Native support for Cauchy model

[tomflexcompute]

Cauchy model for the refractive index is quite commonly used for semiconductors and glass materials, similar to the Sellmeier model. Currently users can define a Cauchy model and use our fitter to fit it, which works well in most cases. Is it possible to convert a Cauchy model to a pole-residue model directly without fitting? If so, we can support this model and avoid asking users to manually do the fitting.

Feel free to close this issue if such a conversion is not possible and fitting is necessary.

[weiliangjin2021]

Is the two-term form sufficient?

[tomflexcompute]

It looks like people often use 3 or 4 terms. Two term form would just be equivalent to Sellmeier?

[tylerflex]

can we close this because of #1974?