How to trap for punctuation when splitting words in target lg

Question

How do I trap for punctuation when splitting words from the source to the target lg? In the following screenshot du-wa-uyu '3p.S-EXIST-INC / they are still there' has a comma after it. In the target language, the incompletive is a separate word, not an enclitic. But the comma ends up coming out in between the two words, and I need it to come out after the two words:

Here's my rule:

Answer 1

In looking at your source text, you have two verbs separated by a comma. Perhaps unexpectedly, your phrasal verb rule is running and it seems to be a quirk of Apertium that when you have some stuff (like a comma) between two words and both words get matched for a rule, any additional thing that gets put out before word 2 is going to come after the stuff.

You probably don't want the phrasal verb rule firing in this case anyway. You can capture the comma by making it part of your sentence ending punctuation. In this way commas will be in their own lexical unit. And a pattern like v v won't match because the data stream will be v sent v. Now if you are interested in matching across a comma (or other sentence punctuation) you would have to include 'sent' in the pattern.

P.S. Your first image was almost unreadable. Please make it a little higher resolution.