bayesian_as_frequentist() fails for brms 0 + Intercept formula

Fixes #646

DESCRIPTION

@@ -1,7 +1,7 @@
 Type: Package
 Package: bayestestR
 Title: Understand and Describe Bayesian Models and Posterior Distributions
-Version: 0.13.2.4
+Version: 0.13.2.5
 Authors@R:
     c(person(given = "Dominique",
              family = "Makowski",

NEWS.md

@@ -13,6 +13,11 @@
 * `bayesian_as_frequentist()` now supports more model families from Bayesian
   models that can be successfully converted to their frequentists counterparts.
 
+## Bug fixes
+
+* Fixed issue in `bayesian_as_frequentist()` for *brms* models with `0 + Intercept`
+  specification in the model formula.
+
 # bayestestR 0.13.2
 
 ## Breaking Changes

R/convert_bayesian_to_frequentist.R

@@ -47,6 +47,14 @@ convert_bayesian_as_frequentist <- function(model, data = NULL, REML = TRUE) {
   model_formula <- insight::find_formula(model)
   model_family <- insight::get_family(model)
 
+  # fix exception: The 0 + Intercept syntax in brms can be used to facilitate
+  # prior specification for the intercept, but but it leads to issues where it
+  # wrongly can be believed that Intercept is a variable and not a special term.
+  f_string <- insight::safe_deparse(model_formula$conditional)
+  if (grepl("0 + Intercept", f_string, fixed = TRUE)) {
+    model_formula$conditional <- stats::as.formula(gsub("0 + Intercept", "1", f_string, fixed = TRUE))
+  }
+
   if (inherits(model_family, "brmsfamily")) {
     insight::check_if_installed("glmmTMB")
     # exception: ordbetareg()