I had a brief look at what needs to be done so we can grab ctags parsers - the history of the Python parser summarizes it quite nicely:
https://github.com/universal-ctags/ctags/commits/master/parsers/python.c
There seem to have been some refactorings/renames in main we'll need to get but there doesn't seem to be anything major fortunately. I'll make a pull request to make these changes first so parser syncing can start.
Question: there seem to be quite massive changes in the lregex.c/h upstream and this one isn't so trivial to merge because we use GRegex and ctags uses GNU Regex. Syncing this would mean taking care of the different semantics of the two regex libraries which might be quite a lot of work. What if we just removed support of regex-based parsers from Geany and #if0'd the bodies of all functions from lregex.c? Currently we use regex parsers for HTML, Cobol and ActionScript - for HTML there's a token-based parser upstream we can use which would mean we'd just kill parsing support for Cobol and ActionScript. I may be biased but I don't think these particular languages are so important to justify all the work related to the regex-based parsers.