Investigate how the translation is split by newlines in the codebase. Look for functions that handle text segmentation before translation. The fix should ensure that sentence boundaries are respected, possibly by using a sentence tokenizer. Check the project's translation pipeline to identify where to implement this change.
