I have done this a few times now and it actually isn't too bad. A parallel installation is a good way to be extra safe, though I might suggest you just use another environment within your existing Miniconda3 installation. In fact, the environment is probably the least of your worries since changing from the old hg repo to the github repos is the larger pain point in my experience (rather than the other python packages like numpy and whatnot). You could realistically just save a copy of your installed packages and versions (just in case) and delete your current environment/installation and start from scratch. You could also duplicate your current environment to save a copy of it, again just in case, and re-use your current installation as well.
Updating to the new repos is a little trickier if you have any local changes, which would obviously need to be manually ported over to the local github repos you install (Note: the developer installs described in the web documentation are the equivalent of how the hg installations worked in the past). Since there have been a lot of updates since the hg repos were abandoned, this can get a little tricky depending on what you may have actually done.
The other pain point is that some of the paths for configuration files and the like have changed. Think things like where the script files/analysis files are stored in particular. Luckily the sub-folder names are all the same as before, just the parent location for the userlib directory moves relative to where it used to be.
Anyway, that's a lot of detail and what applies depends a lot on what your configuration looks like at the moment. My general advice is that so long as you save a backup in a separate location of any custom modifications to labscript with your various config/script/analysis files, there is basically nothing that can't be fixed/undone.