Mercurial > repos > earlhaminst > hcluster_sg_parser

annotate hcluster_sg_parser.py @ 1:17aa68582a05 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
author earlhaminst
date Fri, 20 Jan 2017 06:13:23 -0500
parents
children f9e418125021
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
1
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
1 """
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
2 A simple parser to convert the hcluster_sg 3-column output into lists of IDs, one list for each cluster.
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
3
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
4 When a minimum and/or maximum number of cluster elements are specified, the IDs contained in the filtered-out clusters are collected in the "discarded IDS" output dataset.
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
5
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
6 Usage:
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
7
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
8 python hcluster_sg_parser.py [-m <N>] [-M <N>] <file> <discarded_out>
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
9 """
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
10 import optparse
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
11 import sys
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
12
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
13
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
14 def main():
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
15 parser = optparse.OptionParser()
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
16 parser.add_option('-m', '--min', type='int', default=0, help='Minimum number of cluster elements')
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
17 parser.add_option('-M', '--max', type='int', default=sys.maxsize, help='Maximum number of cluster elements')
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
18 options, args = parser.parse_args()
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
19
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
20 with open(args[1], 'w') as discarded_out:
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
21 with open(args[0]) as fh:
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
22 for line in fh:
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
23 line = line.rstrip()
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
24 (cluster_id, n_ids, id_list) = line.split('\t')
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
25 n_ids = int(n_ids)
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
26 id_list = id_list.replace(',', '\n')
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
27 if n_ids >= options.min and n_ids <= options.max:
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
28 outfile = cluster_id + '_output.txt'
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
29 with open(outfile, 'w') as f:
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
30 f.write(id_list)
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
31 else:
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
32 discarded_out.write(id_list)
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
33
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
34
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
35 if __name__ == "__main__":
17aa68582a05 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/hcluster_sg_parser commit a79c8f1799189754eae80aede6fbe5428570f36b
earlhaminst
parents:
diff changeset
36 main()