library(tidyverse)


site_df <- readr::read_tsv(
  snakemake@input[["station"]],
  col_types = cols(
    MonitoringLocationIdentifier = "f",
    MonitoringLocationTypeName = "f",
    MonitoringLocationName = "f",
    MonitoringLocationDescriptionText = "c",
    LatitudeMeasure = "d",
    LongitudeMeasure = "d",
    .default = "-")
) %>%
  rename(
    location = MonitoringLocationIdentifier,
    location_type = MonitoringLocationTypeName,
    location_name = MonitoringLocationName,
    location_desc = MonitoringLocationDescriptionText,
    lat = LatitudeMeasure,
    long = LongitudeMeasure,
    )

# This has a bunch of crap in it that has nothing to do with chemicals in
# water (which might make amphibians gay). Additionally, there are many
# different units that need to be standardized (eventually to be put in
# terms of ug/ml)
result_df <- readr::read_tsv(
  snakemake@input[["results"]],
  col_types = cols(
    ActivityTypeCode = "f",
    ActivityStartDate = "D",
    ActivityEndDate = "D",
    ActivityMediaName = "f",
    MonitoringLocationIdentifier = "f",
    CharacteristicName = "f",
    ResultMeasureValue = "c",
    "ResultMeasure/MeasureUnitCode" = "c",
    "DetectionQuantitationLimitMeasure/MeasureValue" = "c",
    "DetectionQuantitationLimitMeasure/MeasureUnitCode" = "f",
    .default = "-",
  )
) %>%
  rename(activity = ActivityTypeCode,
         unit = "ResultMeasure/MeasureUnitCode",
         media = ActivityMediaName,
         start = ActivityStartDate,
         end = ActivityEndDate,
         location = MonitoringLocationIdentifier,
         value = ResultMeasureValue,
         limit = "DetectionQuantitationLimitMeasure/MeasureValue",
         limit_unit = "DetectionQuantitationLimitMeasure/MeasureUnitCode",
         species = CharacteristicName) %>%
  arrange(start) %>%
  filter(media == "Water") %>%
  select(-media) %>%
  # select values that are numbers (some are just descriptive strings) assuming
  # blanks are actually zeros (to be filtered out later)
  replace_na(list(value = "0")) %>%
  filter(str_detect(value, "^-?\\d+\\.?\\d*$")) %>%
  mutate(value = as.numeric(value)) %>%
  # select units that are mass concentrations or "counts per something"
  filter(str_detect(unit, "^.*g/.*(l|L|g)$") |
           unit %in% c("%", "ppb", "ppm")) %>%
  # remove a bunch of crap with "%" units
  filter(! str_detect(species, "SSC|Cloud cover|Sediment|solids|demand")) %>%
  filter(! str_detect(species, "Pha?eophytin|Chlorophyll|Alkalinity")) %>%
  filter(species != "Sodium, percent total cations" # not sure what this means
         & species != "Dissolved oxygen saturation" # whatever
         & species != "Water" # ironic...
         & species != "Barometric pressure" # not a chemical
         & species != "Relative humidity" # not a chemical either
         & species != "Extract volume" # ''
         & species != "Volume, total" # ''
         & species != "Acidity, (H+)" # will change later
         & species != "Carbon dioxide" # ditto
         & species != "Dissolved oxygen (DO)" # ditto
         & species != "Total hardness" # not specific
         ) %>%
  # these seems like a typos
  mutate(species = case_when(
    species == "Diazinon0" ~ "Diazinon",
    species == "Phosphate-phosphorus***retired***use Total Phosphorus, mixed forms" ~ "Total Phosphorus, mixed forms",
    species == "Inorganic nitrogen (nitrate and nitrite) ***retired***use Nitrate + Nitrite" ~ "Nitrate + Nitrite",
    TRUE ~ species
  )) %>%
  # collapse units to (f/n/u/m)g/L
  mutate(unit = str_replace(unit, "/l", "/L"),
         unit = str_replace(unit, "kg", "L"),
         unit = case_when(unit == "ppb" ~ "ug/L",
                          unit == "ppm" ~ "mg/L",
                          TRUE ~ unit),
         # percent will just be in mg/L
         value = if_else(unit == "%", value * 10, value),
         unit = if_else(unit == "%", "mg/L", unit)) %>%
  # standardize all values to ug/L
  mutate(std_value = case_when(unit == "g/L" ~ value * 1e6,
                               unit == "mg/L" ~ value * 1e3,
                               unit == "ng/L" ~ value / 1e3,
                               unit == "fg/L" ~ value / 1e6,
                               TRUE ~ value)) %>%
  select(-value, -unit)

result_df %>%
  left_join(site_df, by = "location") %>%
  select(-location) %>%
  readr::write_tsv(snakemake@output[[1]])