Adds fitter#258
Conversation
Adds fitter and instrument_response
samaloney
left a comment
samaloney
left a comment
There was a problem hiding this comment.
This is great, nice to see it finally working!
The biggest things I see missing are docstrings and test, as this is a core part of the package we need to though testing e.g hight level testfit non-spectral data/model, fit spectral data model with out SRM, fits spectral data model with SRM and also more basic unit tests.
On a similar topic an gallery example showing a some of the basic fits you can and some narrarive documentation on the fitter/fitting (this could wait for a separate PR)
| def _set_abledo_angle(self): | ||
|
|
||
| if 'Albedo' in self.model.submodel_names: | ||
|
|
||
| print(len(self._spectrum_object.meta['ph_axis'])) | ||
|
|
||
| replacement_albedo = Albedo(energy_edges=self._spectrum_object.meta['ph_axis'], | ||
| theta=self._spectrum_object.meta['angle']) | ||
| replacement_albedo.theta.fixed = True | ||
|
|
||
| self._model = self._model.replace_submodel('Albedo',replacement_albedo) |
There was a problem hiding this comment.
I think this should be derived from the Spectrum object e.g. spec.observer_location and spec.flare_location which would be SkyCoords
|
|
||
| self._model = self._model.replace_submodel('Albedo',replacement_albedo) | ||
|
|
||
| def _set_observer_distance(self): |
There was a problem hiding this comment.
Derived from spec.observer_location
|
|
||
| mask = self._fit_mask | ||
|
|
||
| self._spectrum_object = self._spectrum_object[mask[0]:mask[-1]+1] |
There was a problem hiding this comment.
What about using the mask / or subscripting the final data before it goes into the fitter? Also what about the SRM?
There was a problem hiding this comment.
The SRM is clipped just below this
| data_spec_units=u.dimensionless_unscaled, | ||
| conversion_factor=1*u.dimensionless_unscaled, |
There was a problem hiding this comment.
Eventually these should probably live the the SRM object (an NDCude with 2 spectral axes in input or photon and one output, count or data)
There was a problem hiding this comment.
This shouldn't show up here needs to merge or rabase main
2 participants
Adds the fitter object
This PR adds the fitter object which intakes the NDcube spectrum object and can perform fitting in count space. This includes the addition of a plotter for fit results.
Tasks:
instrument_response.py