ue4textextractor seems to only pull bits and pieces instead of full text, so I have to feed it more.
I start with ue4textextractor to get a base, use fmodel to export all the assets as jsons, use a script to pull any "Key" and "SourceStrings" out of the jsons and to calculate the string fcrc's, translate with google, another script to reformat the results and create a new locres with ue4textextractor.
Technically the scripts aren't necessary, but I would be hunting through
View attachment 5556506
and this was a "small" game AND I'm too lazy for that. Though if you wanted to be pedantic about it most text is in the data tables, which if the dev kept proper Unreal naming scheme start with "DT_" that and anything that might be a UI element.
Same could be said for the rest. If I were hunting manually I could just add lines to the locres myself with ue4localization tool, but again
View attachment 5556514
(I know I have some duplicates) no thanks.