copy in most unexported funs and hack environment

[brshallo]

Should fix #9 .

Copy in most exported functions from @carlomedina original fork carlomedina/dplyr. Most of these are in R/context.R

This on it's own did not fix things, also did hack that sets mask to dplyr (see pwiser:::context_peek_bare()). This is an absolute hack and means that checks on whether pairwise() is being used in appropriate contexts is no longer being checked...

Also updated README to add note about overlap with more mature package {dplyover} and limitations (pointing to #1).

[IndrajeetPatil]

Awesome, thanks!

Didn't know about dplyover, will check it out.

Are you planning to submit this package to CRAN at some point?

[brshallo]

Are you planning to submit this package to CRAN at some point?

Initially I was thinking after getting to issues # 1, 2 and (maybe) 8 might take steps to set-up with CRAN. (Note though that # 1 is a hurdle as current set-up is just modifying copied code from {dplyr} and neither Carlo nor I really have a firm understanding of how masking etc is working -- e.g. see hack indicated above 😅. So were debating rewriting removing dependencies on unexported funs and other things. (Shared publicly perhaps a bit prematurely but figured may as well see if people like idea of and get some feedback -- thanks for opening issue!)

After discovering {dplyover} am a bit less sure what will do with {pwiser} (for reasons now noted in README) and am starting talks with Tim (owner of package). My plan was in the next few weeks to do a deep dive on {dplyover}. If works the way I expect it to, I will likely just devote energy there and see if can help with some of the documentation and maybe giving feedback on interface etc. Tim is a lot further along and has a lot more comprehensive set of functionality (which I think has the possibility of becoming a relatively popular + useful {dplyr} extension.)

But will have a more firm idea after talking with Tim and reviewing dplyover a bit more.

[IndrajeetPatil]

Makes sense!

If there is an overlap of functionality and none of the packages are on CRAN, maybe it makes more sense to join forces and combine functionalities. :)

Either way, looking forward to using them in my workflow!

[brshallo]

Yeap, pwiser::pairwise() essentially represents a particular use of dplyover::across2(). Yes, Tim has plans of getting {dplyover} on CRAN, will likely devote energy there.