i18n/posplit
author Gregory Szorc <gregory.szorc@gmail.com>
Tue, 22 Apr 2014 13:16:34 -0700
changeset 21521 855f092c96e5
parent 20363 e3ee7ec85a15
child 28074 a1924bc6e267
permissions -rwxr-xr-x
run-tests: move diff generation into TestResult TestResult is the thing that captures all our test results. It's logical for diff viewing to be handled there and not inside Test. While writing this patch, it was discovered that Test.fail() was printing redundant test result output. This has been removed. Arguments controlling diffs have been removed from Test.__init__.
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
11389
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
     1
#!/usr/bin/env python
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
     2
#
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
     3
# posplit - split messages in paragraphs on .po/.pot files
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
     4
#
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
     5
# license: MIT/X11/Expat
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
     6
#
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
     7
20359
ff6ab0b2ebf7 i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents: 11389
diff changeset
     8
import re
11389
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
     9
import sys
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    10
import polib
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    11
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    12
def addentry(po, entry, cache):
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    13
    e = cache.get(entry.msgid)
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    14
    if e:
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    15
        e.occurrences.extend(entry.occurrences)
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    16
    else:
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    17
        po.append(entry)
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    18
        cache[entry.msgid] = entry
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    19
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    20
def mkentry(orig, delta, msgid, msgstr):
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    21
    entry = polib.POEntry()
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    22
    entry.merge(orig)
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    23
    entry.msgid = msgid or orig.msgid
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    24
    entry.msgstr = msgstr or orig.msgstr
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    25
    entry.occurrences = [(p, int(l) + delta) for (p, l) in orig.occurrences]
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    26
    return entry
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    27
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    28
if __name__ == "__main__":
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    29
    po = polib.pofile(sys.argv[1])
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    30
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    31
    cache = {}
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    32
    entries = po[:]
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    33
    po[:] = []
20359
ff6ab0b2ebf7 i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents: 11389
diff changeset
    34
    findd = re.compile(r' *\.\. (\w+)::') # for finding directives
11389
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    35
    for entry in entries:
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    36
        msgids = entry.msgid.split(u'\n\n')
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    37
        if entry.msgstr:
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    38
            msgstrs = entry.msgstr.split(u'\n\n')
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    39
        else:
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    40
            msgstrs = [u''] * len(msgids)
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    41
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    42
        if len(msgids) != len(msgstrs):
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    43
            # places the whole existing translation as a fuzzy
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    44
            # translation for each paragraph, to give the
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    45
            # translator a chance to recover part of the old
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    46
            # translation - erasing extra paragraphs is
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    47
            # probably better than retranslating all from start
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    48
            if 'fuzzy' not in entry.flags:
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    49
                entry.flags.append('fuzzy')
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    50
            msgstrs = [entry.msgstr] * len(msgids)
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    51
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    52
        delta = 0
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    53
        for msgid, msgstr in zip(msgids, msgstrs):
20361
3fe079d3a2b4 i18n: posplit removes the entry "::" from the pot file
Simon Heimberg <simohe@besonet.ch>
parents: 20359
diff changeset
    54
            if msgid and msgid != '::':
11389
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    55
                newentry = mkentry(entry, delta, msgid, msgstr)
20359
ff6ab0b2ebf7 i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents: 11389
diff changeset
    56
                mdirective = findd.match(msgid)
ff6ab0b2ebf7 i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents: 11389
diff changeset
    57
                if mdirective:
20362
1bce1078501d i18n: leave out entries which contain only a rst directive
Simon Heimberg <simohe@besonet.ch>
parents: 20361
diff changeset
    58
                    if not msgid[mdirective.end():].rstrip():
1bce1078501d i18n: leave out entries which contain only a rst directive
Simon Heimberg <simohe@besonet.ch>
parents: 20361
diff changeset
    59
                        # only directive, nothing to translate here
1bce1078501d i18n: leave out entries which contain only a rst directive
Simon Heimberg <simohe@besonet.ch>
parents: 20361
diff changeset
    60
                        continue
20359
ff6ab0b2ebf7 i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents: 11389
diff changeset
    61
                    directive = mdirective.group(1)
20363
e3ee7ec85a15 i18n: leave out entries which contain only rst syntax
Simon Heimberg <simohe@besonet.ch>
parents: 20362
diff changeset
    62
                    if directive in ('container', 'include'):
e3ee7ec85a15 i18n: leave out entries which contain only rst syntax
Simon Heimberg <simohe@besonet.ch>
parents: 20362
diff changeset
    63
                        if msgid.rstrip('\n').count('\n') == 0:
e3ee7ec85a15 i18n: leave out entries which contain only rst syntax
Simon Heimberg <simohe@besonet.ch>
parents: 20362
diff changeset
    64
                            # only rst syntax, nothing to translate
e3ee7ec85a15 i18n: leave out entries which contain only rst syntax
Simon Heimberg <simohe@besonet.ch>
parents: 20362
diff changeset
    65
                            continue
e3ee7ec85a15 i18n: leave out entries which contain only rst syntax
Simon Heimberg <simohe@besonet.ch>
parents: 20362
diff changeset
    66
                        else:
e3ee7ec85a15 i18n: leave out entries which contain only rst syntax
Simon Heimberg <simohe@besonet.ch>
parents: 20362
diff changeset
    67
                            # lines following directly, unexpected
e3ee7ec85a15 i18n: leave out entries which contain only rst syntax
Simon Heimberg <simohe@besonet.ch>
parents: 20362
diff changeset
    68
                            print 'Warning: text follows line with directive' \
e3ee7ec85a15 i18n: leave out entries which contain only rst syntax
Simon Heimberg <simohe@besonet.ch>
parents: 20362
diff changeset
    69
                                  ' %s' % directive
20359
ff6ab0b2ebf7 i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents: 11389
diff changeset
    70
                    comment = 'do not translate: .. %s::' % directive
ff6ab0b2ebf7 i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents: 11389
diff changeset
    71
                    if not newentry.comment:
ff6ab0b2ebf7 i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents: 11389
diff changeset
    72
                        newentry.comment = comment
ff6ab0b2ebf7 i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents: 11389
diff changeset
    73
                    elif comment not in newentry.comment:
ff6ab0b2ebf7 i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents: 11389
diff changeset
    74
                        newentry.comment += '\n' + comment
11389
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    75
                addentry(po, newentry, cache)
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    76
            delta += 2 + msgid.count('\n')
4fd49329a1b5 i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff changeset
    77
    po.save()