Wrong signature in MultiPeakModel

kmnhan commented 6 months ago

Description

After a call to make_params, sigma and amplitude which are not real function parameters appear in function signature detected by lmfit.

Expected behavior

Sigma and amplitude should not appear in parameter names, nor appear in the covariance matrix.

Minimal Complete Verifiable Example

import numpy as np

model = MultiPeakModel(npeaks=1, fd=False, convolve=False)

print(model.param_names)
# ['p0_center', 'p0_width', 'p0_height', 'lin_bkg', 'const_bkg']

model.make_params()

print(model.param_names)
# ['p0_center', 'p0_width', 'p0_height', 'lin_bkg', 'const_bkg', 'p0_sigma', 'p0_amplitude']

MVCE confirmation

[X] Minimal example — the example is as focused as reasonably possible to demonstrate the underlying issue.
[X] Complete example — the example is self-contained, including all data and the text of any traceback.
[X] Verifiable example — the example copy & pastes into an IPython prompt, returning the result.
[X] New issue — a search of GitHub Issues suggests this is not a duplicate.
[X] Recent environment — the issue occurs with the latest version of erlab and its dependencies.

Additional context

No response

kmnhan commented 6 months ago

This also happens for built-in models so it is not our problem, but strangely it doesn't happen for composite models. Maybe a bug in lmfit

kmnhan commented 6 months ago

On second thought, make_params should be called before anyone fits anything anyways. Should just call it before getting the parameter names in the fitting accessor.

kmnhan / erlabpy