Mercurial > repos > devteam > tabular_to_fasta

annotate tabular_to_fasta.py @ 1:7f7a1bea4653 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/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
author devteam
date Wed, 05 Feb 2020 15:51:26 +0000
parents db61adc12770
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
1 #!/usr/bin/env python
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
2 """
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
3 Input: fasta, minimal length, maximal length
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
4 Output: fasta
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
5 Return sequences whose lengths are within the range.
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
6 """
1
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
7 import os
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
8 import sys
0
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
9
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
10
1
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
11 def stop_err(msg):
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
12 sys.exit(msg)
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
13
0
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
14
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
15 def __main__():
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
16 infile = sys.argv[1]
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
17 title_col = sys.argv[2]
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
18 seq_col = sys.argv[3]
1
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
19 outfile = sys.argv[4]
0
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
20
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
21 if title_col == None or title_col == 'None' or seq_col == None or seq_col == 'None':
1
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
22 stop_err("Columns not specified.")
0
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
23 try:
1
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
24 seq_col = int(seq_col) - 1
0
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
25 except:
1
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
26 stop_err("Invalid Sequence Column: %s." % str(seq_col))
0
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
27
1
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
28 title_col_list = title_col.split(',')
0
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
29 skipped_lines = 0
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
30 first_invalid_line = 0
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
31 invalid_line = ""
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
32 i = 0
1
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
33
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
34 with open(outfile, 'w') as out:
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
35 for i, line in enumerate(open(infile)):
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
36 error = False
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
37 line = line.rstrip('\r\n')
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
38 if line and not line.startswith('#'):
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
39 fields = line.split('\t')
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
40 fasta_title = []
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
41 for j in title_col_list:
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
42 try:
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
43 j = int(j) - 1
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
44 fasta_title.append(fields[j])
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
45 except:
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
46 skipped_lines += 1
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
47 if not invalid_line:
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
48 first_invalid_line = i + 1
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
49 invalid_line = line
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
50 error = True
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
51 break
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
52 if not error:
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
53 try:
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
54 fasta_seq = fields[seq_col]
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
55 if fasta_title[0].startswith(">"):
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
56 fasta_title[0] = fasta_title[0][1:]
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
57 print(">%s\n%s" % ("_".join(fasta_title), fasta_seq), file=out)
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
58 except:
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
59 skipped_lines += 1
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
60 if not invalid_line:
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
61 first_invalid_line = i + 1
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
62 invalid_line = line
0
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
63
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
64 if skipped_lines > 0:
1
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
65 print('Data issue: skipped %d blank or invalid lines starting at #%d: "%s"' % (skipped_lines, first_invalid_line, invalid_line))
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
66
0
db61adc12770 Imported from capsule None
devteam
parents:
diff changeset
67
1
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
68 if __name__ == "__main__":
7f7a1bea4653 "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/tabular_to_fasta commit 01140c0ac1a926856c55853a0028e5d44935d9e6"
devteam
parents: 0
diff changeset
69 __main__()