Currently information about each field is scattered around in different dictionaries of MeshSeq and TimePartition: fields, _fs, field_types, etc. I think we should introduce a class Field that would hold all of these at the same place.
So e.g. u = Field(name="velocity", type="unsteady") etc.
Sounds good, thanks. Note that I think fields were originally associated with TimePartition in case we wanted to support having different timesteps for different fields.
Currently information about each field is scattered around in different dictionaries of
MeshSeq
andTimePartition
:fields
,_fs
,field_types
, etc. I think we should introduce a classField
that would hold all of these at the same place.So e.g.
u = Field(name="velocity", type="unsteady")
etc.