endpoint.Rd
R6 class objects of class chent represent chemical entities and can hold a list of information loaded from a chemical yaml file in their chyaml field. Such information is extracted and optionally aggregated by this function.
endpoint( chent, medium = "soil", type = c("degradation", "sorption"), lab_field = c(NA, "laboratory", "field"), redox = c(NA, "aerobic", "anaerobic"), value = c("DT50ref", "Kfoc", "N"), aggregator = geomean, raw = FALSE, signif = 3 ) soil_DT50( chent, aggregator = geomean, signif = 3, lab_field = "laboratory", value = "DT50ref", redox = "aerobic", raw = FALSE ) soil_Kfoc(chent, aggregator = geomean, signif = 3, value = "Kfoc", raw = FALSE) soil_N(chent, aggregator = mean, signif = 3, raw = FALSE) soil_sorption( chent, values = c("Kfoc", "N"), aggregators = c(Kfoc = geomean, Koc = geomean, N = mean), signif = c(Kfoc = 3, N = 3), raw = FALSE )
chent | The chent object to get the information from |
---|---|
medium | The medium for which information is sought |
type | The information type |
lab_field | If not NA, do we want laboratory or field endpoints |
redox | If not NA, are we looking for aerobic or anaerobic data |
value | The name of the value we want. The list given in the usage section is not exclusive |
aggregator | The aggregator function. Can be mean,
|
raw | Should the number(s) be returned as stored in the chent object (could be a character value) to retain original information about precision? |
signif | How many significant digits do we want |
values | The values to be returned |
aggregators | A named vector of aggregator functions to be used |
The result from applying the aggregator function to the values converted to a numeric vector, rounded to the given number of significant digits, or, if raw = TRUE, the values as a character value, retaining any implicit information on precision that may be present.
The functions soil_*
are functions to extract soil specific endpoints.
For the Freundlich exponent, the capital letter N
is used in order to
facilitate dealing with such data in R. In pesticide fate modelling, this
exponent is often called 1/n.