pac_toc_chain: Apply `pac_toc` within a data frame and output a column of `water` class to be chained to other tidywater functions PAC = powdered activated carbon

Description

This function allows pac_toc to be added to a piped data frame. Its output is a `water` class, and can therefore be used with "downstream" tidywater functions.

Usage

pac_toc_chain(
  df,
  input_water = "defined_water",
  output_water = "pac_water",
  dose = 0,
  time = 0,
  type = "bituminous"
)

Value

A data frame containing a water class column with updated DOC, TOC, and UV254 slots

Arguments

df: a data frame containing a water class column, which has already been computed using define_water_chain. The df may include columns named for the dose, time, and type
input_water: name of the column of water class data to be used as the input for this function. Default is "defined_water".
output_water: name of the output column storing updated parameters with the class, water. Default is "pac_water".
dose: Applied PAC dose (mg/L). Model results are valid for doses concentrations between 5 and 30 mg/L.
time: Contact time (minutes). Model results are valid for reaction times between 10 and 1440 minutes
type: Type of PAC applied, either "bituminous", "lignite", "wood".

Details

The data input comes from a `water` class column, as initialized in define_water.

If the input data frame has a dose, time or type column, the function will use those columns. Note: The function can only take dose, time, and type inputs as EITHER a column or from the function arguments, not both.

tidywater functions cannot be added after this function because they require a `water` class input.

For large datasets, using `fn_once` or `fn_chain` may take many minutes to run. These types of functions use the furrr package for the option to use parallel processing and speed things up. To initialize parallel processing, use `plan(multisession)` or `plan(multicore)` (depending on your operating system) prior to your piped code with the `fn_once` or `fn_chain` functions. Note, parallel processing is best used when your code block takes more than a minute to run, shorter run times will not benefit from parallel processing.

Examples

Run this code


library(purrr)
library(furrr)
library(tidyr)
library(dplyr)

example_df <- water_df %>%
  define_water_chain("raw") %>%
  pac_toc_chain(input_water = "raw", dose = 10, time = 20)

example_df <- water_df %>%
  define_water_chain("raw") %>%
  mutate(dose = seq(11, 22, 1), time = 30) %>%
  pac_toc_chain(input_water = "raw")

example_df <- water_df %>%
  define_water_chain("raw") %>%
  mutate(time = 8) %>%
  pac_toc_chain(
    input_water = "raw", dose = 6, type = "wood"
  )

# Initialize parallel processing
plan(multisession, workers = 2) # Remove the workers argument to use all available compute
example_df <- water_df %>%
  define_water_chain("raw") %>%
  pac_toc_chain(input_water = "raw", dose = 4, time = 8)

# Optional: explicitly close multisession processing
plan(sequential)