Merge pull request #50 from Lifebrain/develop

fixed bug with special characters in variable names

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: metagam
 Type: Package
 Title: Meta-Analysis of Generalized Additive Models
-Version: 0.3.0
+Version: 0.3.0.9000
 Authors@R: c(person("Oystein", "Sorensen", 
                     email = "oystein.sorensen@psykologi.uio.no", 
                     role = c("aut", "cre"),

R/metagam.R

@@ -134,8 +134,11 @@ metagam <- function(models, grid = NULL, grid_size = 100, type = "iterms", terms
       ),
     timevar = "term",
     direction = "long",
-    sep = "_")
+    sep = "_",
+    split = list(regexp = "_", include = TRUE, fixed = TRUE)
+    )
   cohort_estimates$id <- NULL
+  names(cohort_estimates)[names(cohort_estimates) %in% c("estimate_", "se_")] <- c("estimate", "se")
   cohort_estimates$model <- as.factor(cohort_estimates$model)
 
   # Now nest the estimates at each grid point

news.md

@@ -1,3 +1,7 @@
+# metagam version 0.3.0.9000
+
+- Fixed a bug which caused metagam to crash when there were special characters (in particular '_' and '.') in variable names.
+
 # metagam version 0.3.0
 
 - This is a major revision of the package containing the changes described since 0.2.0 below.

tests/testthat/test-metagam.R

@@ -15,102 +15,40 @@ test_that("metagam works", {
   metafits <- metagam(fits, grid_size = 10)
   expect_s3_class(metafits, "metagam")
   expect_equal(metafits$terms, "s(x0)")
-  # # Check that predictions are correct
-  # # Vector to compare with has been generate by dput in a manual first run
-  # expect_equal(round(metafits$meta_estimates$estimate, 8),
-  #              c(-0.22194647, -0.13913644, -0.04638216, 0.04226372, 0.10112179,
-  #                0.11856284, 0.09830259, 0.04701996, -0.03726664, -0.14627162))
-  #
-  # # Check that standard errors are correct
-  # expect_equal(round(metafits$meta_estimates$se, 8),
-  #              c(0.42231386, 0.32479295, 0.26294788, 0.23210114, 0.22074271,
-  #                0.22037242, 0.23108217, 0.26216805, 0.32538673, 0.42494948))
+
 
   # Now with second term
   metafits <- metagam(fits, grid_size = 10, terms = "s(x1)")
   expect_s3_class(metafits, "metagam")
   expect_equal(metafits$terms, "s(x1)")
 
-  # # Check that predictions are correct
-  # # Vector to compare with has been generate by dput in a manual first run
-  # expect_equal(round(metafits$meta_estimates$estimate, 8),
-  #              c(-1.90254055, -1.54695893, -1.18290861, -0.74264595, -0.28591313,
-  #                0.40713456, 0.76885555, 1.37551416, 2.08311411, 2.8655438))
-  #
-  # # Check that standard errors are correct
-  # expect_equal(round(metafits$meta_estimates$se, 8),
-  #              c(0.45711194, 0.34464661, 0.28070113, 0.25022086, 0.24067681,
-  #                0.24936844, 0.27250813, 0.3185325, 0.3812979, 0.50396543))
+
 
   # Now with both terms
   grid <- data.frame(x0 = seq(0, 1, by = .1), x1 = seq(0, 1, by = .1))
   metafits <- metagam(fits, grid = grid, terms = c("s(x0)", "s(x1)"))
   expect_s3_class(metafits, "metagam")
   expect_equal(metafits$terms, c("s(x0)", "s(x1)"))
 
-  # Check that predictions are correct
-  # Vector to compare with has been generate by dput in a manual first run
-  # expect_equal(round(metafits$meta_estimates$estimate, 8),
-  #              c(-0.22238192, -1.90845939, -0.14829188, -1.54892801, -0.06563728,
-  #                -1.31104563, 0.01759225, -0.82227955, 0.08236379, -0.537348,
-  #                0.11504366, 0.07442868, 0.1144476, 0.59027953, 0.08619289, 0.93966274,
-  #                0.03228429, 1.51758374, -0.04820967, 2.21008333, -0.14742953,
-  #                2.88682888))
-  #
-  # # Check that standard errors are correct
-  # expect_equal(round(metafits$meta_estimates$se, 8),
-  #              c(0.42295604, 0.45839506, 0.33330303, 0.35355031, 0.27261384,
-  #                0.29159583, 0.23866624, 0.25726103, 0.22347574, 0.24035071, 0.21934645,
-  #                0.24135044, 0.2229785, 0.25485699, 0.23775134, 0.28412049, 0.27228278,
-  #                0.32960658, 0.3344465, 0.38952799, 0.42611871, 0.51139217))
+
 
   # Use link function
   metafits <- metagam(fits, grid_size = 10, type = "link")
   expect_s3_class(metafits, "metagam")
 
 
-  # Check that predictions are correct
-  # Vector to compare with has been generate by dput in a manual first run
-  # expect_equal(round(metafits$meta_estimates$estimate, 8),
-  #              c(5.4670439, 5.46037915, 5.49255883, 5.53565579, 5.56065563,
-  #                5.55408808, 5.51427511, 5.45682166, 5.4022208, 5.37047193, 5.77234916,
-  #                5.83204971, 5.90167633, 5.96530988, 6.00213034, 6.00226181, 5.96604297,
-  #                5.9079231, 5.84119815, 5.77202393, 6.10294759, 6.17831516, 6.25659141,
-  #                6.322799, 6.35921531, 6.35965851, 6.32644533, 6.27139173, 6.20166903,
-  #                6.1198474, 6.47675955, 6.55684504, 6.63719029, 6.70500065, 6.7440403,
-  #                6.74761609, 6.7159831, 6.65709282, 6.57599408, 6.48020266, 6.81718625,
-  #                6.92221445, 7.01954777, 7.09604684, 7.13919389, 7.14320217, 7.10459342,
-  #                7.02466736, 6.90695279, 6.77011341, 7.75841574, 7.79355039, 7.83062972,
-  #                7.86849486, 7.89089952, 7.88699169, 7.8562124, 7.81060755, 7.76039564,
-  #                7.70799628, 7.99200066, 8.08583347, 8.17387255, 8.24139827, 8.27613891,
-  #                8.27366052, 8.23374719, 8.16380394, 8.07067864, 7.96398062, 8.6347526,
-  #                8.72592665, 8.80718192, 8.86597393, 8.89224555, 8.88268038, 8.83721235,
-  #                8.76390866, 8.67098748, 8.56761759, 9.40907753, 9.48478576, 9.55602755,
-  #                9.6132786, 9.64492716, 9.64456517, 9.61156831, 9.55484575, 9.48169577,
-  #                9.39741113, 10.17843352, 10.21479917, 10.26564162, 10.31793464,
-  #                10.3525207, 10.35735471, 10.32883243, 10.27644202, 10.21197469,
-  #                10.1488514))
-
-  # Check that standard errors are correct
-  # expect_equal(round(metafits$meta_estimates$se, 8),
-  #              c(0.62125251, 0.54734162, 0.50675527, 0.48851356, 0.48130605,
-  #                0.47980643, 0.48460574, 0.5031818, 0.54790212, 0.62801304, 0.52279316,
-  #                0.44423438, 0.39858858, 0.3766332, 0.36771751, 0.36638181, 0.37307217,
-  #                0.39492924, 0.44316837, 0.5251978, 0.47845495, 0.39378464, 0.34320052,
-  #                0.31845653, 0.30843931, 0.30700639, 0.3147795, 0.33973125, 0.3930118,
-  #                0.4806462, 0.46770855, 0.37812971, 0.32271933, 0.29482284, 0.28315425,
-  #                0.28105084, 0.28955724, 0.31763749, 0.37598225, 0.46848224, 0.47121043,
-  #                0.37811471, 0.31893477, 0.28863955, 0.27610191, 0.27422384, 0.28415058,
-  #                0.31507888, 0.37685648, 0.47184022, 0.46800644, 0.37795971, 0.32114599,
-  #                0.2922408, 0.28074594, 0.27986466, 0.29070015, 0.3219654, 0.38333894,
-  #                0.47790479, 0.47665762, 0.38992139, 0.33623719, 0.30915261, 0.29850066,
-  #                0.29783363, 0.30787888, 0.33628698, 0.39257267, 0.48167837, 0.50397737,
-  #                0.42347752, 0.37444496, 0.35000036, 0.34069043, 0.34056594, 0.35029184,
-  #                0.37699855, 0.43017712, 0.51553861, 0.53747731, 0.46553916, 0.4243672,
-  #                0.40528701, 0.3987934, 0.39934706, 0.40736265, 0.42890294, 0.47319647,
-  #                0.5481428, 0.64774137, 0.58012839, 0.54134183, 0.52325335, 0.51702804,
-  #                0.51756521, 0.5251356, 0.54501404, 0.58560428, 0.65469007))
 
 })
 
 
+test_that("metagam accepts strange variable names", {
+  fits <- lapply(1:ndat, function(x){
+    dat <- gamSim(n = n, verbose = FALSE)
+    dat$var__with.many3things <- dat$x1
+    b <- gam(y ~ s(x0, bs = "cr") + s(var__with.many3things, bs = "cr"), data = dat)
+    strip_rawdata(b)
+  })
+
+  m <- metagam(fits, nsim = 3)
+
+})