Mercurial > repos > immport-devteam > extract_pop

annotate extractpop.py @ 1:4f28ee74079b draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
author azomics
date Fri, 24 Jul 2020 19:24:49 -0400
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
1
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
1 #!/usr/bin/env python
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
2
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
3 ######################################################################
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
4 # Copyright (c) 2016 Northrop Grumman.
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
5 # All rights reserved.
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
6 ######################################################################
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
7
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
8 from __future__ import print_function
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
9 import sys
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
10 import pandas as pd
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
11
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
12 from argparse import ArgumentParser
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
13
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
14
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
15 def is_int(s):
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
16 try:
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
17 int(s)
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
18 return True
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
19 except ValueError:
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
20 return False
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
21
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
22
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
23 def extract_pop(in_file, pop_list, out_file):
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
24 df = pd.read_table(in_file, dtype={'Population': object})
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
25 dfout = df.loc[df['Population'].isin(pop_list)]
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
26 dfout.to_csv(out_file, sep="\t", index=False)
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
27 return
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
28
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
29
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
30 def remove_pop(in_file, pop_list, out_file):
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
31 df = pd.read_table(in_file, dtype={'Population': object})
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
32 dfout = df.loc[~df['Population'].isin(pop_list)]
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
33 dfout.to_csv(out_file, sep="\t", index=False)
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
34 return
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
35
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
36
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
37 if __name__ == "__main__":
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
38 parser = ArgumentParser(
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
39 prog="ExtractPop",
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
40 description="Extract events associated to given population numbers.")
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
41
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
42 parser.add_argument(
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
43 '-i',
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
44 dest="input_file",
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
45 required=True,
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
46 help="File location for the text file.")
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
47
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
48 parser.add_argument(
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
49 '-p',
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
50 dest="pops",
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
51 required=True,
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
52 help="List of populations to extract.")
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
53
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
54 parser.add_argument(
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
55 '-o',
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
56 dest="output_file",
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
57 required=True,
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
58 help="Name of the output file.")
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
59
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
60 parser.add_argument(
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
61 '-m',
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
62 dest="method",
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
63 required=True,
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
64 help="What to do with the populations.")
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
65
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
66 args = parser.parse_args()
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
67
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
68 # check populations
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
69 default_values = ["i.e.:2,3,11,25", "default", "Default"]
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
70 populations = []
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
71 if args.pops:
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
72 if args.pops not in default_values:
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
73 tmp_pops = args.pops.split(",")
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
74 for popn in tmp_pops:
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
75 populations.append(popn.strip())
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
76 else:
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
77 sys.exit(2)
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
78 for pops in populations:
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
79 if not is_int(pops):
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
80 sys.exit(3)
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
81 if args.method == "selected":
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
82 extract_pop(args.input_file, populations, args.output_file)
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
83 else:
4f28ee74079b "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/extract_pop commit 117c0e4a5c00dfd2e190359badf77d5643fccefa"
azomics
parents:
diff changeset
84 remove_pop(args.input_file, populations, args.output_file)