Call for co-maintainers :-)

[maelle]

On behalf of @tonyfischetti 😉, cc @ropensci/admin

Co-maintainer(s) would take some of the load off submitted issues. 😸

[phewson]

Are you still looking for co-maintainers, and if so what would be good issues to look at? I was interested in #130 but saw there were a few opinions around that (for example, I quite liked the idea of yaml but I wouldn't for one minute suggest my way of handling meta-data was correct or optimal).

[maelle]

Pinging @tonyfischetti

[tonyfischetti]

@phewson Yes :) And I'm glad you're interested in #130 . Do you think you can make a small example of what the yaml might look like? If we like it, we can go ahead with development.

Another thing to think about is whether assertr could/should read user-defined $metadata fields in JSON schema documents and be able to translate those directives into an assertr chain. Just throwing a crazy idea out there

[phewson]

I'm using something like this as a config.yaml. I added three lines at the top to illustrate roughly what I was thinking of with assertr:

mtcars:
  min_mpg: 0
  max_mpg: Inf
source: https://www.data.gov.uk/dataset/cb7ae6f0-4be6-4935-9277-47e5ce24a11f/road-accidents-safety-data
downloaded: 14/2/2025
publisher: Department for Transport 
source_update: 27/1/2025 
licence: Open Government Licence
summary:  |
  These files provide detailed road safety data about the circumstances of personal injury road collisions in Great Britain from 1979, the types of vehicles involved and the consequential casualties. The statistics relate only to personal injury collisions on public roads that are reported to the police, and subsequently recorded, using the STATS19 collision reporting form. This data contains all the non-sensitive fields that can be made public. Sensitive data fields, for example contributory factors data, can be requested by completing the sensitive data form and contacting the road safety statistics team at roadacc.stats@dft.gov.uk  All the data variables are coded rather than containing textual strings. The lookup tables are available in the supporting documents section towards the bottom of the table. Data relating to the casualty and collision severity adjustment to account for changes in police reporting of severity is provided in separate files and can be joined using the appropriate record identifiers.

  Final annual data is released annually in late September following the publication of the annual reported road casualties Great Britain statistical publication. Individual years data is available for each of the last 5 years, with earlier years available as part of a single download.
  
  In addition, un-validated provisional mid-year data (covering January to June) is released at end November, to provide more up to date information
  
  Data revisions
  
  Except for the severity adjustments, data are not routinely revised those occasionally minor amendments to previous years can be made. Details of recent revisions are available, together with a request for any feedback on the approach to revising the data. The files published here represent the latest data.

As it happens, that needs a couple of helper functions to turn "Inf" into Inf, e.g.,

convert_specials <- function(x) {
  if (is.character(x)) {
    if (x == "Inf") return(Inf)
    if (x == "-Inf") return(-Inf)
    if (x == "NaN") return(NaN)
  }
  return(x)
}

convert_list_specials <- function(lst) {
  lapply(lst, function(x) {
    if (is.list(x)) {
      convert_list_specials(x)
    } else {
      convert_specials(x)
    }
  })
}

(I haven't thought that through at all, I never set Inf as a bound like that, I was just trying to make the examples from the vignette work). Basically I can read and convert that yaml with config <- yaml::read_yaml(here::here("config.yaml")) and config <- convert_list_specials(config) which gives me (at the moment) some clunky ways of coding from the config file e.g.

our.data %>%
  assert(within_bounds(config$mtcars$min_mpg, config$mtcars$max_mpg), mpg) %>%
  group_by(cyl) %>%
  summarise(avg.mpg=mean(mpg))

As it happens, I'm no longer so sure I want to put everything in a single everything config file (yaml or otherwise), so would definitely consider json instead.

I'm guessing it would be tidier if assert could look for the config values for data frame / tibble our.data in a section of the config called our.data so you don't have to type config$mtcars$max_mpg but I haven't looked at the code a while yet and can't really remember what train of thought I was on back then.