Mercurial > repos > devteam > fasta_to_tabular

annotate fasta_to_tabular.py @ 3:5b998d13de71 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/fasta_to_tabular commit 34a6c9f94a5722bb7d2f887618aafa410a770e91"
author devteam
date Mon, 02 Mar 2020 11:46:30 +0000
parents ff4751ce764d
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
1 #!/usr/bin/env python
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
2 # This code exists in 2 places: ~/datatypes/converters and ~/tools/fasta_tools
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
3 """
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
4 Input: fasta (input file), tabular (output file), int (truncation of id), int (columns from description)
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
5 Output: tabular
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
6 format convert: fasta to tabular
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
7 """
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
8
2
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
9 import sys
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
10
0
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
11
2
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
12 def stop_err(msg):
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
13 sys.exit(msg)
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
14
0
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
15
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
16 def __main__():
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
17 if len(sys.argv) != 5:
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
18 stop_err("Wrong number of argument. Expect four (fasta, tabular, truncation, columns)")
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
19 infile = sys.argv[1]
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
20 outfile = sys.argv[2]
2
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
21 keep_first = int(sys.argv[3])
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
22 descr_split = int(sys.argv[4])
0
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
23 if keep_first == 0:
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
24 keep_first = None
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
25 elif descr_split == 1:
2
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
26 # Added one for the ">" character
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
27 # (which is removed if using descr_split > 1)
0
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
28 keep_first += 1
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
29 if descr_split < 1:
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
30 stop_err("Bad description split value (should be 1 or more)")
2
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
31 with open(outfile, 'w') as out, open(infile) as in_fh:
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
32 for i, line in enumerate(in_fh):
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
33 line = line.rstrip('\r\n')
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
34 if not line or line.startswith('#'):
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
35 continue
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
36 if line.startswith('>'):
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
37 # Don't want any existing tabs to trigger extra columns:
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
38 line = line.replace('\t', ' ')
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
39 if i > 0:
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
40 out.write('\n')
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
41 if descr_split == 1:
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
42 out.write(line[1:keep_first])
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
43 else:
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
44 words = line[1:].split(None, descr_split - 1)
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
45 # apply any truncation to first word (the id)
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
46 words[0] = words[0][0:keep_first]
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
47 # pad with empty columns if required
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
48 words += [""] * (descr_split - len(words))
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
49 out.write("\t".join(words))
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
50 out.write('\t')
0
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
51 else:
2
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
52 out.write(line)
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
53 if i > 0:
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
54 out.write('\n')
0
ae709fd50581 Imported from capsule None
devteam
parents:
diff changeset
55
2
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
56
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
57 if __name__ == "__main__":
ff4751ce764d "planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/tools/fasta_to_tabular commit cd1ed08574b749eee2a3f6e6151dbb0c8ca15bbf"
devteam
parents: 0
diff changeset
58 __main__()