Description of problem: Transifex now have the ability to give suggestions and translation history in its web translate UI. So each project keeps a lot of translation suggestions and history. The history function can be used to identify who translated the string and translated time. The suggestion function is very useful for the translation of duplicated strings as well as update the old translation. In addition, the powerful searching tool can be used to search many details of the translation process. For example, we can use it to find which string is not translate, which is translated after certain time, or who translated the strings. This really helps the translation efficiency and quality. So I want to know, are the functions available in Zanata? If Zanata has, can these data from Transifex be moved to Zanata as well? If Zanata don't have, will you consider to develop any similar tools for us?
- RE: translation history in Transifex Unless Transifex provides an API for us to download/export those history, there is no way Zanata can do to retain those. What Zanata can do is to pull down the latest translation for each project and make them available. - RE: translation suggestion Zanata also reuse translations with Zanata itself, in the form of TM (translation memory) or copyTran( copy translation). - RE: powerful search Zanata provide project wide search and replace, filter by translation state, translator, translation time, partial text search, comment search and resource id search, plus validation error search (i.e. strings that has failed certain validation rule such as missing variables).
Some of the translation suggestion/reuse is covered in http://zanata.org/help/ We will be adding more contents to that. Please help us to identify what's missing there.
In addition to what Patrick already mentioned about translation memory, Zanata offers the option to import external translation memories in a TMX format. If there is a way we can extract such a file from Transifex, then it can be imported into Zanata, and used in conjunction with Zanata's own internal translation memory, which works as Patrick mentioned.
TianShixiong and Zanata team * 'suggestions and translation history' feature This is also called TM (aka translation memory) both in transifex and zanata. The feature serves similar way in both. So the answer is 'YES' Zanata already has this feature. * 'searching tool' feature Again 'YES" Zanata already has this feature. last-modified-after serves same as 'translated after certain time'. last-modified-by serves similar to 'who translated the strings'. Also text, resource-id, last-modified-before, source-comment, translation-comment and msgctxt are available in Search field. In another fields, filtering Untranslated, Fuzzy, Rejected, Translated, Approved or Invalid is available. Here Untranslated can serve same as 'which string is not translate'. TianShixiong, which particular data you like to migrate from Transifex? If you can list all data items you want but can't find in Zanata, it is helpful for Zanata team to evaluate one by one. Once an item is identified that Zanata only needs API from Transifex, I can ask Transifex if they can provide.
(In reply to Noriko Mizumoto from comment #4) > TianShixiong and Zanata team > TianShixiong, > which particular data you like to migrate from Transifex? > If you can list all data items you want but can't find in Zanata, it is > helpful for Zanata team to evaluate one by one. Once an item is identified > that Zanata only needs API from Transifex, I can ask Transifex if they can > provide. That's great. Now I just hope Transifex can provide the API to move these data. Keep the history is important to translators.
(In reply to Patrick Huang from comment #2) > Some of the translation suggestion/reuse is covered in > http://zanata.org/help/ > > We will be adding more contents to that. Please help us to identify what's > missing there. I think a shortcuts list should be added there.
Just to clarify a few concepts here as I fear we might be mixing them up: When I say history, in Zanata that means the sequence of states and translations that every string has gone through time to to finally get to the current translation. This history is NOT taken into account by Zanata for Translation Memory matches. It is there just for reference, comparison and overall tracking. Translation memory (TM) are the suggestions that Zanata makes to translators. These suggestions come from two places: other projects' current translations (not history), and any externally imported translation memories. So, if you want to see historical data from Transifex used in TM suggestions, and IF Transifex can export this data as a TMX file, we can import it into Zanata and it will show up in the TM suggestions.
Created attachment 932010 [details] Screenshot of suggestion and history feature in Transifex
Created attachment 932011 [details] Screenshot of search feature in Transifex
(In reply to Carlos Munoz from comment #7) > Just to clarify a few concepts here as I fear we might be mixing them up: > > When I say history, in Zanata that means the sequence of states and > translations that every string has gone through time to to finally get to > the current translation. This is what I meant. > This history is NOT taken into account by Zanata for Translation Memory > matches. It is there just for reference, comparison and overall tracking. Yes, you are right. The historical data would be imported to Zanata history for continuous reference, tracking and etc. If Transifex will give the API access as you said. > > Translation memory (TM) are the suggestions that Zanata makes to > translators. These suggestions come from two places: other projects' current > translations (not history), and any externally imported translation memories. > If the suggestion data can be imported as well, then it's great. > So, if you want to see historical data from Transifex used in TM > suggestions, and IF Transifex can export this data as a TMX file, we can > import it into Zanata and it will show up in the TM suggestions. No, historical data should still be history records in Zanata, suggestion should be used in TM. In addition, I made two screenshots for guys who may not understand what I am concerning.
To sum up, The searc function, translation history and suggesting (TM match) are in our production servers. However, we have not yet implement methods to extract history data from transifex. So the migration only migrate the source documents (e.g. .pot files) and translation documents (e.g. .po files).
Migrated; check JIRA for bug status: http://zanata.atlassian.net/browse/ZNTA-250