band aid for characters with unknown encodings

closes #12

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: paint
 Title: paint data.frames summaries in colour 
-Version: 0.1.6
+Version: 0.1.7
 Authors@R: 
     person(given = "Miles",
            family = "McBain",
@@ -24,6 +24,7 @@ Imports:
     pryr,
     purrr,
     RColorBrewer,
+    stringi,
     utils,
     vctrs,
     viridisLite

NEWS.md

@@ -1,7 +1,6 @@
-# 0.3.2
+# 0.1.7
 
-* Minor usability fix for RStudio
-* Make compatible with more strict type checking in dev `{dplyr}`.
+* Handle characters with invalid encoding by attempting conversion with `stringi::stri_trans_general`. This will result invalid characters being replaces by a missing char question mark. Previously these crashed out `{paint}`, failing in `{crayon}`.
 
 # 0.1.6
 

R/paint.R

@@ -50,7 +50,7 @@ paint.data.frame <- function(
     utils::head(
       offset(df, start_row),
       getOption("paint_n_rows", length(palette))
-    )
+    ) 
   cols <- mapply(
     paint_col,
     rows_to_paint,
@@ -61,10 +61,11 @@ paint.data.frame <- function(
     col_names <- align_str(col_names)
     col_types <- align_str(col_types)
   }
-  col_lines <- paste0(col_names, " ", col_types, " ", cols)
+  sanitised_cols <- vapply(cols, sanitise_col, character(1))
+  col_lines <- paste0(col_names, " ", col_types, " ", sanitised_cols)
   cropped_lines <- crop_lines(col_lines, getOption("paint_max_width", 60))
   col_block <- paste0(
-    sanitise_text(cropped_lines),
+    cropped_lines,
     collapse = "\n"
   )
   name <- paint_name(name)

R/sanitise_col.R

@@ -0,0 +1,7 @@
+#' Attempt to transform chars in encodings we don't understand, rather than crash crayon.
+#' @author Miles McBain
+sanitise_col <- function(col) {
+      sanitise_text(
+        ifelse(!validEnc(col), stringi::stri_trans_general(col, "latin-ascii"), col)
+      )
+}

tests/testthat/_snaps/invalid-encoding.md

@@ -0,0 +1,10 @@
+# invalid encodings are handled
+
+    Code
+      paint(df)
+    Output
+      data.frame [1, 3]
+      bad       chr 23�C
+      good      chr dummy
+      the_worst chr 2015 President and Vice-Chancellor�s Alumni S~
+

tests/testthat/test-invalid-encoding.R

@@ -0,0 +1,9 @@
+test_that("invalid encodings are handled", {
+df <- data.frame(
+    bad = "23\xbfC",
+    good = "dummy",
+    the_worst = "2015 President and Vice-Chancellor\x92s Alumni Scholarship Appeal"
+  )
+
+  expect_snapshot(paint(df))
+})