From 5c01119de298763d7871035452a51fd9a5f66cba Mon Sep 17 00:00:00 2001
From: Andrew Bruce <andrewbruce.himni@gmail.com>
Date: Mon, 29 Jul 2024 01:06:02 -0700
Subject: [PATCH] construct_regex testing

---
 R/regex.R              | 79 ++++++++++++++++++++++++------------------
 man/construct_regex.Rd | 11 ++++++
 2 files changed, 57 insertions(+), 33 deletions(-)

diff --git a/R/regex.R b/R/regex.R
index 08dcb14..cb115fe 100644
--- a/R/regex.R
+++ b/R/regex.R
@@ -5,6 +5,17 @@
 #' @examples
 #' construct_regex(search_descriptions()$hcpcs_code)
 #'
+#' construct_regex(search_hcpcs()$hcpcs_code)
+#'
+#' # Incorrect:
+#' construct_regex(search_adjustments()$adj_code)
+#'
+#' # Should be:
+#' "^[A-DM-PWY1-9][AIOR0-9]?[0-9]{0,3}?$"
+#'
+#' # Test adj codes
+#' c("4", "CO", "P6", "100", "B19", "MA22", "MA124", "N766")
+#'
 #' @returns `<chr>` vector
 #'
 #' @autoglobal
@@ -12,11 +23,14 @@
 #' @export
 construct_regex <- function(x) {
 
-  x <- collapse::funique(
-    collapse::na_rm(
-      gsub(" ", "", x)
-    )
-  )
+  uniq_nona <- \(x) collapse::funique(collapse::na_rm(x))
+
+  x <- gsub(" ", "", uniq_nona(x))
+
+  # stringi::stri_split_boundaries(
+  # x,
+  # type = "character",
+  # simplify = TRUE)
 
   vecs <- stringr::str_split_fixed(
     x, "",
@@ -28,14 +42,14 @@ construct_regex <- function(x) {
     purrr::map(dplyr::na_if, y = "")
 
   to_brackets <- vecs |>
-    purrr::map(collapse::na_rm) |>
-    purrr::map(collapse::funique) |>
+    purrr::map(uniq_nona) |>
     purrr::map(pos_re)
 
   qmark <- names(which(purrr::map_lgl(vecs, anyNA)))
 
   if (!vctrs::vec_is_empty(qmark)) {
-    to_brackets[qmark] <- purrr::map(to_brackets[qmark], \(x) paste0(x, "?"))
+    to_brackets[qmark] <- purrr::map(
+      to_brackets[qmark], \(x) paste0(x, "?"))
   }
 
   to_vec <- to_brackets |>
@@ -44,8 +58,8 @@ construct_regex <- function(x) {
 
   if (collapse::any_duplicated(to_vec)) {
 
-    # TODO probably need to vectorize this,
-    # will surely have more than one unique duplicate
+    # TODO probably need to vectorize this, will surely
+    # have more than one unique duplicate out of order
 
     dupe_idx <- which(collapse::fduplicated(to_vec, all = TRUE))
 
@@ -62,6 +76,27 @@ construct_regex <- function(x) {
   return(x)
 }
 
+#' Internal function for `construct_regex()`
+#'
+#' @param x `<chr>` vector
+#'
+#' @returns `<chr>` vector
+#'
+#' @autoglobal
+#'
+#' @noRd
+pos_re <- function(x) {
+
+  sorted   <- stringr::str_sort(x, numeric = TRUE)
+  alphabet <- purrr::list_c(strex::str_extract_non_numerics(sorted))
+  numbers  <- purrr::list_c(strex::str_extract_numbers(sorted))
+
+  paste0(
+    fuimus::collapser(alphabet),
+    fuimus::collapser(numbers)
+  )
+}
+
 #' Internal function for `construct_regex()`
 #'
 #' @param x `<chr>` vector
@@ -77,7 +112,7 @@ id_runs <- function(x) {
 
   vec <- rlang::set_names(rep(0, length(vec)), vec)
 
-  test <- strsplit(x, "")[[1]]
+  test <- fuimus::splitter(x)
 
   vecna <- vec[test]
 
@@ -150,28 +185,6 @@ id_runs <- function(x) {
   paste0("[", res, "]")
 }
 
-#' Internal function for `construct_regex()`
-#'
-#' @param x `<chr>` vector
-#'
-#' @returns `<chr>` vector
-#'
-#' @autoglobal
-#'
-#' @noRd
-pos_re <- function(x) {
-
-  sorted   <- stringr::str_sort(x, numeric = TRUE)
-  alphabet <- purrr::list_c(strex::str_extract_non_numerics(sorted))
-  numbers  <- purrr::list_c(strex::str_extract_numbers(sorted))
-
-  paste0(
-    fuimus::collapser(alphabet),
-    fuimus::collapser(numbers)
-    )
-
-}
-
 #' Internal function for `construct_regex2()`
 #'
 #' @param x `<chr>` vector
diff --git a/man/construct_regex.Rd b/man/construct_regex.Rd
index bcedb5e..f997432 100644
--- a/man/construct_regex.Rd
+++ b/man/construct_regex.Rd
@@ -18,4 +18,15 @@ Construct regex patterns
 \examples{
 construct_regex(search_descriptions()$hcpcs_code)
 
+construct_regex(search_hcpcs()$hcpcs_code)
+
+# Incorrect:
+construct_regex(search_adjustments()$adj_code)
+
+# Should be:
+"^[A-DM-PWY1-9][AIOR0-9]?[0-9]{0,3}?$"
+
+# Test adj codes
+c("4", "CO", "P6", "100", "B19", "MA22", "MA124", "N766")
+
 }