# Tests for curve_interval # # Author: mjskay ############################################################################### suppressPackageStartupMessages({ library(dplyr) library(tidyr) }) test_that("curve_interval works with lineribbon", { skip_if_no_vdiffr() skip_if_not_installed("posterior") k = 11 # number of curves n = 101 curve_df = tibble( .draw = 1:k, mean = seq(-5,5, length.out = k), x = list(seq(-15,15,length.out = n)), ) %>% unnest(x) %>% mutate(y = dnorm(x, mean, 3)/max(dnorm(x, mean, 3))) %>% group_by(x) vdiffr::expect_doppelganger("curve_interval with mhd", curve_df %>% curve_interval(y, .width = c(.5, .8), .interval = "mhd") %>% ggplot(aes(x = x, y = y)) + geom_lineribbon(aes(ymin = .lower, ymax = .upper)) + geom_line(aes(group = .draw), alpha = 0.15, data = curve_df) + scale_fill_brewer() ) # conditioning works set.seed(1234) b_draw = sample.int(n) ab_curve_df = bind_rows( mutate(curve_df, group = "a"), mutate(curve_df, group = "b", .draw = b_draw[.draw]) ) %>% arrange(.draw) vdiffr::expect_doppelganger("conditional curve_interval with mhd", ab_curve_df %>% group_by(group) %>% curve_interval(y, .along = x, .width = c(.5, .8), .interval = "mhd") %>% ggplot(aes(x = x, y = y)) + geom_lineribbon(aes(ymin = .lower, ymax = .upper)) + geom_line(aes(group = .draw), alpha = 0.15, data = curve_df) + scale_fill_brewer() + facet_wrap(~ group) ) # joint works # TODO: fix # vdiffr::expect_doppelganger("joint curve_interval with mhd", # ab_curve_df %>% # arrange(.draw) %>% # TODO: this should not be needed # curve_interval(y, .along = c(group, x), .width = c(.5, .8), .interval = "mhd") %>% # ggplot(aes(x = x, y = y)) + # geom_lineribbon(aes(ymin = .lower, ymax = .upper)) + # geom_line(aes(group = .draw), alpha = 0.15, data = curve_df) + # scale_fill_brewer() + # facet_wrap(~ group) # ) skip_if_not_installed("fda") vdiffr::expect_doppelganger("curve_interval with bd-mbd", curve_df %>% curve_interval(y, .width = c(.5, .8), .interval = "bd-mbd") %>% ggplot(aes(x = x, y = y)) + geom_lineribbon(aes(ymin = .lower, ymax = .upper)) + geom_line(aes(group = .draw), alpha = 0.15, data = curve_df) + scale_fill_brewer() ) }) # basic cases: 95%, 0%, 100%, > 100% --------------------------------------------- test_that("basic cases on single interval work", { df = data.frame(value = ppoints(1000)) ref = tibble( value = 0.4995, .lower = NA, .upper = NA, .actual_width = NA, .width = NA, .point = "mhd", .interval = "mhd" ) expect_equal(curve_interval(df, .width = .95), mutate(ref, .lower = .0255, .upper = .9745, .actual_width = .95, .width = .95)) expect_equal(curve_interval(df, .width = 0), mutate(ref, .lower = 0.4995, .upper = .5005, .actual_width = 0.002, .width = 0)) expect_equal(curve_interval(df, .width = 1), mutate(ref, .lower = 0.0005, .upper = .9995, .actual_width = 1, .width = 1)) expect_equal(curve_interval(df, .width = 1.1), mutate(ref, .lower = 0.0005, .upper = .9995, .actual_width = 1, .width = 1.1)) }) test_that("basic cases on single curve work", { df = data.frame(x = 1:3, y = rep(ppoints(1000), each = 3) + 1:3) ref = tibble( x = rep(1:3, 3), y = rep(0.4995 + 1:3, 3), .lower = rep(c(.0255, 0.4995, 0.0005), each = 3) + rep(1:3, 3), .upper = rep(c(.9745, 0.5005, 0.9995), each = 3) + rep(1:3, 3), .actual_width = rep(c(.95, 0.002, 1), each = 3), .width = rep(c(.95, 0, 1), each = 3), .point = "mhd", .interval = "mhd" ) # data frame of draws expect_equal(curve_interval(df, .along = x, .width = c(.95, 0, 1)), ref) expect_equal(curve_interval(group_by(df, x), .width = c(.95, 0, 1)), ref) skip_if_not_installed("posterior") # data frame of rvars y_rvar = rep(posterior::rvar(ppoints(1000)), 3) + 1:3 df = data.frame(x = 1:3, y = y_rvar) expect_equal(curve_interval(df, .along = x, .width = c(.95, 0, 1)), ref) expect_equal(curve_interval(group_by(df, x), .width = c(.95, 0, 1)), ref) # rvar expect_equal(curve_interval(y_rvar, .width = c(.95, 0, 1)), select(ref, -x, .value = y)) # matrix expect_equal(curve_interval(posterior::draws_of(y_rvar), .width = c(.95, 0, 1)), select(ref, -x, .value = y)) }) test_that("basic cases on multiple variables", { df = data.frame( x = 1:3, y1 = rep(ppoints(1000), each = 3) + 1:3, y2 = rep(rev(ppoints(1000)), each = 3) + 2:4 ) ref = tibble( x = rep(1:3, 3), y1 = rep(0.4995 + 1:3, 3), y1.lower = rep(c(.0255, 0.4995, 0.0005), each = 3) + rep(1:3, 3), y1.upper = rep(c(.9745, 0.5005, 0.9995), each = 3) + rep(1:3, 3), .actual_width = rep(c(.95, 0.002, 1), each = 3), .width = rep(c(.95, 0, 1), each = 3), y2 = rep(0.5005 + 2:4, 3), y2.lower = y1.lower + 1, y2.upper = y1.upper + 1, .point = "mhd", .interval = "mhd" ) expect_equal(curve_interval(df, .along = x, .width = c(.95, 0, 1)), ref) expect_equal(curve_interval(group_by(df, x), y1, y2, .width = c(.95, 0, 1)), ref) }) # errors ------------------------------------------------------------------ test_that("error is thrown when no columns found to summarize", { df = data.frame(value = ppoints(10)) expect_error(curve_interval(df, .exclude = "value"), "No columns found to calculate point and interval summaries for") }) test_that("error is thrown with groups of different sizes", { df = data.frame(value = ppoints(9), group = c("a", "a", "b")) expect_error(curve_interval(df, .along = group), "Must have the same number of values in each group") }) test_that("curve_interval() and curve_interval() do not support along", { expect_error( curve_interval(matrix(1:4, nrow = 2), .along = "x"), 'does\\s+not\\s+support\\s+the\\s+[^a-zA-Z]*\\.along[^a-zA-Z]*\\s+argument' ) skip_if_not_installed("posterior") expect_error( curve_interval(posterior::rvar(), .along = "x"), 'does\\s+not\\s+support\\s+the\\s+[^a-zA-Z]*\\.along[^a-zA-Z]*\\s+argument' ) })