Lat-Lon-Naming

[Karinon]

@wachsylon had a irregular-grid dataset which named their dimensions latitude and longitude instead of the omni-present lat and lon. I think that instead of dealing with this issue in every single Globe*-File separately, like we do that now, and in GlobeView aswell for good measure, we should try to to this somewhere in a common place and pass the dimensions to the globes. An alternative approach would be two global functions getLon and getLat which deal with all the different variations, but also need to take care of rlon, in which case we also need the information that something is regional. Not sure if there are more hidden information behind the naming of lon and lat somewhere....

[d70-t]

It's quite likely that people will create datasets which don't follow conventions, but people in our field tend to like CF-conventions. I'd suggest that we should as a first measure support these, specifically §4.1 and 4.2. That is:

• the latitude coordinate is a variable which has a units attribute, which is set to any of the following strings: degrees_north, degree_north, degree_N, degrees_N, degreeN or degreesN
• the longitude coordinate is a variable which has a units attribute, which is set to any of the following strings: degrees_east, degree_east, degree_E, degrees_E, degreeE or degreesE

Notable there are no other strict formal criteria.

We might want at some point extract tools to handle the meaning of variables into a separate library, as I guess there could be more projects which could benefit from these. In particular we've had to do something similar for the ORCESTRA data browser, where we combined the CF rules with a variable-name heuristic: https://github.com/orcestra-campaign/ipfsui/blob/d9e255c5f255aa4535b2106a4c87c6106dcf30ec/src/utils/cf/index.ts#L84