sparse matrix methods

[topepo]

It would be nice to havedgCMatrix support (now that I have an application that can really use that)

[DavisVaughan]

I would need a little bit more information about this. Currently mold() forces any matrix input into a data frame. That makes the output from mold() consistent across all interfaces (i.e. the predictors and outcomes are always data frames).

It would be pretty inconsistent to return a dgCMatrix in the predictors slot even if had performance improvements

[juliasilge]

I looked at this a bit yesterday and realized that it is going to be a significant undertaking. Being able get to something like this:

new_default_recipe_blueprint(
    mold = mold,
    forge = forge,
    intercept = intercept,
    allow_novel_levels = allow_novel_levels,
    composition = composition,
    fresh = fresh
  )

is going to be a significant improvement in quality of life in several domains, though.

[DavisVaughan]

Closed by #150 (closed enough for now, at least). We can revisit in the future as needed

[github-actions]

This issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue.