Mercurial > repos > iuc > ncbi_eutils_egquery

annotate ecitmatch.py @ 5:20a86bfb54eb draft

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-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
author iuc
date Thu, 07 Jul 2016 02:38:00 -0400
parents 22fed2340d2b
children 2d54a753d1f1
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
1 #!/usr/bin/env python
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
2 import argparse
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
3 import eutils
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
4
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
5
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
6 if __name__ == '__main__':
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
7 parser = argparse.ArgumentParser(description='ECitMatch', epilog='')
5
20a86bfb54eb planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents: 0
diff changeset
8 parser.add_argument('--file', type=argparse.FileType('r'), help='Tabular file containing citations to search')
0
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
9
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
10 parser.add_argument('--key', nargs='*', help='Citation Key')
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
11 parser.add_argument('--journal_title', nargs='*', help='Journal Title')
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
12 parser.add_argument('--year', nargs='*', help='Year')
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
13 parser.add_argument('--volume', nargs='*', help='Volume')
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
14 parser.add_argument('--first_page', nargs='*', help='First Page')
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
15 parser.add_argument('--author_name', nargs='*', help='Author name')
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
16
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
17 # Emails
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
18 parser.add_argument('--user_email', help="User email")
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
19 parser.add_argument('--admin_email', help="Admin email")
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
20 args = parser.parse_args()
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
21
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
22 c = eutils.Client(user_email=args.user_email, admin_email=args.admin_email)
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
23
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
24 citations = []
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
25 if args.file is None:
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
26 for key, journal, year, volume, first_page, author_name in \
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
27 zip(args.key, args.journal_title, args.year, args.volume, args.first_page, args.author_name):
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
28 citations.append({
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
29 'key': key,
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
30 'journal': journal,
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
31 'year': year,
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
32 'volume': volume,
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
33 'first_page': first_page,
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
34 'author_name': author_name,
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
35 })
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
36 else:
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
37 for line in args.file:
5
20a86bfb54eb planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents: 0
diff changeset
38 line = line.strip()
0
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
39 if not line.startswith('#'):
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
40 tmp = line.split('\t')
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
41 try:
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
42 citations.append({
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
43 'journal': tmp[0],
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
44 'year': tmp[1],
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
45 'volume': tmp[2],
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
46 'first_page': tmp[3],
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
47 'author_name': tmp[4],
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
48 'key': tmp[5],
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
49 })
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
50 except KeyError:
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
51 print "Could not parse line: %s" % line
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
52
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
53 payload = {
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
54 'db': 'pubmed',
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
55 'bdata': citations
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
56 }
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
57
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
58 results = c.citmatch(**payload)
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
59 # We get data back as pipe separated, so just replace those with tabs
22fed2340d2b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
60 print results.replace('|', '\t')