Mercurial > repos > immport-devteam > auto_collapse_pop

annotate auto_collapse_pops.py @ 1:08b71aee9b80 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/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
author azomics
date Mon, 22 Jun 2020 17:22:03 -0400
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
1
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
1 #!/usr/bin/env python
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
2
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
3 ######################################################################
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
4 # Copyright (c) 2016 Northrop Grumman.
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
5 # All rights reserved.
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
6 ######################################################################
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
7 from __future__ import print_function
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
8 import sys
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
9 import pandas as pd
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
10 from argparse import ArgumentParser
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
11
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
12
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
13 def auto_collapse(input_file, profile_file, output, report):
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
14 profile_pop_list = {}
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
15 pop_to_collapse = []
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
16 markers = []
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
17 with open(profile_file, "r") as pf:
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
18 pffl = pf.readline()
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
19 markers = pffl.strip().split("\t")
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
20 for pfline in pf:
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
21 line = pfline.strip().split("\t")
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
22 pop = line[0]
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
23 profil = "\t".join(line[1:-2])
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
24 if profil in profile_pop_list:
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
25 profile_pop_list[profil].append(pop)
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
26 else:
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
27 profile_pop_list[profil] = [pop]
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
28 i = 1
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
29 with open(report, "w") as rt:
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
30 rt.write("New_Population\tFormer_Populations\t")
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
31 rt.write("\t".join(markers[1:-2]) + "\n")
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
32 for profs in profile_pop_list:
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
33 pop_to_collapse.append(profile_pop_list[profs])
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
34 pop_ls = ", ".join(profile_pop_list[profs])
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
35 rt.write("\t".join([str(i), pop_ls, profs]) + "\n")
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
36 i += 1
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
37 df = pd.read_table(input_file, dtype={'Population': object})
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
38 df['new_population'] = df.Population
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
39 for i, sets_pop in enumerate(pop_to_collapse):
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
40 df.loc[df['Population'].isin(sets_pop), ['new_population']] = i + 1
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
41
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
42 df.Population = df.new_population
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
43 df.drop(['new_population'], inplace=True, axis=1)
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
44 df.to_csv(output, sep="\t", index=False)
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
45
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
46
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
47 if __name__ == "__main__":
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
48 parser = ArgumentParser(
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
49 prog="auto_pop_collapse_from_profile",
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
50 description="collapses FLOCK populations based on profile.")
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
51
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
52 parser.add_argument(
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
53 '-i',
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
54 dest="input_file",
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
55 required=True,
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
56 help="FLOCK output file")
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
57
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
58 parser.add_argument(
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
59 '-o',
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
60 dest="output",
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
61 required=True,
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
62 help="Name of the output file.")
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
63
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
64 parser.add_argument(
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
65 '-r',
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
66 dest="report",
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
67 required=True,
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
68 help="Name of the report file.")
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
69
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
70 parser.add_argument(
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
71 '-p',
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
72 dest="profile",
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
73 required=True,
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
74 help="File location for the profile.txt from FLOCK.")
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
75
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
76 args = parser.parse_args()
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
77
08b71aee9b80 "planemo upload for repository https://github.com/ImmPortDB/immport-galaxy-tools/tree/master/flowtools/auto_collapse_pop commit c4cd54a5b7a57f494a82d2a5e9b94f24473fff24"
azomics
parents:
diff changeset
78 auto_collapse(args.input_file, args.profile, args.output, args.report)