Mercurial > repos > guerler > dbkit

comparison dbkit_create.py @ 0:1914107cc967 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload commit 3632646bec5edbe47e06c894e32bfd215b895555-dirty"
author guerler
date Wed, 25 Nov 2020 17:22:48 +0000
parents
children 987e55ea29b8
comparison
equal deleted inserted replaced
-1:000000000000 0:1914107cc967
1 #! /usr/bin/env python3
2 import argparse
3 from os import system
4 from os.path import getsize
5
6
7 def getIdentifiers(args):
8 entries = set()
9 with open(args.list) as file:
10 for line in file:
11 entry = line.split()[0]
12 idLength = int(args.idlength)
13 if idLength > 0:
14 entry = entry[:idLength]
15 if args.idcase == "lower":
16 entry = entry.lower()
17 elif args.idcase == "upper":
18 entry = entry.upper()
19 if args.idextension is not None:
20 entry = "%s.%s" % (entry, args.idextension)
21 entries.add(entry)
22 return sorted(entries)
23
24
25 def main(args):
26 entries = getIdentifiers(args)
27 logFile = open(args.log, "w")
28 logFile.write("Found %s entries.\n" % len(entries))
29 outputIndex = args.index
30 outputDatabase = args.database
31 tempPath = args.temp.rstrip("/")
32 tempFile = "%s/temp.pdb" % tempPath
33 system("mkdir -p %s" % tempPath)
34 system("rm -f %s" % outputDatabase)
35 indexFile = open(outputIndex, 'w')
36 start = 0
37 for entryId in entries:
38 logFile.write("Loading %s.\n" % entryId)
39 system("wget -q -O %s %s%s" % (tempFile, args.url, entryId))
40 tempSize = getsize(tempFile)
41 if tempSize == 0:
42 logFile.write("Entry `%s` not found.\n" % entryId)
43 else:
44 indexFile.write("%s\t%d\t%d\n" % (entryId, start, tempSize))
45 start = start + tempSize + 1
46 system("cat %s >> %s" % (tempFile, outputDatabase))
47 logFile.flush()
48 logFile.close()
49
50
51 if __name__ == "__main__":
52 parser = argparse.ArgumentParser(description='DBKit - Download and Merge files into a single file.')
53 parser.add_argument('-l', '--list', help='List of entries', required=True)
54 parser.add_argument('-u', '--url', help='Source Url', required=True)
55 parser.add_argument('-t', '--temp', help='temp', required=True)
56 parser.add_argument('-il', '--idlength', help='Format Identifier Length (integer)', required=False, default="0")
57 parser.add_argument('-ic', '--idcase', help='Format Identifier Case (lower, upper)', required=False, default=None)
58 parser.add_argument('-ie', '--idextension', help='Format Identifier Extension', required=False, default=None)
59 parser.add_argument('-o', '--index', help='Output Database Index', required=True)
60 parser.add_argument('-d', '--database', help='Output Database', required=True)
61 parser.add_argument('-g', '--log', help="Log file", required=True)
62 args = parser.parse_args()
63 main(args)