Mercurial > repos > galaxyp > percolator

annotate nested_collection.py @ 1:23141085ca9e draft

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/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
author galaxyp
date Sat, 04 Mar 2017 20:35:44 -0500
parents
children b20ea66c01f0
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
1
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
1 import argparse
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
2 import os
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
3 import re
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
4 from collections import OrderedDict
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
5
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
6
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
7 def get_filename_index_with_identifier(realnames, pool_id):
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
8 pool_indices = []
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
9 for index, fn in enumerate(realnames):
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
10 if re.search(pool_id, fn) is not None:
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
11 pool_indices.append(index)
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
12 return pool_indices
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
13
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
14
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
15 def get_batches_of_galaxyfiles(realnames, batchsize, pool_ids):
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
16 """For an amount of input files, pool identifiers and a batch size,
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
17 return batches of files for a list of lists"""
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
18 if pool_ids:
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
19 filegroups = OrderedDict([(p_id, get_filename_index_with_identifier(
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
20 realnames, p_id)) for p_id in pool_ids])
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
21 else:
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
22 filegroups = {1: range(len(realnames))}
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
23 batch = []
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
24 for pool_id, grouped_indices in filegroups.items():
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
25 if pool_id == 1:
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
26 pool_id = 'pool0'
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
27 for index in grouped_indices:
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
28 batch.append(index)
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
29 if batchsize and len(batch) == int(batchsize):
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
30 yield pool_id, batch
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
31 batch = []
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
32 if len(batch) > 0:
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
33 yield pool_id, batch
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
34 batch = []
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
35
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
36
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
37 def main():
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
38 parser = argparse.ArgumentParser()
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
39 parser.add_argument('--batchsize', dest='batchsize', default=False)
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
40 parser.add_argument('--real-names', dest='realnames', nargs='+')
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
41 parser.add_argument('--galaxy-files', dest='galaxyfiles', nargs='+')
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
42 parser.add_argument('--pool-ids', dest='poolids', nargs='+', default=False)
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
43 args = parser.parse_args()
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
44 for batchcount, (pool_id, batch) in enumerate(get_batches_of_galaxyfiles(
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
45 args.realnames, args.batchsize, args.poolids)):
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
46 for fncount, batchfile in enumerate([args.galaxyfiles[index] for index in batch]):
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
47 dsetname = '{}___batch{}_inputfn{}.mzid'.format(pool_id, batchcount, fncount)
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
48 print('producing', dsetname)
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
49 os.symlink(batchfile, dsetname)
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
50
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
51 if __name__ == '__main__':
23141085ca9e planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tools/percolator commit 0a5f9eb82877545be1c924357e585b17e01cfd1c
galaxyp
parents:
diff changeset
52 main()