Recognition of line breaks is a crucial feature in making our output more legible without forcing the user to create a series of small phraserects.
An algorithm to decide which words are on the same line should not be hard to put together based on the new heuristic, and with that we could either split the detected text into lines and send them to google as separate strings, or possibly just insert linebreaks into the rawtext, if google will retain those in the translatedtext (i imagine they will.)
Recognition of line breaks is a crucial feature in making our output more legible without forcing the user to create a series of small phraserects.
An algorithm to decide which words are on the same line should not be hard to put together based on the new heuristic, and with that we could either split the detected text into lines and send them to google as separate strings, or possibly just insert linebreaks into the rawtext, if google will retain those in the translatedtext (i imagine they will.)