Expose find_chain_start() for View() clones

[gvelasq]

This is an amazing and time-saving innovation, thank you.

Would you consider exposing find_chain_start()? I'm specifically thinking of the use case for View() clones like my br() function, which I pepper into EDA chains.

It would be great to lift the beginning of the chain for a programmatically-generated title as opposed to .(e.g., something like mtcars (piped).

[MilesMcBain]

Happy to expose anything you think you'll find helpful but be aware this is probably a never-CRAN package.

[MilesMcBain]

Actually I decided I might try for CRAN. We'll see what happens.

[gvelasq]

Great! Fingers crossed. If it makes it on CRAN I would be happy to use {breakerofchains} as a dependency. Otherwise if it will remain on GitHub I'm happy to adapt your code directly with attribution/thanks, since we are both releasing under MIT licenses.

Bigger picture, this may be out of the scope of what you were going for, but I think it would be interesting to have something like a {chaintools} where you can print the AST of a chain, find the objects in first/second/n links in the chain, the verbs that bind them, break the chain as you implemented, or even pause a chain to be replayed/reused later (like a superpowered %T>%).

A good first step would be to improve View(), tibble:view(), and friends including the VS Code version to be more informative when they are piped into. A more involved idea than mtcars (piped) would be to identify the chain steps, e.g. mtcars>filter>select>count as the title for a piped View().