Mercurial > repos > iuc > ncbi_eutils_efetch

annotate esearch.py @ 15:cb4d6f20aacc 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/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 851b88d92fa7740880a5edf407f629bf3f7f7bf0
author iuc
date Fri, 04 Oct 2024 14:33:32 +0000
parents b151cd22921f
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
1b4ac594d02a 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
6
2ff5369b4b51 planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit a4b0969b33a68a0ea9ba12291f6694aec24f13ed
iuc
parents: 0
diff changeset
2
2ff5369b4b51 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
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
4 import json
8
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
5 import logging
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
6
6
2ff5369b4b51 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
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
8 import eutils
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
9
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
10
8
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
11 logging.basicConfig(level=logging.INFO)
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
12
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
13
0
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
14 if __name__ == '__main__':
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
15 parser = argparse.ArgumentParser(description='ESearch', epilog='')
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
16 parser.add_argument('db', help='Database to use')
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
17 parser.add_argument('term', help='Query')
1b4ac594d02a 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('--history_file', help='Filter existing history')
1b4ac594d02a 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('--datetype', help='Date type')
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
20 parser.add_argument('--reldate', help='In past N days')
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
21 parser.add_argument('--mindate', help='Minimum date')
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
22 parser.add_argument('--maxdate', help='maximum date')
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
23 # History
8
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
24 parser.add_argument('--history_out', action="store_true", help='Output history file')
0
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
25 parser.add_argument('--user_email', help="User email")
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
26 parser.add_argument('--admin_email', help="Admin email")
8
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
27
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
28 parser.add_argument('--version', action='version', version=eutils.Client.getVersion(), help='Version (reports Biopython version)')
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
29
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
30 # Output
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
31 parser.add_argument('--retmode', help='Retmode')
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
32 parser.add_argument('--rettype', help='Rettype')
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
33 parser.add_argument('--retstart', type=int, default=0, help='Retstart - Starting rec number (0)')
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
34 parser.add_argument('--retmax', type=int, default=20, help='Retmax - max number of recs returned (20, max 100000)')
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
35
0
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
36 args = parser.parse_args()
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
37
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
38 c = eutils.Client(history_file=args.history_file, user_email=args.user_email, admin_email=args.admin_email)
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
39
8
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
40 max_retmax = 100000
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
41 min_retmax = 1
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
42 max = max(min(args.retmax, max_retmax), min_retmax)
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
43
0
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
44 payload = {
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
45 'db': args.db,
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
46 'term': args.term,
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
47 }
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
48 if args.history_file is not None:
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
49 payload.update(c.get_history())
8
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
50
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
51 # if args.history_out is not None:
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
52 if args.history_out:
0
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
53 payload['usehistory'] = 'y'
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
54
8
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
55 payload['retmode'] = args.retmode
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
56
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
57 for attr in ('datetype', 'reldate', 'mindate', 'maxdate', 'rettype', 'retmax', 'retstart'):
0
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
58 if getattr(args, attr, None) is not None:
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
59 payload[attr] = getattr(args, attr)
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
60
8
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
61 logging.info("Payload used for query:" + json.dumps(payload, indent=4))
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
62
0
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
63 results = c.search(**payload)
1b4ac594d02a planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit 780c9984a9c44d046aadf1e316a668d1e53aa1f0
iuc
parents:
diff changeset
64
8
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
65 # We're going to infer that rettype being uilist means convert to text format (which esearch does not do)
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
66 if args.retmode == 'text':
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
67 ids = c.xmlstring2UIlist(results)
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
68 for id in ids:
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
69 print(id)
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
70 elif args.retmode == 'json':
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
71 json_data = c.jsonstring2jsondata(results)
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
72 print(json.dumps(json_data, indent=4))
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
73 else:
b151cd22921f "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_entrez_eutils commit dae34e5e182b4cceb808d7353080f14aa9a78ca9"
iuc
parents: 6
diff changeset
74 print(results)