Mercurial > repos > guerler > dbkit

annotate dbkit_extract.py @ 2:81c7d4668a7e draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
author guerler
date Wed, 16 Dec 2020 12:02:48 +0000
parents
children 03e124ff7e26
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
2
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
1 #! /usr/bin/env python3
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
2 import argparse
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
3 from os import system
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
4 from os.path import getsize
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
5
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
6 from dbkit_package.DBKit import DBKit
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
7
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
8
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
9 def main(args):
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
10 logFile = open(args.log, "w")
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
11 outIndex = args.outindex
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
12 outData = args.outdata
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
13 entries = list()
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
14 with open(args.list, "r") as f:
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
15 for line in f:
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
16 name = line.split()[0]
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
17 entries.append(name)
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
18 logFile.write("Detected %s entries.\n" % len(entries))
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
19 tempFile = "temp.dat"
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
20 count = 0
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
21 dbkit = DBKit(args.index, args.database)
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
22 for entry in sorted(entries):
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
23 success = dbkit.createFile(entry, tempFile)
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
24 if success:
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
25 currentSize = getsize(outData)
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
26 entrySize = getsize(tempFile)
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
27 system("cat %s >> %s" % (tempFile, outData))
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
28 system("echo '%s\t%s\t%s' >> %s" % (entry, currentSize, entrySize, outIndex))
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
29 count = count + 1
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
30 else:
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
31 logFile.write("Entry %s not found.\n" % entry)
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
32 logFile.write("Extracted %s entries.\n" % count)
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
33 logFile.close()
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
34
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
35
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
36 if __name__ == "__main__":
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
37 parser = argparse.ArgumentParser(description='DBKit - Merge database pair.')
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
38 parser.add_argument('-l', '--list', help='List of entries to be extracted', required=True)
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
39 parser.add_argument('-i', '--index', help='Database Index file (ffindex)', required=True)
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
40 parser.add_argument('-d', '--database', help='Database Data file (ffdata)', required=True)
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
41 parser.add_argument('-oi', '--outindex', help='Output Index file', required=True)
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
42 parser.add_argument('-od', '--outdata', help='Output Data file', required=True)
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
43 parser.add_argument('-g', '--log', help='Log file', required=True)
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
44 args = parser.parse_args()
81c7d4668a7e "planemo upload commit bd03b7888eab0b010acfc3affd38bf4d4e2bb1ef-dirty"
guerler
parents:
diff changeset
45 main(args)