Hi,

I use Studio Pro 2017 on Win 10 pro. I have a JSON file with exported content from a simple database (I parse it using Supertext's plugin). The export functionality is not incremental, so I always get all the source content, not just new/updated JSON subpaths. The "strings" are product descriptions and may get a little wordy, thus some contain line breaks (not paragraphs), which are not a default segmentation break point. This is unfortunate, since the TM contains every single sentence, but split (the content was originally translated by copying the strings into a text file, so the line break was forced into the segmentaton). As a result, I get little pretranslation whereas I know for a fact that 98% of the translations already exist in the TMs (I can get them through concordance and autosuggests).

I dug out a number of articles on the topic and, following (Paul Filkin's) suggestions found, for instance, here and here, I created a segmentation rule using the regular expressions .[\n]+ before the break and . after. I also tried omitting the leading point in the first expression as suggested in the comments to one of the post. Unfortunately, the result doesn't change: line breaks get included in larger segments, and don't act as break characters.

My feeling is that the JSON parser prevails, and the content is treated as paragraph-based, with each JSON subpath treated as a single paragraph; so, no matter what I try, the segmentation rule will be overridden.

However, if there IS a way to force the segmentation as I desire, I'd like to hear it before I go manually re-translating the files using concordance and forcing the paragraph-based segments in TM.

Thank you!

Stefano