Mercurial > hg
annotate i18n/posplit @ 49249:db19f6be0442
debugindex: rename to debugindex debug-revlog-index
The command dump some content of the revlog index and omit a lot of
information. I am going to make it display the missing information.
For clarity, we rename the command to explicitly mention revlog.
author | Pierre-Yves David <pierre-yves.david@octobus.net> |
---|---|
date | Mon, 30 May 2022 11:30:48 +0200 |
parents | 6000f5b25c9b |
children | f4733654f144 |
rev | line source |
---|---|
45830
c102b704edb5
global: use python3 in shebangs
Gregory Szorc <gregory.szorc@gmail.com>
parents:
43691
diff
changeset
|
1 #!/usr/bin/env python3 |
11389
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 |
29152
c5057b7780dc
py3: make i18n/posplit use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents:
28074
diff
changeset
|
8 |
c5057b7780dc
py3: make i18n/posplit use absolute_import
Pulkit Goyal <7895pulkit@gmail.com>
parents:
28074
diff
changeset
|
9 import polib |
20359
ff6ab0b2ebf7
i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents:
11389
diff
changeset
|
10 import re |
11389
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
11 import sys |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
12 |
43691
47ef023d0165
black: blacken scripts
Gregory Szorc <gregory.szorc@gmail.com>
parents:
41759
diff
changeset
|
13 |
11389
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
14 def addentry(po, entry, cache): |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
15 e = cache.get(entry.msgid) |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
16 if e: |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
17 e.occurrences.extend(entry.occurrences) |
39269
d0e8933d6dad
i18n: merge i18n comments of translatable texts correctly
FUJIWARA Katsunori <foozy@lares.dti.ne.jp>
parents:
29153
diff
changeset
|
18 |
d0e8933d6dad
i18n: merge i18n comments of translatable texts correctly
FUJIWARA Katsunori <foozy@lares.dti.ne.jp>
parents:
29153
diff
changeset
|
19 # merge comments from entry |
d0e8933d6dad
i18n: merge i18n comments of translatable texts correctly
FUJIWARA Katsunori <foozy@lares.dti.ne.jp>
parents:
29153
diff
changeset
|
20 for comment in entry.comment.split('\n'): |
d0e8933d6dad
i18n: merge i18n comments of translatable texts correctly
FUJIWARA Katsunori <foozy@lares.dti.ne.jp>
parents:
29153
diff
changeset
|
21 if comment and comment not in e.comment: |
d0e8933d6dad
i18n: merge i18n comments of translatable texts correctly
FUJIWARA Katsunori <foozy@lares.dti.ne.jp>
parents:
29153
diff
changeset
|
22 if not e.comment: |
d0e8933d6dad
i18n: merge i18n comments of translatable texts correctly
FUJIWARA Katsunori <foozy@lares.dti.ne.jp>
parents:
29153
diff
changeset
|
23 e.comment = comment |
d0e8933d6dad
i18n: merge i18n comments of translatable texts correctly
FUJIWARA Katsunori <foozy@lares.dti.ne.jp>
parents:
29153
diff
changeset
|
24 else: |
d0e8933d6dad
i18n: merge i18n comments of translatable texts correctly
FUJIWARA Katsunori <foozy@lares.dti.ne.jp>
parents:
29153
diff
changeset
|
25 e.comment += '\n' + comment |
11389
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
26 else: |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
27 po.append(entry) |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
28 cache[entry.msgid] = entry |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
29 |
43691
47ef023d0165
black: blacken scripts
Gregory Szorc <gregory.szorc@gmail.com>
parents:
41759
diff
changeset
|
30 |
11389
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
31 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
|
32 entry = polib.POEntry() |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
33 entry.merge(orig) |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
34 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
|
35 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
|
36 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
|
37 return entry |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
38 |
43691
47ef023d0165
black: blacken scripts
Gregory Szorc <gregory.szorc@gmail.com>
parents:
41759
diff
changeset
|
39 |
11389
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
40 if __name__ == "__main__": |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
41 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
|
42 |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
43 cache = {} |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
44 entries = po[:] |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
45 po[:] = [] |
43691
47ef023d0165
black: blacken scripts
Gregory Szorc <gregory.szorc@gmail.com>
parents:
41759
diff
changeset
|
46 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
|
47 for entry in entries: |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
48 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
|
49 if entry.msgstr: |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
50 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
|
51 else: |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
52 msgstrs = [u''] * len(msgids) |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
53 |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
54 if len(msgids) != len(msgstrs): |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
55 # 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
|
56 # 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
|
57 # 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
|
58 # translation - erasing extra paragraphs is |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
59 # 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
|
60 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
|
61 entry.flags.append('fuzzy') |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
62 msgstrs = [entry.msgstr] * len(msgids) |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
63 |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
64 delta = 0 |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
65 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
|
66 if msgid and msgid != '::': |
11389
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
67 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
|
68 mdirective = findd.match(msgid) |
ff6ab0b2ebf7
i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents:
11389
diff
changeset
|
69 if mdirective: |
43691
47ef023d0165
black: blacken scripts
Gregory Szorc <gregory.szorc@gmail.com>
parents:
41759
diff
changeset
|
70 if not msgid[mdirective.end() :].rstrip(): |
20362
1bce1078501d
i18n: leave out entries which contain only a rst directive
Simon Heimberg <simohe@besonet.ch>
parents:
20361
diff
changeset
|
71 # only directive, nothing to translate here |
28074
a1924bc6e267
i18n: calculate correct line number in source of messages to be translated
FUJIWARA Katsunori <foozy@lares.dti.ne.jp>
parents:
20363
diff
changeset
|
72 delta += 2 |
20362
1bce1078501d
i18n: leave out entries which contain only a rst directive
Simon Heimberg <simohe@besonet.ch>
parents:
20361
diff
changeset
|
73 continue |
20359
ff6ab0b2ebf7
i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents:
11389
diff
changeset
|
74 directive = mdirective.group(1) |
20363
e3ee7ec85a15
i18n: leave out entries which contain only rst syntax
Simon Heimberg <simohe@besonet.ch>
parents:
20362
diff
changeset
|
75 if directive in ('container', 'include'): |
e3ee7ec85a15
i18n: leave out entries which contain only rst syntax
Simon Heimberg <simohe@besonet.ch>
parents:
20362
diff
changeset
|
76 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
|
77 # only rst syntax, nothing to translate |
28074
a1924bc6e267
i18n: calculate correct line number in source of messages to be translated
FUJIWARA Katsunori <foozy@lares.dti.ne.jp>
parents:
20363
diff
changeset
|
78 delta += 2 |
20363
e3ee7ec85a15
i18n: leave out entries which contain only rst syntax
Simon Heimberg <simohe@besonet.ch>
parents:
20362
diff
changeset
|
79 continue |
e3ee7ec85a15
i18n: leave out entries which contain only rst syntax
Simon Heimberg <simohe@besonet.ch>
parents:
20362
diff
changeset
|
80 else: |
e3ee7ec85a15
i18n: leave out entries which contain only rst syntax
Simon Heimberg <simohe@besonet.ch>
parents:
20362
diff
changeset
|
81 # lines following directly, unexpected |
43691
47ef023d0165
black: blacken scripts
Gregory Szorc <gregory.szorc@gmail.com>
parents:
41759
diff
changeset
|
82 print( |
47ef023d0165
black: blacken scripts
Gregory Szorc <gregory.szorc@gmail.com>
parents:
41759
diff
changeset
|
83 'Warning: text follows line with directive' |
47ef023d0165
black: blacken scripts
Gregory Szorc <gregory.szorc@gmail.com>
parents:
41759
diff
changeset
|
84 ' %s' % directive |
47ef023d0165
black: blacken scripts
Gregory Szorc <gregory.szorc@gmail.com>
parents:
41759
diff
changeset
|
85 ) |
20359
ff6ab0b2ebf7
i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents:
11389
diff
changeset
|
86 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
|
87 if not newentry.comment: |
ff6ab0b2ebf7
i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents:
11389
diff
changeset
|
88 newentry.comment = comment |
ff6ab0b2ebf7
i18n: posplit writes a warning for translators before rst directives
Simon Heimberg <simohe@besonet.ch>
parents:
11389
diff
changeset
|
89 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
|
90 newentry.comment += '\n' + comment |
11389
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
91 addentry(po, newentry, cache) |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
92 delta += 2 + msgid.count('\n') |
4fd49329a1b5
i18n: script for splitting large messages on .po/.pot files
Wagner Bruna <wbruna@yahoo.com>
parents:
diff
changeset
|
93 po.save() |