If we were to autogenerate maps of the h5 data tree with variable types for each granule type then we could replace a lot of the repetitive code with loops. It may also allow us to move to a single points function for all granule types.

Both #55, #57 and most others consist of a lot of find and replace as everything is hardcoded. If we were to autogenerate maps of the h5 data tree with variable types and save the output in the package then a lot of code could be replaced with loops. We could autogenerate new data trees each time a new version of the package was created.

This would also allow the user to pass a list of variables to the points function if they wanted more/less than the defaults.

This would also make it easier acomodate new granule types.