Add the split transformation

[AlexandreKempf]

Hello everyone,

This is my first PR ever, but I'm willing to learn, so don't hesitate to be picky.

I was doing the advent of code 2022 with a constraint to use only Pyfunctional so that I could learn the library. I was blocked on the first problem because I felt it was missing a split function.

The goal of such a function is to take a sequence and split it into several lists given a function. The function maps over the sequence and splits if it returns True or continues if it returns False. An example is worth a million words, so here it is:

seq([1,2,3,None,4,5,6,7,None,8,9]).split(lambda x: x is None) 
>>> seq([ [1,2,3], [4,5,6,7], [8,9] ])

I tried to respect the rules for PR, but if I missed something feel free to comment ;)

Have a good day!

[EntilZha]

Hi @AlexandreKempf, thanks for the PR! One thing you could look into is using functional.extend to register your own functions https://github.com/EntilZha/PyFunctional/issues/113.

For adding functions to the API, I like to at least know it has precedent elsewhere (e.g., in Apache Spark, Scala collections, Rust iterables, or other similar FP libraries). Do you know if there is a split function somewhere? Not to say it is required, but would be great to know :).

Thanks!

[AlexandreKempf]

Hi EntilZha, and thank you for your reply !

I discovered the extend method earlier today and I was amazed :+1: ! I'm not sure I would have done the PR knowing extend but here we are :p

I couldn't find exactly such functions, but there are similar functions in Elixir (chunk_by) and in toolz (partition_by). However, they are splitting not based on a boolean but on a change of output from the function.

If you're interested in a split function, what behavior would you like to have ?

Best,

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[EntilZha]

Been busy, sorry for late reply. I'm not aware of prior functions that do that, so would lean towards using this via extend for now. If its a pretty common use case, then could reconsider.