test type.convert=fun

DESCRIPTION

@@ -3,7 +3,7 @@ Authors@R: c(
     person("Toby", "Hocking",
      email="[email protected]",
      role=c("aut", "cre")))
-Version: 2024.12.16
+Version: 2024.12.17
 License: GPL-3
 Title: Named Capture to Data Tables
 Description: User-friendly functions for extracting a data

NEWS

@@ -1,6 +1,6 @@
-Changes in version 2024.12.16
+Changes in version 2024.12.17
 
-- type.convert=TRUE means to use type.convert() as default conversion function.
+- type.convert=TRUE means to use utils::type.convert(x,as.is=TRUE) as default conversion function (as.is=TRUE means to return character instead of factor), FALSE means identity, and otherwise can be any function to use as default conversion.
 
 Changes in version 2024.9.20
 

R/capture_all_str.R

@@ -16,8 +16,12 @@ capture_all_str <- structure(function # Capture all matches in a single subject
 ### separator string for combining elements of subject into a single
 ### string, used as collapse argument of base::paste.
   type.convert=getOption("nc.type.convert", FALSE)
-### If TRUE, use utils::type.convert instead of base::identity for the
-### default conversion.
+### Default conversion function, which will be used on each capture
+### group, unless a specific conversion is specified for that
+### group. If TRUE, use utils::type.convert; if FALSE, use
+### base::identity; otherwise must be a function of at least one
+### argument (character), returning an atomic vector of the same
+### length.
 ){
   stop_for_engine(engine)
   L <- subject_var_args(..., type.convert=type.convert)

tests/testthat/test-CRAN-vec.R

@@ -133,6 +133,51 @@ test_engines("capture_first_vec(type.convert=TRUE) returns two int columns", {
   expect_identical(computed, expected)
 })
 
+test_engines("capture_first_vec(type.convert=constant fun) returns all num columns", {
+  computed <- capture_first_vec(
+    "chr1:2-3,000",
+    chrom="chr.*?",
+    ":",
+    chromStart=".*?",
+    "-",
+    chromEnd="[0-9,]*",
+    type.convert=function(...)5)
+  expected <- data.table(
+    chrom=5,
+    chromStart=5,
+    chromEnd=5)
+  expect_identical(computed, expected)
+})
+
+test_engines("capture_first_vec(type.convert=as.integer) returns all int columns", {
+  computed <- capture_first_vec(
+    "chr1:2-3,000",
+    chrom="chr.*?",
+    ":",
+    chromStart=".*?",
+    "-",
+    chromEnd="[0-9,]*",
+    type.convert=as.integer)
+  expected <- data.table(
+    chrom=NA_integer_,
+    chromStart=2L,
+    chromEnd=NA_integer_)
+  expect_identical(computed, expected)
+})
+
+test_engines("capture_first_vec(type.convert=invalid) errors", {
+  expect_error({
+    capture_first_vec(
+      "chr1:2-3,000",
+      chrom="chr.*?",
+      ":",
+      chromStart=".*?",
+      "-",
+      chromEnd="[0-9,]*",
+      type.convert=function()5)
+  }, "type conversion functions should take one argument (character vector of captured text) and return an atomic vector of the same size; function for group 1(chrom) raised an error: unused argument (match.vec)", fixed=TRUE)
+})
+
 test_engines("capture_all_str(type.convert=TRUE) returns one int column", {
   computed <- capture_all_str(
     "chr1:2-3,000 chr4:5-6,000",