Mercurial > repos > jay > pdaug_tsvtofasta

annotate PDAUG_Merge_Dataframes/PDAUG_Merge_Dataframes.py @ 6:391e7e836fe9 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
author jay
date Tue, 12 Jan 2021 18:40:09 +0000
parents c3f0b3a6339e
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
c3f0b3a6339e "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit a9bd83f6a1afa6338cb6e4358b63ebff5bed155e"
jay
parents:
diff changeset
1 import glob
c3f0b3a6339e "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit a9bd83f6a1afa6338cb6e4358b63ebff5bed155e"
jay
parents:
diff changeset
2 import pandas as pd
c3f0b3a6339e "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit a9bd83f6a1afa6338cb6e4358b63ebff5bed155e"
jay
parents:
diff changeset
3 import sys
c3f0b3a6339e "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit a9bd83f6a1afa6338cb6e4358b63ebff5bed155e"
jay
parents:
diff changeset
4
6
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
5
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
6 def MergeData(infiles, add_class_label, class_label, OutPut):
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
7
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
8 data_frame = pd.DataFrame()
0
c3f0b3a6339e "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit a9bd83f6a1afa6338cb6e4358b63ebff5bed155e"
jay
parents:
diff changeset
9
6
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
10 if add_class_label == 'True' or add_class_label == 'true':
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
11 for i, file in enumerate(infiles.split(',')):
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
12 df1 = pd.read_csv(file,sep='\t')
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
13 df2 = pd.DataFrame(df1.shape[0]*[i], columns=[class_label])
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
14 df3 = pd.concat([df1,df2], axis=1)
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
15 data_frame = pd.concat([data_frame,df3])
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
16 final_DF = data_frame.fillna(0)
0
c3f0b3a6339e "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit a9bd83f6a1afa6338cb6e4358b63ebff5bed155e"
jay
parents:
diff changeset
17
6
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
18 else:
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
19
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
20 for file in infiles.split(','):
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
21 df1 = pd.read_csv(file,sep='\t')
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
22 data_frame = pd.concat([data_frame,df1])
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
23 final_DF = data_frame.fillna(0)
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
24
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
25 final_DF.to_csv(OutPut, sep="\t", index=False)
0
c3f0b3a6339e "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit a9bd83f6a1afa6338cb6e4358b63ebff5bed155e"
jay
parents:
diff changeset
26
c3f0b3a6339e "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit a9bd83f6a1afa6338cb6e4358b63ebff5bed155e"
jay
parents:
diff changeset
27
6
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
28 if __name__=="__main__":
0
c3f0b3a6339e "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit a9bd83f6a1afa6338cb6e4358b63ebff5bed155e"
jay
parents:
diff changeset
29
6
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
30 import argparse
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
31 parser = argparse.ArgumentParser()
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
32 parser.add_argument("-I", "--infiles", required=True, default=None, help=".tsv")
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
33 parser.add_argument("-L", "--add_class_label", required=False, default=False, help="Path to target tsv file")
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
34 parser.add_argument("-C", "--class_label", required=False, default='class_label', help="Path to target tsv file")
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
35 parser.add_argument("-O", "--OutPut", required=False, default='Out.tsv', help="Path to target tsv file")
0
c3f0b3a6339e "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit a9bd83f6a1afa6338cb6e4358b63ebff5bed155e"
jay
parents:
diff changeset
36
6
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
37 args = parser.parse_args()
0
c3f0b3a6339e "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit a9bd83f6a1afa6338cb6e4358b63ebff5bed155e"
jay
parents:
diff changeset
38
6
391e7e836fe9 "planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit 45ebf32dcaa1eed91670d3a2491f9cf3dfb535ef"
jay
parents: 0
diff changeset
39 MergeData(args.infiles, args.add_class_label, args.class_label, args.OutPut)