Objective
Produce tooling for translators and measurement of Linux translation progress. By extracting information coming from source files used in RPMs.
How to achieve this objective?
As RPMs change all the time, either the system would extract the full list of RPM for a particular Fedora version at a given time (is it doable?) or would register to RPM package activity using datagrepper.
From the source file, use https://pypi.org/project/translation-finder/ to identify translation file and python to copy the files.
From the translation file, use http://docs.translatehouse.org/projects/translate-toolkit/ to interact with content (statistics, aggregation, conversion...).
FAQ
Why not using list of ISO images? It was used for first tests, but it only contains a sub set of packages.
Which package updates are meaningful for this goal? What interest us is upstream code change, independently if this source code is part of one or many RPM package. Only new package name and version interests us, the package release is not meaningful.
Which Fedora versions do we want? We only want actual branches, rawhide activity is not mandatory.
Open questions
If the system registers to RPM activity notification, how to run first initialization?
Notes
about how to register to package activity
Way to do it:
- Extract build information from Datagrepper
- probably this topic https://apps.fedoraproject.org/datagrepper/raw?rows_per_page=1&delta=800&topic=org.fedoraproject.prod.buildsys.build.state.change with new state=1 (documentation: https://fedora-fedmsg.readthedocs.io/en/latest/topics.html#id145)
- Ask koji the list of related RPMs using listBuildRPMs(Build) (https://koji.fedoraproject.org/koji/api)
- Ask koji files related to this RPM using getRPMDeps(rpmID, depType=None, queryOpts=None, strict=False)
- Ask koji files details about each file using getRPMFile(rpmID, filename, strict=False)
- If file was never processed before, then apply translation processing
Example:
Step | Link | Retrieved information | Comment/questions |
---|---|---|---|
1 | https://apps.fedoraproject.org/datagrepper/id?id=2020-eca3c072-1b32-483c-8e74-84ced83ef77e&is_raw=true&size=extra-large | koji build id=1589180 | can we have more than one build? |
2 | https://koji.fedoraproject.org/koji/buildinfo?buildID=1589180 | get linked rpm | Texte de la cellule |
3/4 | https://koji.fedoraproject.org/koji/rpminfo?rpmID=22787621 | Texte de la cellule | Texte de la cellule |