Refine and refactor UKAQ import functions

This PR fixes #346 plus more. Three key changes are contained within:

When data_type is "annual", "monthly" or "daqi", pollutant and site are now respected. The default behaviour has not changed. This is important if we're trying to make {openair} more accessible to those without R skills (i.e., those who won't be able to filter it after the fact).

library(openair)

importAURN(data_type = "annual")
#> # A tibble: 129 × 38
#>    uka_code code  site    date                   o3 o3_capture o3.summer_capture
#>    <fct>    <chr> <chr>   <dttm>              <dbl>      <dbl>             <dbl>
#>  1 UKA00399 ABD   Aberde… 2009-01-01 00:00:00  42.0      0.944             0.990
#>  2 UKA00513 ABD7  Aberde… 2009-01-01 00:00:00  NA       NA                NA    
#>  3 UKA00451 ACTH  Auchen… 2009-01-01 00:00:00  55.5      0.975             0.953
#>  4 UKA00137 AH    Aston … 2009-01-01 00:00:00  63.8      0.968             0.976
#>  5 UKA00541 ARM6  Armagh… 2009-01-01 00:00:00  NA       NA                NA    
#>  6 UKA00236 BAR2  Barnsl… 2009-01-01 00:00:00  NA       NA                NA    
#>  7 UKA00353 BAR3  Barnsl… 2009-01-01 00:00:00  43.3      0.969             0.997
#>  8 UKA00306 BATH  Bath R… 2009-01-01 00:00:00  NA       NA                NA    
#>  9 UKA00212 BEL2  Belfas… 2009-01-01 00:00:00  37.8      0.936             0.907
#> 10 UKA00238 BEX   London… 2009-01-01 00:00:00  NA       NA                NA    
#> # ℹ 119 more rows
#> # ℹ 31 more variables: o3.daily.max.8hour <dbl>, o3.aot40v <int>,
#> #   o3.aot40f <int>, somo35 <dbl>, somo35_capture <dbl>, no <dbl>,
#> #   no_capture <dbl>, no2 <dbl>, no2_capture <dbl>, nox <dbl>,
#> #   nox_capture <dbl>, so2 <dbl>, so2_capture <dbl>, co <dbl>,
#> #   co_capture <dbl>, pm10 <dbl>, pm10_capture <dbl>, nv10 <dbl>,
#> #   nv10_capture <dbl>, v10 <dbl>, v10_capture <dbl>, pm2.5 <dbl>, …

importAURN(
  data_type = "annual",
  site = c("my1", "kc1"),
  year = 2020,
  pollutant = c("no2", "pm2.5"),
)
#> # A tibble: 2 × 8
#>   uka_code code  site  date                  no2 no2_capture pm2.5 pm2.5_capture
#>   <fct>    <chr> <chr> <dttm>              <dbl>       <dbl> <dbl>         <dbl>
#> 1 UKA00253 KC1   Lond… 2020-01-01 00:00:00  20.6       0.994  8.05         0.996
#> 2 UKA00315 MY1   Lond… 2020-01-01 00:00:00  43.7       0.968  9.00         0.787

importAURN(
  data_type = "annual",
  site = c("my1", "kc1"),
  year = 2020,
  pollutant = c("no2", "pm2.5"),
  to_narrow = TRUE
)
#> # A tibble: 4 × 6
#>   code  site                   date                species value data_capture
#>   <chr> <chr>                  <dttm>              <chr>   <dbl>        <dbl>
#> 1 KC1   London N. Kensington   2020-01-01 00:00:00 no2     20.6         0.994
#> 2 KC1   London N. Kensington   2020-01-01 00:00:00 pm2.5    8.05        0.996
#> 3 MY1   London Marylebone Road 2020-01-01 00:00:00 no2     43.7         0.968
#> 4 MY1   London Marylebone Road 2020-01-01 00:00:00 pm2.5    9.00        0.787

^{Created on 2023-05-26 with reprex v2.0.2}

importAQE() and the other devolved networks can now access DAQI statistics. Trev's mentioned they only exist from 2022 for now, but could be back-dated to 2010 in the future.

library(openair)

importAQE(year = 2022, data_type = "daqi")
#> # A tibble: 138,363 × 7
#>    code  site             pollutant date                concentration poll_index
#>    <chr> <chr>            <chr>     <dttm>                      <dbl>      <int>
#>  1 AB001 VofWH Abingdon … no2       2022-01-01 00:00:00         10.3           1
#>  2 BAI2  Birmingham Airp… no2       2022-01-01 00:00:00         12.0           1
#>  3 BAI2  Birmingham Airp… o3        2022-01-01 00:00:00         59             2
#>  4 BAI2  Birmingham Airp… pm10      2022-01-01 00:00:00         16             1
#>  5 BAI2  Birmingham Airp… pm2.5     2022-01-01 00:00:00         10             1
#>  6 BAI2  Birmingham Airp… so2       2022-01-01 00:00:00          3.04          1
#>  7 BAR11 Barnsley A628 R… no2       2022-01-01 00:00:00         22.1           1
#>  8 BAR9  Barnsley A635 K… pm10      2022-01-01 00:00:00         27             2
#>  9 BN1   Barnet Tally Ho  no2       2022-01-01 00:00:00         26.2           1
#> 10 BN1   Barnet Tally Ho  pm10      2022-01-01 00:00:00         25             2
#> # ℹ 138,353 more rows
#> # ℹ 1 more variable: measurement_period <chr>

^{Created on 2023-05-26 with reprex v2.0.2}

There was a lot of repetition between importAURN(), importAQE(), importNI(), importWAQN() and importSAQN(), so much of their code has been combined into one helper function. The actual import*() functions are now just wrappers around this internal helper and live in a single R script. The key advantages of this are:
- It is easier to update these functions and for them to remain consistent. Indeed, I did this refactor because of how repetitive it was to
- importMeta() can import multiple networks at once, so this is a good first step to having an exported importUKAQ() function which can import sites across, e.g., AURN, AQE and LMAM in one go. When that happens, importAURN() et al. can probably be deprecated, and you'd do something like importUKAQ(source = "aurn", site = "my1", year = 2020) or importUKAQ(source = meta$source, site = meta$site, year = 2020).

Notes:

For now, importLocal() is left on its own. This is because, while it is very similar in a lot of ways, it behaves differently under the hood and has fewer stats than the other functions.
importKcl() and importEurope() are also very different and arguably legacy functions so they've been left alone too.
The non-aq import*() functions like importTraj() are untouched.

davidcarslaw / openair

Refine and refactor UKAQ import functions #347