Mercurial > repos > sanbi-uwc > vcf2neo

annotate vcf2neo_wrapper.py @ 0:3e14eda348d3 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/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
author sanbi-uwc
date Mon, 19 Jun 2017 00:08:18 -0400
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
1 #!/usr/bin/env python
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
2
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
3 from __future__ import print_function
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
4 import argparse
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
5 import os
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
6 import shlex
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
7 import subprocess
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
8 import uuid
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
9 import sys
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
10
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
11 parser = argparse.ArgumentParser(
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
12 description="Call vcf2neo on VCF inputs from Galaxy")
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
13
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
14 parser.add_argument('--vcf_dataset_names', nargs='+',
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
15 help='Names of VCF datasets')
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
16 parser.add_argument('--neo4j_db_path', help='Neo4j database directory')
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
17 parser.add_argument('--user', help='Email of Galaxy user running this tool')
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
18 parser.add_argument('--variantset_name',
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
19 help='Name for the VariantSet containing all the variants')
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
20 parser.add_argument('--vcf_files',
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
21 help='VCF format variant file', nargs='+')
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
22
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
23 args = parser.parse_args()
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
24
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
25 os.mkdir(args.variantset_name)
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
26
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
27 print("VCF names:", len(args.vcf_dataset_names), args.vcf_dataset_names, file=sys.stderr)
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
28 print("VCF files:", len(args.vcf_files), args.vcf_files, file=sys.stderr)
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
29 for i, vcf_file in enumerate(args.vcf_files):
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
30 print("XXXX I:", i, vcf_file, file=sys.stderr)
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
31 callset_name = args.vcf_dataset_names[i]
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
32 os.symlink(vcf_file, os.path.join(args.variantset_name,
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
33 callset_name) + '.vcf')
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
34 # Usage: vcf2neo init [OPTIONS] VCF_DIR OWNER [HISTORY_ID] [OUTPUT_DIR]
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
35 #
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
36 # Copy reference database and load VCF to Neo4j Graph database. :param
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
37 # vcf_dir: :param refdb_dir: :param d: :return:
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
38 #
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
39 # Options:
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
40 # -d / -D Run Neo4j docker container.
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
41 # --help Show this message and exit.
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
42
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
43 history_id = str(uuid.uuid4())
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
44 cmd_str = ('vcf2neo init -d ' +
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
45 '{input_vcf_dir} {email} {history_id} {neo4j_db_path}'.format(
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
46 input_vcf_dir=args.variantset_name,
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
47 email=args.user,
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
48 history_id=history_id,
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
49 neo4j_db_path=args.neo4j_db_path))
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
50 cmd = shlex.split(cmd_str)
3e14eda348d3 planemo upload for repository https://github.com/sanbi-sa/tools-sanbi-uwc commit 0245b4efed8ced6d06d76f1249a47d2178285385
sanbi-uwc
parents:
diff changeset
51 subprocess.check_call(cmd)