Improve documentation around available solvers

[ablaom]

I found I had to dig a bit to get this. Perhaps each doc-string states explicitly the solvers that are allowed (including full path for MLJ users, eg, MLJLinearModels.Newton())

Also there is a broken link on the Solvers section of the manual to table of models (which I don't think specifies the available solvers for each model as the docs intimate??).

[tlienart]

Yes this is fair enough, I need to spend a weekend on this package and close off a few of the long standing issues

[ablaom]

I think the recent doc-PRs address this.