tests/testthat/test_onestation.r

In mtalluto/NSmetabolism: N-Station Stream Metabolism Modelling

# DEPRECATED TESTS; THESE FUNCTIONS HAVE BEEN REMOVED
# should use method-stan now
# 
# 
# context("One Station")
# library("NSmetabolism")
# library(lubridate)

# disso <- read_minidot(system.file("testdata/metab/minidot", package="NSmetabolism"))
# light <- read_hobo(system.file("testdata/metab/hobo.csv", package="NSmetabolism"))
# ## the elevation and depth at this site
# elev <- 447.3237
# z <- 0.722
# # filter to only two days, april 23 and 24
# startTime <- ymd_hms("2018-04-23 00:00:00", tz = tz(disso$timestamp[1]))
# endTime <-  ymd_hms("2018-04-25 00:00:00", tz = tz(disso$timestamp[1]))
# disso <- disso[disso$timestamp >= startTime & disso$timestamp <  endTime,]
# light <- light[light$timestamp >= startTime & light$timestamp <  endTime,]

# ## approximate conversion for light
# ## from http://bccp.berkeley.edu/o/Academy/workshop08/08%20PDFs/Inv_Square_Law.pdf
# light$light <- light$light * 0.0079

# # add a time column in elapsed minutes
# disso$minutes <- (as.integer(disso$timestamp) - as.integer(startTime)) / 60
# light$minutes <- (as.integer(light$timestamp) - as.integer(startTime)) / 60

# initParams <- c(P1=69911.063207, P2=220.062287, k600 = 2/(24*60), ER24_20=-5.386056, sd=1)



# i <- which.max(light$light)
# j <- which(disso$minutes > light$minutes[i] - 5 & disso$minutes < light$minutes[i] + 5)

# test_that("dDOdt function working for one station", {
# 	DOdata <- list(
# 		PAR = approxfun(x = light$minutes, y = light$light, rule = 2),
# 		temp = approxfun(x = disso$minutes, y = disso$temperature, rule = 2),
# 		P = pressureFromElevation(elev),
# 		z = z)

# 	# dDOdt is negative at night when light == 0 (if no RF); the first point works
# 	initParams2 <- initParams
# 	initParams2['k600'] <- 0
# 	ddo <- oneStation_dDOdt(disso$minutes[1], disso$DO[1], initParams2, DOdata)
# 	expect_lt(ddo, 0)
# 	# expect an absolute value less than the max in the data
# 	expect_lt(abs(ddo), max(disso$DO))
# 	# expect a positive value when light is at it's maximum
# 	ddo <- oneStation_dDOdt(disso$minutes[j], disso$DO[j], initParams, DOdata)
# 	expect_gt(ddo, 0)
# 	# expect an absolute value less than the max in the data
# 	expect_lt(ddo, max(disso$DO))

# })


# test_that("DOpredict working for one station", {
# 	DOdata <- list(
# 		PAR = light[,c('light', 'minutes')],
# 		temp = disso[,c('temperature', 'minutes')],
# 		P = pressureFromElevation(elev),
# 		z = z)

# 	expect_error(doPr_euler <- oneStation_DOPredict(disso$DO[1], c(0, disso$minutes[1],   
# 		disso$minutes[j]), initParams, DOdata, dt = 0.1, method='euler'), regex=NA)
# 	expect_error(doPr_lsoda <- oneStation_DOPredict(disso$DO[1], c(0, disso$minutes[1],
# 		 disso$minutes[j]), initParams, DOdata, method='lsoda'), regex=NA)

# 	# quantitative predictions don't make a ton of sense, but we can try qualitative ones
# 	# similar to each other
# 	expect_lt(sum(abs(doPr_euler - doPr_lsoda)), 1e-2)
# 	# should be close to the initial value (within 5x for sure!)
# 	expect_true(all((doPr_euler / doPr_euler[2]) > 1/5 & (doPr_euler / doPr_euler[2]) < 5))

# 	# strictly nonnegative
# 	expect_true(all(doPr_euler >= 0))
# 	expect_true(all(doPr_lsoda >= 0))
# })