Added check function for weighted data

[briederer]

I ran into a small problem, that curve_fit(model, xdata::AbstractArray, ydata::AbstractArray, wt::AbstractArray{T}, p0::AbstractArray; inplace = false, kwargs...) where T tried to fit my function for a weight 1/0 = Inf.
So I thought it would be useful to also check weights for NaN, Inf or Missing-values if provided.

[briederer]

@pkofod Any chance to get this merged soon(ish)?
LsqFit.jl is currently breaking my full analysis-program and it would be nice to have this fixed 😅

If there are any objections let me know too so I can adapt it.

[pkofod]

I'm very sorry for taking so long here. I think it would be a bit simpler if you just gave wt a default value that would succeed

check_data_health(xdata, ydata, wt=[])

or something similar

[briederer]

Thanks for the comment.
Your suggestion seems to be the cleanest possibility.
I added the modification and also already rebased onto master.
Let me know if there is anything else I can do.

[briederer]

Sorry for all the force-pushes but the commit history was kind of confusing and I wanted to clean it up.

[pkofod]

I'm sorry for abandoning this discussion. Is this still relevant to you? If so, I would ask of you to add some tests, and then I will merge and release it asap.

[briederer]

No worries 😉
So far I've simply used my fork.
IMHO I still think that also the weights should be checked for consistency unless a Inf weight has a special meaning which I've missed so far?
I'll add tests and rebase the PR as soon as I find some time.