Mercurial > repos > galaxyp > fragpipe

annotate generate_manifest.py @ 8:012191b79fda 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/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
author galaxyp
date Fri, 17 Oct 2025 16:22:03 +0000
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
8
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
1 #!/usr/bin/env python3
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
2
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
3 #
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
4 # Generates a FragPipe Manifest file.
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
5 #
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
6
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
7 import argparse
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
8 import csv
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
9
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
10 # The three columns for each scanfile are "Experiment, Bioreplicate, and Data type
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
11 column_types = ('exp', 'bio', 'type')
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
12 output_filename = 'fp.manifest'
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
13
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
14
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
15 # Add column values to a list of rows for each scan file.
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
16 def add_column(column_type, args, rows):
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
17 nfiles = len(args.scanfiles)
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
18
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
19 # Each scan file is numbered 1 through n in column
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
20 if getattr(args, f'{column_type}_consec'):
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
21 vals = range(1, nfiles + 1)
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
22
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
23 # All scan files have same value in column
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
24 elif getattr(args, f'{column_type}_assign_all'):
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
25 vals = [getattr(args, f'{column_type}_assign_all')] * nfiles
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
26
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
27 # Values are provided for scan files in a comma-delimited list
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
28 elif getattr(args, f'{column_type}_col'):
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
29 vals = getattr(args, f'{column_type}_col').split(',')
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
30 if len(vals) != nfiles:
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
31 raise ValueError((f'Incorrect number of values entered for column {column_type}. '
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
32 'Exactly one value must be entered for each scan file.'))
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
33
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
34 # Otherwise, this column remains empty.
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
35 else:
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
36 vals = [''] * nfiles
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
37
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
38 for i, row in enumerate(rows):
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
39 row.append(vals[i])
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
40
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
41
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
42 def main():
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
43 parser = argparse.ArgumentParser()
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
44
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
45 # Each column has the same methods for populating
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
46 for column_type in column_types:
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
47 parser.add_argument(f'--{column_type}-consec', action='store_true')
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
48 parser.add_argument(f'--{column_type}-assign-all')
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
49 parser.add_argument(f'--{column_type}-col')
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
50
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
51 # Scanfile names, which should be identical to history identifiers
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
52 parser.add_argument('scanfiles', nargs='+')
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
53
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
54 args = parser.parse_args()
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
55
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
56 # Create and populate data structure for tabular output
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
57 rows = [[scanfile] for scanfile in args.scanfiles]
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
58 for column_type in column_types:
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
59 add_column(column_type, args, rows)
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
60
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
61 # Write out manifest file.
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
62 # Use mode=a as the script will be called once for each scan group.
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
63 with open(output_filename, mode='a') as outf:
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
64 manifest_writer = csv.writer(outf, delimiter='\t')
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
65 for row in rows:
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
66 manifest_writer.writerow(row)
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
67
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
68
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
69 if __name__ == "__main__":
012191b79fda planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/fragpipe commit 6413a461059c4a421a7812a08f244c224cde8ee2
galaxyp
parents:
diff changeset
70 main()