Mercurial > repos > petr-novak > various_galaxy_tools

annotate extract_GFF_Features.py @ 0:696e702ebf74 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
author petr-novak
date Mon, 09 May 2022 08:26:30 +0000
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
1 #!/usr/bin/env python
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
2 # Guruprasad Ananda
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
3 """
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
4 Extract features from GFF file.
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
5
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
6 usage: %prog input1 out_file1 column features
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
7 """
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
8 from __future__ import print_function
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
9
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
10 import sys
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
11
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
12 from bx.cookbook import doc_optparse
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
13
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
14
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
15 def stop_err(msg):
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
16 sys.stderr.write(msg)
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
17 sys.exit()
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
18
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
19
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
20 def main():
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
21 # Parsing Command Line here
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
22 options, args = doc_optparse.parse(__doc__)
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
23
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
24 try:
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
25 inp_file, out_file, column, features = args
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
26 except ValueError:
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
27 stop_err("One or more arguments is missing or invalid.\nUsage: prog input output column features")
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
28 try:
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
29 column = int(column)
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
30 except ValueError:
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
31 stop_err("Column %s is an invalid column." % column)
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
32
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
33 if features is None:
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
34 stop_err("Column %d has no features to display, select another column." % (column + 1))
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
35
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
36 fo = open(out_file, "w")
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
37 for line in open(inp_file):
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
38 line = line.rstrip("\r\n")
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
39 if line and line.startswith("#"):
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
40 # Keep valid comment lines in the output
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
41 fo.write("%s\n" % line)
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
42 else:
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
43 try:
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
44 if line.split("\t")[column] in features.split(","):
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
45 fo.write("%s\n" % line)
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
46 except Exception:
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
47 pass
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
48 fo.close()
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
49
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
50 print("Column %d features: %s" % (column + 1, features))
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
51
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
52
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
53 if __name__ == "__main__":
696e702ebf74 "planemo upload commit 0f6eca49bafc3c946189d793161a7f81d595e1a1-dirty"
petr-novak
parents:
diff changeset
54 main()