regexp question (my solution)

library(pipeR)
library(stringr)

set.seed(20)
str_candidates <- c(
  str_c("{", replicate(30L, str_c(sample(letters, sample(5:20, 5L), TRUE), collapse = "")), "}"),
  str_c("{{", replicate(30L, str_c(sample(letters, sample(5:20, 5L), TRUE), collapse = "")), "}}"),
  str_c("{{", replicate(30L, str_c(sample(letters, sample(5:20, 5L), TRUE), collapse = "")), ":",
        str_c(rep(" ", sample(4, 1L)), collapse = ""),
        replicate(30L, str_c(sample(c(letters, str_split("/*-+=~!@#$%^&*`?", "")[[1]]), sample(5:20, 5), TRUE),
                            collapse = "")), "}}"),
  str_c("<", replicate(30L, str_c(sample(letters, sample(5:20, 5L), TRUE), collapse = "")), ">"),
  letters, LETTERS
)

N <- 10000L
n <- 50L

a <- replicate(N, sample(str_candidates, n, TRUE) %>>% str_c(collapse=" "), simplify = TRUE)
a2 <- a %>>% str_replace_all("\\{\\{([^: ]+)([^{}]+)?\\}\\}", "\\{\\{\\1\\}\\}") %>>%
  str_replace_all("([^{}]+)\\{[^{}]+\\}", "\\1") %>>%
  str_replace_all("^\\{[^{}]+\\}\\s?([^{}]+)?", "\\1") %>>%
  str_replace_all("([^<]+)?<[^<>]*>\\s?", "\\1")

all.equal(str_split(a, " ") %>>% laply(function(x) sum(str_detect(x, "^[a-zA-Z]$"))),
          str_split(a2, " ") %>>% laply(function(x) sum(str_detect(x, "^[a-zA-Z]$")))) # TRUE
str_detect(a2, "<[^<>]*>") %>>% any # FALSE
str_detect(a2, "[^\\{]\\{[^{}]*\\}[^\\}]") %>>% any # FALSE
str_detect(a2, "^\\{[^{}]*\\}[^\\}]") %>>% any # FALSE
str_detect(a2, "[^\\{]\\{[^{}]*\\}$") %>>% any # FALSE