find_variables() & brms: Distributional component variables missing

DESCRIPTION

@@ -1,7 +1,7 @@
 Type: Package
 Package: insight
 Title: Easy Access to Model Information for Various Model Objects
-Version: 0.20.2.11
+Version: 0.20.2.12
 Authors@R:
     c(person(given = "Daniel",
              family = "Lüdecke",

NEWS.md

@@ -8,6 +8,9 @@
 
 ## Bug fixes
 
+* Fixed issue in `find_formula()`, `find_predictors()` and `find_variables()`
+  for models from package *brms* with custom formulas.
+
 * Fixed issues in `find_response()` for *brms* models with `mi()` function in
   the response variable.
 

R/find_formula.R

@@ -526,7 +526,7 @@
     id <- attr(x, "id")
 
     within_variables <- names(attr(x, "within"))
     within_variables <- paste0(within_variables, collapse = "*")
     within_variables <- paste0("(", within_variables, ")")
     e <- paste0("Error(", id, "/", within_variables, ")")
 
@@ -1578,6 +1578,33 @@
   f_bias <- f$pforms$bias
   f_bs <- f$pforms$bs
 
+  # brms formulas can also have custom names, based on variable names, e.g.:
+  # brm(
+  #   bf(carb ~ gear * vs) + lf(disc ~ 0 + mo(cyl)),
+  #   data = mtcars,
+  #   family = cumulative("probit"),
+  # )
+  # the lf() part is in "f$pforms" with name "disc".
+  #
+  # we therefore need to check whether we have additional names not yet covered
+  # by the above exceptions.
+
+  # auxiliary names
+  auxiliary_names <- c(
+    "sigma", "mu", "nu", "shape", "beta", "phi", "hu", "ndt", "zoi", "coi",
+    "kappa", "bias", "bs", "zi"
+  )
+
+  # check if any further pforms exist
+  if (!all(names(f$pforms) %in% auxiliary_names)) {
+    custom_names <- setdiff(names(f$pforms), auxiliary_names)
+    if (length(custom_names)) {
+      f_custom <- f$pforms[custom_names]
+    }
+  } else {
+    f_custom <- NULL
+  }
+
   f_sigmarandom <- NULL
   f_betarandom <- NULL
 
@@ -1633,7 +1660,7 @@
   }
 
 
-  compact_list(list(
+  compact_list(c(list(
     conditional = f_cond,
     random = f_random,
     zero_inflated = f_zi,
@@ -1653,7 +1680,7 @@
     zero_one_inflated = f_zoi,
     conditional_one_inflated = f_coi,
     kappa = f_kappa
-  ))
+  ), f_custom))
 }
 
 

R/find_predictors.R

@@ -82,7 +82,7 @@ find_predictors.default <- function(x,
 
   f <- find_formula(x, verbose = verbose)
   is_mv <- is_multivariate(f)
-  elements <- .get_elements(effects, component)
+  elements <- .get_elements(effects, component, model = x)
 
 
   # filter formulas, depending on requested effects and components

R/helper_functions.R

@@ -244,7 +244,7 @@
   )
 }
 
-.get_elements <- function(effects, component) {
+.get_elements <- function(effects, component, model = NULL) {
   # all elements of a model
   elements <- .all_elements()
 
@@ -280,6 +280,14 @@
     return(auxiliary_parameters)
   }
 
+  # if we have brms-models with custom formulas, we have element-names
+  # that are not covered by the standard elements. We then just do not
+  # filter elements.
+  if (inherits(model, "brmsfit") && component == "all") {
+    f <- insight::find_formula(model)
+    elements <- unique(c(elements, names(f)))
+  }
+
   elements <- switch(effects,
     all = elements,
     fixed = elements[!elements %in% random_parameters],

tests/testthat/test-brms.R

@@ -889,3 +889,14 @@
   out <- get_modelmatrix(m9)
   expect_identical(dim(out), c(32L, 2L))
 })
+
+test_that("get_modelmatrix", {
+  m10 <- insight::download_model("brms_lf_1")
+  expect_identical(
+    find_variables(m10),
+    list(
+      response = "carb",
+      conditional = c("gear", "vs"),disc = c("disc", "cyl")
+    )
+  )
+})