If I'm reading you right, you have something like this:
some text here in some language
and maybe some more
and then a line ending in a comma,
with the comma-terminated line followed by something like the above
and then some more text
and another comma,
or something of the sort
You say "comma", and I have assumed comma (,) but you show a full stop (.). Whichever it is, I'm sure you can adjust to suit.
You should be able to use a Find and Replace ...
firstly press the "More" button if need be and check "Use Wildcards"
to find a comma, followed by a paragraph mark, followed by some text (which does not include a paragraph mark so that it is all on the same line), followed by a semicolon ... put (,^13[!^13]@); in the "find what" box
to replace it with just the text, as found, the same and the semicolon replaced by two colons ... put \1:: in the "replace with" box
and press "Replace All"