Allow 'fuzzy' access to Magpie translation cache
Description
Activity
Former user 27 April 2018 at 05:42
that's right, it would have to have been already cached by (I expect) a non-private project. As the database grows, it should be more likely to provide results with smaller strings (considering the segmentation of input), same as regular TM.
Sean Flanigan 27 April 2018 at 01:12
Oh I see, so full text search for previous machine translations. Interesting.
But I'm not sure it would get much use if we start running the editor's "current string" through MT - ie "MT suggestions", since that will provide a 100% match at the cost of translating one string (unless already cached).
Former user 27 April 2018 at 00:27
this came from a discussion that and I had. Simply put, it could be basically another source of data for the editor's Suggestions panel. Fuzzy match the MT database, and provide match percentages.
Former user 27 April 2018 at 00:09
Could potentially be the same endpoint, but with a parameter to say "offline" or similar.
Sean Flanigan 19 April 2018 at 05:30
Can you explain this one please? Is this in the context of Zanata Platform translators using MT assistance? Or something else?
Details
Details
Assignee
Reporter
Former userPriority
unspecified
Use the indexing of the Magpie database to provide Translation Memory like matching behaviour, for additional translator assistance.
Requires a new endpoint to access the database, and an efficient index search.