Okay so i wanted (and still want) to translate prime plugin and i did get the same question, so i made researches the best source of translation is effectively lotro it self. And more precisely the dat files.
There is 2 solutions to extract data from it
the first are this old extractors
LOCALDATAEXTRACTOR and DAT UNPACKER TOOL
I tested them with my antivirus first, then i got it working, the unpacker for what y remember (it's been several months since then).
What you get is directories with numbers and files with numbers too.
some contains lotro translations. But you have to convert them again with the LOCALDATAEXTRACTOR to obtain text files.
The huge problem with this solution is that the first extract all of lotro data, it's long and take a LOT(RO) of disk space.
The other solution is lotro-companion that is extracting lotro's data the same way, but not the translations at current time (i think). But the specific java library source has vanished from internet. And Lotro-companion is a two stepped software, the dev extrat lotro data, convert it to Xml and publish the final software.
So for the two what you looking for are resources id, a non real example is [11200][1120] that point to a specific translation the first will point to english, the second the text it self, with that a [11200 + 1][1120] will point to the deutsche text.
As you can imagine doing that by hand is fastidious and time consuming.
So... I'm trying to automate this. I will just give few details as it's not finished yet and too long of an interesting explanation. So just let say i've coded an additional java library attached to lotro-companion containing a lua interpreter with the same specifications of lotro. And with metalua i parse homeopatix's code like, modify it with translations then rewrite it as it. The missing part is the translations from lotro with is unfortunate...
I've linked my github to my profile, but you need to be mentally ready for stack manipulations...