Higher order fit when the secondary is in Gaia

[gmbrandt]

Currently, we assume that if the companion is in Gaia, then it will have an astrometric fit with the same number of parameters:

orvara/orvara/orbit.pyx

Lines 975 to 979 in 8e51a3f

	# TODO for cases with companion in gaia, the npar might be different. Need to handle this.
	# can fudge it a bit, by just doing the higher parameter fit and throwing away the higher order fit.
	for i in range(Gaia.npar):
	for j in range(Gaia.npar):
	chi2mat_Gaia_B[i*Gaia.npar + j] = Gaia.chi2_matrix[i, j]

We then just throw away the higher order parameters of that fit, and only compare and care about the first 5 parameters:

orvara/orvara/orbit.pyx

Lines 1012 to 1021 in 8e51a3f

	if data.Cinv_G_B[0, 0] != 0:
	# TODO this is the case where the companion is in Gaia (e.g., its some widely separated stellar object)
	# in this case, we should ideally have a separate fitter for this gaia object, which pulls the
	# fit degree (5, 7 or 9 parameters) for this object and uses an appropriate htof fitter for it.
	# for now though... just doing a high order fit and throwing away the higher order parameters is OK.
	lstsq_C(chi2mat_Gaia_B, b_Gaia_B, Gaia.npar, Gaia.npar, res_Gaia_B)
	RA_G_B = res_Gaia_B[1]
	Dec_G_B = res_Gaia_B[2]
	pmra_G_B = res_Gaia_B[3]
	pmdec_G_B = res_Gaia_B[4]

Ideally, we should update this so that if the companion exists in gaia, it will have a Fitter object with the correct number of parameters.