Consider adding convenience functions for loading data

[JamesVarndell]

Is your feature request related to a problem? Please describe.

At present, data is loaded into earthkit-data using the from_source(source, *args, **kwargs) function. This function is heavily overloaded, as the value of the source argument changes the valid args and kwargs.

For example, when loading a file, the call signature is essentially from_source(source, file_name) - but when loading data from the CDS, the signature becomes from_source(source, dataset, request).

This makes the function impossible to document in a standard way because the valid arguments change depending on the source. Given the wide (and increasing) range of valid source arguments, users will have to scan through a lot of documentation to find the arguments that apply to their particular source.

Describe the solution you'd like

I would like to see a from_{source_name} function for each source type, e.g. from_file, from_cds, from_url, from_fdb. The benefits of this approach are:

• much easier to document - each function only takes one set of arguments
• easier to quickly find functionality in modern text editors and Jupyter-notebooks - depending on string arguments for source in the current implementation relies on the user reading the documentation to find all valid source values, but function names work well with autocomplete to help users quickly find the functionality they need
• easier to read - especially for sources like the CDS which take a longer list of arguments

Note that I'm not suggesting to remove the from_source function, but rather create convenience functions which essentially call from_source with the source argument pre-populated.

Describe alternatives you've considered

No response

Additional context

No response

Organisation

ECMWF