test_that("datSumTree Aboveground Biomass w/out Extrapolation", { skip_on_cran() # Using datSumTree - nonTPA test1 <- datSumTree(tree = FIESTA::WYtree, ## tsumvarlst = "DRYBIO_AG", ##tsumvarlst == tree-level variables to aggregate: in this case it is aboveground biomass lbs2tons = FALSE, ##if TRUE converts biomass or carbon variables from lbs to tons TPA = FALSE, ##Trees per acre - tfilter = "STATUSCD == 1") ## Set to alive trees treedat_nonTPA <- test1$treedat # Using datSumTree for one plot CN <- 40404730010690 ## Randomly selecting one plot ? How to best go about selecting this? treedat_nonTPA[treedat_nonTPA$PLT_CN == CN,] treedat_nonTPA[treedat_nonTPA$PLT_CN == CN,][[2]]#Selecting a single plot within the treedat df, selecting second variable which is the # Using WYplot data input1 <- WYtree[WYtree$PLT_CN == CN, ] sum(input1[input1$STATUSCD == 1, "DRYBIO_AG"], na.rm = TRUE) # Expect datSumTree & WYplot data match expect_equal(sum(input1[input1$STATUSCD == 1, "DRYBIO_AG"], na.rm=TRUE), treedat_nonTPA[treedat_nonTPA$PLT_CN == CN, ][[2]]) #expect_snapshot(treedat_nonTPA) # Using datSumTree - TPA test2 <- datSumTree(WYtree, tsumvarlst = "DRYBIO_AG", lbs2tons = FALSE, TPA = TRUE, tfilter = "STATUSCD == 1") treedat_TPA <- test2$treedat # Test output for 1 plot CN <- 40404730010690 treedat_TPA[treedat_TPA$PLT_CN == CN, ] [[2]] input1 <- WYtree[WYtree$PLT_CN == CN & WYtree$STATUSCD == 1, ] output1 <- with(input1, sum(DRYBIO_AG * TPA_UNADJ, na.rm=TRUE)) # Expecting same output when summed for TPA expect_equal(output1, treedat_TPA[treedat_TPA$PLT_CN == CN, ][[2]]) #expect_snapshot(treedat_TPA) })