Mercurial > repos > iuc > ncbi_eutils_esearch

annotate esearch.py @ 5:79bc38fca3d9 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 dd9b889d83f058d2ddb87a4e7709ee3d93dddd69"
author iuc
date Fri, 27 Aug 2021 11:28:25 +0000
parents e267701c187b
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
1 #!/usr/bin/env python
1
30150bd36c9a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit a4b0969b33a68a0ea9ba12291f6694aec24f13ed
iuc
parents: 0
diff changeset
2
30150bd36c9a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit a4b0969b33a68a0ea9ba12291f6694aec24f13ed
iuc
parents: 0
diff changeset
3 import argparse
0
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
4 import json
3
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
5 import logging
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
6
1
30150bd36c9a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit a4b0969b33a68a0ea9ba12291f6694aec24f13ed
iuc
parents: 0
diff changeset
7
0
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
8 import eutils
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
9
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
10
3
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
11 logging.basicConfig(level=logging.INFO)
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
12
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
13
0
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
14 if __name__ == '__main__':
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
15 parser = argparse.ArgumentParser(description='ESearch', epilog='')
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
16 parser.add_argument('db', help='Database to use')
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
17 parser.add_argument('term', help='Query')
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
18 parser.add_argument('--history_file', help='Filter existing history')
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
19 parser.add_argument('--datetype', help='Date type')
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
20 parser.add_argument('--reldate', help='In past N days')
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
21 parser.add_argument('--mindate', help='Minimum date')
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
22 parser.add_argument('--maxdate', help='maximum date')
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
23 # History
3
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
24 parser.add_argument('--history_out', action="store_true", help='Output history file')
0
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
25 parser.add_argument('--user_email', help="User email")
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
26 parser.add_argument('--admin_email', help="Admin email")
3
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
27
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
28 parser.add_argument('--version', action='version', version=eutils.Client.getVersion(), help='Version (reports Biopython version)')
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
29
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
30 # Output
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
31 parser.add_argument('--retmode', help='Retmode')
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
32 parser.add_argument('--rettype', help='Rettype')
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
33 parser.add_argument('--retstart', type=int, default=0, help='Retstart - Starting rec number (0)')
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
34 parser.add_argument('--retmax', type=int, default=20, help='Retmax - max number of recs returned (20, max 100000)')
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
35
0
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
36 args = parser.parse_args()
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
37
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
38 c = eutils.Client(history_file=args.history_file, user_email=args.user_email, admin_email=args.admin_email)
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
39
3
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
40 max_retmax = 100000
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
41 min_retmax = 1
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
42 max = max(min(args.retmax, max_retmax), min_retmax)
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
43
0
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
44 payload = {
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
45 'db': args.db,
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
46 'term': args.term,
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
47 }
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
48 if args.history_file is not None:
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
49 payload.update(c.get_history())
3
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
50
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
51 # if args.history_out is not None:
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
52 if args.history_out:
0
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
53 payload['usehistory'] = 'y'
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
54
3
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
55 payload['retmode'] = args.retmode
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
56
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
57 for attr in ('datetype', 'reldate', 'mindate', 'maxdate', 'rettype', 'retmax', 'retstart'):
0
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
58 if getattr(args, attr, None) is not None:
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
59 payload[attr] = getattr(args, attr)
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
60
3
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
61 logging.info("Payload used for query:" + json.dumps(payload, indent=4))
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
62
0
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
63 results = c.search(**payload)
c8a84e910e6b planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 15bcc5104c577b4b9c761f2854fc686c07ffa9db
iuc
parents:
diff changeset
64
3
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
65 # We're going to infer that rettype being uilist means convert to text format (which esearch does not do)
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
66 if args.retmode == 'text':
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
67 ids = c.xmlstring2UIlist(results)
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
68 for id in ids:
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
69 print(id)
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
70 elif args.retmode == 'json':
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
71 json_data = c.jsonstring2jsondata(results)
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
72 print(json.dumps(json_data, indent=4))
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
73 else:
e267701c187b "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 1
diff changeset
74 print(results)