This list is generated by running "tox -e zanata-users-sync"
command to sync translation members up-to-date similar as
I5a309564fae7229fdc097715bdbb1ce85d109753
Change-Id: I00b39bb1d6e6a87a6105ab1bf298b9046ffe6d2d
Added weblate API codes to get ready for
Zanata to Weblate migration
Implements: blueprint weblate-tools
Change-Id: I8d1a2393cc50541a29bd9581c4ac2574b9dc9813
Rewrites existing ZanataUtils.py to work with Weblate as part of migration work.
Key summary:
- IniConfig class name is the same
- ZanataRestService class name is renamed as WeblateRestService class
with the same functionality
- ProjectConfig class is not implemented since Weblate does not
have project configuration file like zanata.xml.
Implements: blueprint weblate-tools
Change-Id: Ie4cf861cc6ca59456c5d563d7673fbf54b3bb42f
Added weblate API codes to get ready for
Zanata to Weblate migration.
WeblateUtils.py has been deleted from this PR, as the
i18n main contributor decided it would be better to
seperate the PR with other files.
weblate_userinfo.py creates .csv file as zanata_userinfo.py does.
Implements: blueprint weblate-tools
Change-Id: I71bb234ea5cc5356c12245379c607340ea97249d
(g)awk in Ubuntu 20.04 or later (perhaps gawk >=5.0) has more strict
regexp check and the regexp passed to awk in generatepot.sh cannot
pass the regexp check. This means our script does not work on
Ubuntu 20.04 or later.
This regexp was introduced to strip lines with UUID in a POT file
(e.g., "# 55a7796f27cb4468820fa48b90515add"), but the recent versions
of sphinx does not output such information at least with the sphinx
configuration used in OpenStack CI.
As a result this awk expression in generatepot.sh does nothing,
and I think we can drop this awk expression.
Change-Id: Ia61f74d0a73cc012afde6a4e16ac4431687a65a1
doc-pot-filter.sh is prepared to drop numbers from *.csv.
Howover, the path of CSV files in POT files were changed from some version
of Sphinx (from "doc/source/data/*.csv" to "../../source/data/*.csv") and
doc-pot-filter.sh does not work expectedly.
This commit adjusts the path of CSV files in POT files.
The number of entries in the target POT file (atc-stats.pot) is printed
before and after the filter is applied. I believe this helps us debug
the similar issue in future if needed.
Change-Id: I32d760a735ab1ae491d7e5efb36fd5059da0cd11
As part of commit 9b939e1b40,
"bash" was added when calling tools/generatepot.sh to avoid tox warning,
but as a result "-xe" option specified at the beginning of generatepot.sh
is no longer considered and we cannot see trace of generatepot.sh
in logs of upstream-translation-update job :-(
This commit explicitly set these options in the shell script
instead of the magic line (line 1).
doc-pot-filter.sh is also updated in the same way.
Change-Id: Ia2ca111cbb386ae5333d8a9a8622f7ac9b19746e
This list is generated by running "tox -e zanata-users-sync"
command to sync translation members up-to-date similar as
I5a309564fae7229fdc097715bdbb1ce85d109753
Change-Id: Ica09d0348863aca9822eef7d9978a36b2ff3aac6
This patch enhances the speed of Zanata stats by using
user-based Zanata APIs rather than project-based APIs,
which took >1 days as more Zanata users and translation data
were accumulated.
Note that the API scheme follows with Stackalytics
: https://opendev.org/x/stackalytics/src/branch/master/stackalytics/processor/zanata.py#L38
Change-Id: Ia2462447f7a0cc2534e877976838de09e8683e89
The path of ATC stats CSV files in the generated POT seems to have changed.
As a result, strings from ATC stats are included to the generated POT file.
There is no need to translate them and the translation percentage will go
down. This commit adjust the filepath in doc-pot-filter.sh to make the POT
filter work as expected again.
Change-Id: Ieeee86619a561fcca216b568797bd6c3c1de4d62
This list is generated by running "tox -e zanata-users-sync"
command to sync translation members up-to-date.
Change-Id: I5a309564fae7229fdc097715bdbb1ce85d109753
Zanata >=4 which is now on production utilizes REST API [1].
This commit fixes zanata_users.py to make compatible with
Zanata REST API and syncs translation_team.yaml up to date.
[1] http://zanata.org/zanata-platform/rest-api-docs/resource_LocalesResource.html
Change-Id: Ia99db310cff236797ef403ba5534d853ed63e3e3
To generate pot files using Sphinx, just using "-b gettext"
is sufficient and there is no need to additionally specify
"-b html".
Change-Id: I2eebeb5c575ed99be25980fbd3ec560f3bce2da3
- Changes generatepot logic which is general for all OpenStack
project documentation translation support
- Generalizes generatepot.sh for all genereatepot jobs
for project documentmentation translation by separating
exclusion logic into doc-pot-filter.sh
- Utilizes msggrep to filter non-translation targets
and not to include in translation source pot files
doc-pot-filter.sh
Change-Id: Ibe0dcf599ad99db15af1ba50c9ddf0af5487fa24
We are discussing an approach on translation support in
project documents in https://review.openstack.org/#/c/543220/.
In this approach POT files for RST files in subdirectory are
treated differently. In the i18n doc, atc-stats directory is
the only exception and the thing would be much simpler if it
is moved to a top directory.
tools/generatepot.sh is updated accordingly.
Also generatepot.sh is added to pep8 tox env to confirm
generatepot.sh works successfully.
Change-Id: Ia227358cf9681e8d1745f4c260fe9c97ab18a3e7
The new PTI doesn't work for the way the i18n docs are built. Add a new
doc build job that runs tox -edocs for until we figure out a better path
to success.
Remove the add-marker code since that's done in the doc publication
jobs. Also stop rsyncing the code to a different directory, since it'll
make the normal build case wrong and is otherwise handled by the
publication jobs.
Change-Id: I5237e750b4bc4899d71952a14c8d2ca92d9985a0
The previous version did not care the case where documents are placed
in any sub directories. It assumes all documents are placed
in the top doc directory.
We now have atc-stats/***.html and the previous build script
does not work for these pages.
This commit fixes the problem.
Change-Id: Ic19c2de6e4bf3f4f5c7ae6dd79372a19fddcc7c3
*.csv files which contain translation statistics
do not need to be translated to local languages.
Change-Id: I88f11b8fef7221129c81493a17830f271b484767
Co-Authored-By: Ian Y. Choi <ianyrchoi@gmail.com>
Closes-Bug: #1710787
Previously we kept the order of members of individual language teams
in the same order as that of Zanata language team page.
It was mainly to make it the maintenance of the YAML file
as the file is manually maintained.
We use zanata_users.py script to sync translation_team.yaml now,
so the need to keep the order has decreased much.
In addition, it seems Zanata changes the order of members
even when members of a specific language team is not changed.
This happens during the review https://review.openstack.org/#/c/485966/.
This commit change the script to sort members in the alphabetical order.
Change-Id: I09ddc202e0a3021bbc5433e37a5d815c760697ff
It is a boring job to review a patch to update the team member list.
To avoid this, this commit proposes a bit improved approach.
- When you want to propose an update of the team member list,
you need to sync all members with Zanata.
If you want to add you, you need to propose a latest member list.
The new list may contains someone other than you.
- pep8 job checks a proposed member list is up-to-date.
If not, the job fails.
To make it easy to sync all members with Zanata, a tox target
'zanata-users-sync' is added. To download the latest member list
of all language teams, just run:
tox -e zanata-users-sync
Change-Id: I393a6b77261f6a3d7788a4c4f57bd6fdc95b5146