Add match chains data

[jimmyday12]

Noting down some general steps and thoughts on adding the match chain data from @DataByJosh.

https://github.com/DataByJosh/AFL-Data/blob/main/AFLM_Match_Chains/Scraper.R

I'll assume you are relatively new to package dev based on our convo but if not, you can ignore some of this stuff!

Setup and Overview

• This is a great resource and tends to be how I've tried to develop - https://r-pkgs.org/
• The usethis package is massively helpful to setup a lot of boilerplate
• Making a Pull Request is also great because we can review, comment, edit and update it don't stress too much about the first submission (see here for example - https://github.com/jimmyday12/fitzRoy/pull/159)

General structure

• I suggest trying to follow one of the existing files in the /R folder. In general - each main group of functions lives in one file. And there is a kind of parent function and then each 'source' (afl.com.au, footywire, afltables) has it's own function that the parent function calls
• Have a look through existing ones but the fetch_lineup.R file would be a good example to follow since it only has one source like you'll have here (https://github.com/jimmyday12/fitzRoy/blob/main/R/fetch_lineup.R)
• Try use the same naming structure. Something like fetch_match_chains and fetch_match_chains_afl
• Similarly - try use a similar argument structure for consistency
• You can make a new R file easily with usethis::use_R("file_name.R") and it'll set stuff up
• You can also add stuff to helper files that is only going to be internal. There are already a few helper files setup for each source so you could add things in there if you need (e.g. https://github.com/jimmyday12/fitzRoy/blob/main/R/helpers-afl.R) but you may also want to create a new helper file for stuff just related to this family of functions (e.g. see https://github.com/jimmyday12/fitzRoy/blob/main/R/helpers-afltables-playerstats.R)
• Name your helper functions in a way that makes it obvious what they doing in context of the rest of the package (e.g. instead of get_players you could use fetch_players_afl)

Code stuff

• It can take a little while to get used to writing code for packages
• Main thing to remember is if you are using an external package, you either need to import a particular function directly or use the package name (e.g. dplyr::mutate). I tend to find that the easiest thing
• If you are using an external package like above - it also needs to live in the DESCRIPTION file. Most of the ones you need should be in there but if you need to add another you can use usethis::use_package("pkgname") and it will add it automatically. That does add an external dependancy so I try to not add too many more (and really need to purge some that are there!)
• You need to add some roxygen code above functions to document them properly - good explainer here but also you can copy/borrow from existing files https://r-pkgs.org/man.html
• If you need to reference a column in any kind of dplyr function, use .data$col_name. I'm not exactly sure why but it throws warning messages in tests that CRAN doesn't like (e.g. df %>% dplyr::mutate(season = as.numeric(.data$season)

Tests

• You'll have to write tests - some instructions here - https://r-pkgs.org/tests.html
• Again - you can just copy what I've got. In general - 1 test file per R file you create
• usethis::use_test("name_of_file.R" will set it up for you

Workflow

• The best way to work with the code is following a pattern outlined here - https://r-pkgs.org/r.html#constant-health-checks
• Generally - use load_all regularly to run code and test
• Less frequently but still often, run devtools::test
• After all major work, run devtools::check. The goal here is 0 notes, warnings or errors!
• There is almost always a warning when using bare column names in things like mutate. The way to get around this is to use .data$col_name (for example df %>% dplyr::mutate(season = as.numeric(.data$season)

General comments

• For getting a token you should be able to use get_afl_cookie which is in the AFL helpers file - https://github.com/jimmyday12/fitzRoy/blob/main/R/helpers-afl.R
• Your access_api function is genius - I'd love if you added that to the AFL helpers file and at some point I might re-use that in other functions!

Other stuff

• Add a note to the NEWS.md file outlining stuff you've added
• Add yourself as a contributor
• It might be worth writing a vignette, although that can come a bit later if needed. See https://r-pkgs.org/vignettes.html or check out existing ones (usethis::use_vignette())

[jimmyday12]

Hey @DataByJosh - just reviewing some old tickets and noticed this one. No stress or pressure, just thought I'd ping to see if you had a chance to look into this at all?

[DataByJosh]

Hey @jimmyday12 - apologies, has been a hectic couple of months and just never quite made it to the top of the priorities pile. Moving house at the moment but should have some time to tackle it in February.

[jimmyday12]

No stress at all @DataByJosh - I have been in the same boat. Thanks for the message!