f384865fd932ff"
author bgruening
date Tue, 20 Apr 2021 16:59:56 +0000
parents
children 6c23a44a3c4f
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
37
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
1 import argparse
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
2 import json
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
3 import warnings
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
4
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
5 import numpy as np
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
6 import pandas as pd
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
7 from sklearn.preprocessing import LabelEncoder
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
8
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
9
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
10 def main(inputs, infile, outfile):
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
11 """
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
12 Parameter
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
13 ---------
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
14 input : str
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
15 File path to galaxy tool parameter
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
16
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
17 infile : str
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
18 File paths of input vector
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
19
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
20 outfile : str
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
21 File path to output vector
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
22
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
23 """
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
24 warnings.simplefilter('ignore')
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
25
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
26 with open(inputs, 'r') as param_handler:
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
27 params = json.load(param_handler)
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
28
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
29 input_header = params['header0']
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
30 header = 'infer' if input_header else None
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
31
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
32 input_vector = pd.read_csv(infile, sep='\t', header=header)
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
33
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
34 le = LabelEncoder()
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
35
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
36 output_vector = le.fit_transform(input_vector)
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
37
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
38 np.savetxt(outfile, output_vector, fmt="%d", delimiter='\t')
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
39
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
40
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
41 if __name__ == '__main__':
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
42 aparser = argparse.ArgumentParser()
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
43 aparser.add_argument("-i", "--inputs", dest="inputs", required=True)
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
44 aparser.add_argument("-y", "--infile", dest="infile")
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
45 aparser.add_argument("-o", "--outfile", dest="outfile")
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
46 args = aparser.parse_args()
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
47
8ed14b9a1777 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ca87db9c038a6fcf96aa39da50f384865fd932ff"
bgruening
parents:
diff changeset
48 main(args.inputs, args.infile, args.outfile)