find an automatic way of keeping scenario source labels up to date

[jdhoffa]

          NIT: I think we may just want to functionalize this? But out of scope for this PR since it is a pure refactor

_Originally posted by @jdhoffa in https://github.com/RMI-PACTA/pacta.scenario.data.preparation/pull/6#discussion_r1494496338_

AB#10859

[cjyetman]

We're talking about these lines here? https://github.com/RMI-PACTA/pacta.scenario.data.preparation/blob/6d77aeb79cad080730d3474a76611120357ce288/R/utils.R#L43-L48

What is the suggestion? Something like...

p4i_label <- function() { c("WEO2022", "GECO2022") }
p4b_label <- function() { c("weo_2022", "geco_2022") }

Seems a bit overkill honestly

[jdhoffa]

No, I think the idea is more to create a function that maps: ABCDYYYY -> abcd_yyyy

The former being what P4I expects, the latter being what P4B expects.

(In general, my preference is to just align the expectation from both tools, and not deal with the reformatting at all)

[cjyetman]

relevant work was done in

• 59

• 60

[cjyetman]

There's already dictionary_p4i_p4b(), which was updated in #59 and #60 https://github.com/RMI-PACTA/pacta.scenario.data.preparation/blob/5c3925555358b13e06698a5b3fa30a912775a64c/R/utils.R#L43-L53

Is that enough, or did you want more? like...

dictionary_p4i_p4b <- function() {
  dplyr::tribble(
    ~p4i_label, ~p4b_label,
    "ISF2021", "isf_2021",
    "WEO2022", "weo_2022",
    "GECO2022", "geco_2022",
    "GECO2023", "geco_2023",
    "WEO2023", "weo_2023",
    "ISF2023", "isf_2023"
  )
}

convert_scenario_sources <-
  function(scenario_sources, label = c("p4i_label", "p4b_label")) {
    dict <- dictionary_p4i_p4b()
    to_label <- names(dict)[label != names(dict)]
    dict[[to_label]][match(scenario_sources, dict[[label]])]
  }

scenario_sources <- c("ISF2021", "WEO2022", "GECO2022")
convert_scenario_sources(scenario_sources, label = "p4i_label")
#> [1] "isf_2021"  "weo_2022"  "geco_2022"

scenario_sources <- c("isf_2021", "weo_2022", "geco_2022")
convert_scenario_sources(scenario_sources, label = "p4b_label")
#> [1] "ISF2021"  "WEO2022"  "GECO2022"

[jdhoffa]

The goal of this issue would be to create a function that programmatically maps e.g. ABC1234 -> abc_1234, so that we don't need to manually update the dictionary_p4i_p4b() function whenever we add a new scenario

[cjyetman]

@jdhoffa does this achieve what you were thinking of?

p4i_p4b <- function(x) {
  tolower(sub(pattern = "([:alpha:]*)(\\d{4})", replacement = "\\1_\\2", x = x))
}

p4b_p4i <- function(x) {
  toupper(sub(pattern = "([:alpha:]*)_(\\d{4})", replacement = "\\1\\2", x = x))
}

p4i_p4b("WEO2023")
#> [1] "weo_2023"

p4i_p4b("GECO2023")
#> [1] "geco_2023"

p4b_p4i("weo_2023")
#> [1] "WEO2023"

p4b_p4i("geco_2023")
#> [1] "GECO2023"

[jdhoffa]

It does! Then the next step would be to actually implement that function in any of the P4B processing functions