updated tests after changing example data

Author: osorensen

tests/testthat/test-all_topological_sorts.R

@@ -19,10 +19,9 @@ test_that("precompute_topological_sorts works", {
 
   expect_equal(
     sorts$sort_matrix,
-    structure(c(1L, 2L, 3L, 5L, 4L, 1L, 2L, 5L, 3L, 4L, 1L, 3L, 5L,
-                2L, 4L, 1L, 5L, 3L, 4L, 2L, 1L, 5L, 4L, 2L, 3L, 1L, 5L, 4L, 3L,
-                2L), dim = 5:6)
+    structure(c(5L, 1L, 2L, 3L, 4L, 5L, 1L, 2L, 4L, 3L, 5L, 1L, 4L,
+                2L, 3L), dim = c(5L, 3L))
   )
 
-  expect_equal(sorts$sort_count, 12L)
+  expect_equal(sorts$sort_count, 8L)
 })

tests/testthat/test-compute_sequentially_complete.R

@@ -7,8 +7,8 @@ test_that("compute_sequentially works with complete data", {
   )
   expect_s3_class(mod, "BayesMallowsSMC2")
   alpha_hat <- weighted.mean(x = as.numeric(mod$alpha), w = mod$importance_weights)
-  expect_gt(alpha_hat, .99)
-  expect_lt(alpha_hat, 1.05)
+  expect_gt(alpha_hat, .03)
+  expect_lt(alpha_hat, .08)
 
   set.seed(2)
   mod <- compute_sequentially(
@@ -18,8 +18,8 @@ test_that("compute_sequentially works with complete data", {
                                   resampler = "residual")
   )
   alpha_hat <- weighted.mean(x = as.numeric(mod$alpha), w = mod$importance_weights)
-  expect_gt(alpha_hat, .99)
-  expect_lt(alpha_hat, 1.06)
+  expect_gt(alpha_hat, .04)
+  expect_lt(alpha_hat, .06)
 
   set.seed(2)
   mod <- compute_sequentially(
@@ -29,8 +29,8 @@ test_that("compute_sequentially works with complete data", {
                                   resampler = "stratified")
   )
   alpha_hat <- weighted.mean(x = as.numeric(mod$alpha), w = mod$importance_weights)
-  expect_gt(alpha_hat, .99)
-  expect_lt(alpha_hat, 1.1)
+  expect_gt(alpha_hat, .02)
+  expect_lt(alpha_hat, .05)
 
   set.seed(2)
   mod <- compute_sequentially(
@@ -40,6 +40,6 @@ test_that("compute_sequentially works with complete data", {
                                   resampler = "systematic")
   )
   alpha_hat <- weighted.mean(x = as.numeric(mod$alpha), w = mod$importance_weights)
-  expect_gt(alpha_hat, .99)
-  expect_lt(alpha_hat, 1.06)
+  expect_gt(alpha_hat, .02)
+  expect_lt(alpha_hat, .05)
 })

tests/testthat/test-compute_sequentially_partial.R

@@ -8,8 +8,8 @@ test_that("compute_sequentially works with partial data", {
       max_particle_filters = 30, max_rejuvenation_steps = 5)
   )
   alpha_hat <- weighted.mean(x = as.numeric(mod$alpha), w = mod$importance_weights)
-  expect_gt(alpha_hat, .95)
-  expect_lt(alpha_hat, 1.06)
+  expect_gt(alpha_hat, .06)
+  expect_lt(alpha_hat, .09)
 
   set.seed(2)
   mod <- compute_sequentially(
@@ -20,8 +20,8 @@ test_that("compute_sequentially works with partial data", {
       resampler = "residual", max_rejuvenation_steps = 5)
   )
   alpha_hat <- weighted.mean(x = as.numeric(mod$alpha), w = mod$importance_weights)
-  expect_gt(alpha_hat, .99)
-  expect_lt(alpha_hat, 1.06)
+  expect_gt(alpha_hat, .02)
+  expect_lt(alpha_hat, .05)
 
   set.seed(2)
   mod <- compute_sequentially(
@@ -35,8 +35,8 @@ test_that("compute_sequentially works with partial data", {
       max_rejuvenation_steps = 5)
   )
   alpha_hat <- weighted.mean(x = as.numeric(mod$alpha), w = mod$importance_weights)
-  expect_gt(alpha_hat, .94)
-  expect_lt(alpha_hat, 1.05)
+  expect_gt(alpha_hat, .02)
+  expect_lt(alpha_hat, .05)
 
   set.seed(2)
   mod <- compute_sequentially(
@@ -50,8 +50,8 @@ test_that("compute_sequentially works with partial data", {
       max_rejuvenation_steps = 5)
   )
   alpha_hat <- weighted.mean(x = as.numeric(mod$alpha), w = mod$importance_weights)
-  expect_gt(alpha_hat, .99)
-  expect_lt(alpha_hat, 1.08)
+  expect_gt(alpha_hat, .02)
+  expect_lt(alpha_hat, .05)
 })
 
 test_that("compute_sequentially works with partial data and pseudolikelihood proposal", {
@@ -67,7 +67,7 @@ test_that("compute_sequentially works with partial data and pseudolikelihood pro
       latent_rank_proposal = "pseudo")
   )
   alpha_hat <- weighted.mean(x = as.numeric(mod$alpha), w = mod$importance_weights)
-  expect_gt(alpha_hat, .98)
-  expect_lt(alpha_hat, 1.06)
+  expect_gt(alpha_hat, .05)
+  expect_lt(alpha_hat, .1)
 
 })

tests/testthat/test-compute_sequentially_preferences.R

@@ -23,8 +23,8 @@ test_that("compute_sequentially works with preference data", {
     topological_sorts = topological_sorts
   )
 
-  expect_gt(mean(mod$alpha), 1)
-  expect_lt(mean(mod$alpha), 1.02)
+  expect_gt(mean(mod$alpha), .2)
+  expect_lt(mean(mod$alpha), .3)
 })
 
 test_that("compute_sequentially works with preference data and tracing", {
@@ -55,8 +55,8 @@ test_that("compute_sequentially works with preference data and tracing", {
 
   expect_equal(length(mod$alpha_traces), 3)
   expect_equal(length(mod$alpha_traces[[2]]), 100)
-  expect_gt(mod$alpha_traces[[2]][[3]], .24)
-  expect_lt(mod$alpha_traces[[2]][[3]], .25)
+  expect_gt(mod$alpha_traces[[2]][[3]], .49)
+  expect_lt(mod$alpha_traces[[2]][[3]], .51)
 
   set.seed(3)
   mod <- compute_sequentially(
@@ -72,13 +72,13 @@ test_that("compute_sequentially works with preference data and tracing", {
 
   expect_equal(length(mod$alpha_traces), 3)
   expect_equal(length(mod$alpha_traces[[2]]), 100)
-  expect_gt(mod$alpha_traces[[2]][[3]], .04)
-  expect_lt(mod$alpha_traces[[2]][[3]], .05)
+  expect_gt(mod$alpha_traces[[2]][[3]], .4)
+  expect_lt(mod$alpha_traces[[2]][[3]], .5)
 
   expect_equal(length(mod$latent_rankings_traces), 3)
   expect_equal(length(mod$latent_rankings_traces[[2]]), 100)
   expect_equal(
     mod$latent_rankings_traces[[2]][[3]],
-    c(1, 3, 5, 4, 2, 2, 1, 4, 3, 5)
+    c(5, 4, 1, 3, 2, 5, 4, 2, 3, 1)
   )
 })