Tyler Hughes

Might be a big change, so let's flesh it out a bit more before digging in: So the thought here would be perhaps to define a mapping in `tidy3d.constants` for...

> ... maybe we want to do just CustomFieldSource for now rather than also custom current source, which is less likely to be needed. One point related to that is...

I see your point about the centering of the monitor. I think I missed the bit about the center of the data being placed at the source center. I instead...

Sounds good to me overall ========= I forgot to add a note that to the processing (centering, normal field removal, etc) should be done on backend, what you wrote makes...

Ok, next week I can set this up: * remove `CustomCurrentSource` (?) * use`field_data : FieldData` as a `Field()`? or how do you want to proceed? Note that we can...

My last commit changed things according to our discussions: * Removed `CustomCurrentSource` * Added `field_data : FieldData` as the data container in `CustomFieldSource`. * Added validator assuring a single frequency...

> The only place where self.monitor is now used in a FieldData is in self.field_components. I wonder how we can replace this to automatically check what data attributes are present...

Ok, so I just made a few changes: * monitors are optional in the MonitorData, reshuffled things around to avoid needing monitors as much as possible. will wait until we...

* Rebased against develop, fixed merge conflicts * Fixed tests, note that I underscored `_test_data_performance`, it doesn't test much, breaks due to memory requirements of the GitHub actions and is...

> So that's just because the tests set things up "wrong" no? The SimulationData is wrongly constructed for its associated Simulation? Yea it was set up incorrectly (by accident) in...