feat: allow UKAQ to guess source of site

[jack-davison]

This PR extends the capabilities introduced in #350.

In short, the default source in importUKAQ() is now NULL. This allows importUKAQ() to work out the source on its own, where possible.

See examples:

devtools::load_all()
#> ℹ Loading openair

set.seed(123)

meta_sample <-
  importMeta(year = 2023, source = "ukaq") |>
  dplyr::slice_sample(n = 1, by = source)

meta_sample$code
#> [1] "TALL"  "WL1"   "NL11"  "NPT4A" "BEL5"  "WA7"
meta_sample$source
#> [1] "aurn"  "aqe"   "saqn"  "waqn"  "ni"    "local"

aq <-
  importUKAQ(
    year = 2023,
    site = meta_sample$code
  )

dplyr::count(aq, source, site, code)
#> # A tibble: 5 × 4
#>   source site                            code      n
#>   <fct>  <chr>                           <chr> <int>
#> 1 aurn   Tallington                      TALL    504
#> 2 saqn   N Lanarkshire Kirkshaws         NL11   8760
#> 3 aqe    Waltham Forest Dawlish Rd       WL1    8760
#> 4 ni     Belfast Newtownards Road        BEL5   8760
#> 5 local  Wandsworth - Putney High Street WA7    8760

^{Created on 2024-08-12 with reprex v2.1.1}

Caveats:

• Will error if the code is unknown.

• Will error if the code is ambiguous ("AD1").

devtools::load_all()
#> ℹ Loading openair

# ambiguous codes
importUKAQ(c("my1", "kc1", "ad1"))
#> Error in `guess_source()` at openair/R/importUKAQ.R:215:7:
#> ✖ Ambiguous site codes detected. Please specify source in
#>   `importUKAQ()`.
#> ℹ Ambiguous codes: AD1 (saqn, local)

# non existant codes
importUKAQ(c("my1", "kc1", "foo", "bar", "falk"))
#> Error in `guess_source()` at openair/R/importUKAQ.R:215:7:
#> ✖ Unknown site codes detected. Please ensure all site codes can be found
#>   in `importMeta()`.
#> ℹ Unknown site codes: FOO and BAR

^{Created on 2024-08-12 with reprex v2.1.1}

Ambiguous codes are few and far between, so it may just have to be an unfortunate feature of the function for those local to "AD1" for now! Unless you can think of a better treatment of them.

[jack-davison]

Update - now just prefers AURN to LMAM data.

devtools::load_all()
#> ℹ Loading openair

importUKAQ(c("ad1", "el3"))
#> Warning: ✖ Ambiguous site codes detected. National networks are imported preferentially
#>   to locally managed networks.
#> ! AD1 (could also be 'Adur - Shoreham-by-Sea' from the source: 'local'.)
#> ! EL3 (could also be 'Elmbridge Hampton Court Parade' from the source:
#>   'local'.)
#> ℹ Specify source to import sites from specific monitoring networks.
#> # A tibble: 17,520 × 12
#>    source site     code  date                  nox   no2    no  pm10 pm2.5    ws
#>    <fct>  <chr>    <chr> <dttm>              <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#>  1 saqn   Aberdee… AD1   2022-01-01 00:00:00    NA    NA    NA  19.0 14.2    4.7
#>  2 saqn   Aberdee… AD1   2022-01-01 01:00:00    NA    NA    NA  14.6  8.76   4.6
#>  3 saqn   Aberdee… AD1   2022-01-01 02:00:00    NA    NA    NA  19.6  9.44   5.9
#>  4 saqn   Aberdee… AD1   2022-01-01 03:00:00    NA    NA    NA  14.6  7.86   6.4
#>  5 saqn   Aberdee… AD1   2022-01-01 04:00:00    NA    NA    NA  13.6  7.74   8  
#>  6 saqn   Aberdee… AD1   2022-01-01 05:00:00    NA    NA    NA  15.9  8.93   9.5
#>  7 saqn   Aberdee… AD1   2022-01-01 06:00:00    NA    NA    NA  21.1 11.9   10.6
#>  8 saqn   Aberdee… AD1   2022-01-01 07:00:00    NA    NA    NA  22.8 12.6   10  
#>  9 saqn   Aberdee… AD1   2022-01-01 08:00:00    NA    NA    NA  23.5 13.3    8.9
#> 10 saqn   Aberdee… AD1   2022-01-01 09:00:00    NA    NA    NA  25.6 13.3   10.8
#> # ℹ 17,510 more rows
#> # ℹ 2 more variables: wd <dbl>, air_temp <dbl>

^{Created on 2024-08-12 with reprex v2.1.1}
12 with reprex v2.1.1