That is extremely unfortunate.
I also don't have a backup of that repo.

They also removed https://github.com/Financial-Times/useragent_parser

Maintaining a UA parser and/or normalizer is not something we have capacity for :(

Here are the closest forks I've found: https://github.com/christianhaller3000/polyfill-useragent-normaliser and https://github.com/EthicalSecurity-Agency/Financial-Times_useragent_parser. We could compare these to what's on NPM to see what they are missing.

It looks like https://github.com/EthicalSecurity-Agency/Financial-Times_useragent_parser has also been removed. I can try to find another recent fork, but I'm not sure what we're thinking of doing with it, given #92.

How large of an effort do we think it will be to maintain a normalizer? It seems like the UA parser libraries already do a kind of normalization, which would then need to be mapped to our list of browsers and versions.

As a test, I just took a list of user agents (link) and ran them through @financial-times/polyfill-useragent-normaliser and a simple normalizer that I just wrote that uses ua-parser-js; the simple normalizer matches 90% of the user agents in the list.