Allow Non UTF-8 source

- Add unit test to check_translation_csv()
- closes #61

NAMESPACE

@@ -18,6 +18,7 @@ import(shiny)
 import(tmap)
 importFrom(DBI,dbConnect)
 importFrom(RPostgres,Postgres)
+importFrom(dplyr,mutate_all)
 importFrom(golem,activate_js)
 importFrom(golem,add_resource_path)
 importFrom(golem,bundle_resources)

R/utils_csv_check.R

@@ -1,32 +1,50 @@
 #' csv_check
-#'
+#' 
 #' @description Check the CSV written by hand
 #'
+#' @param path path to file to check
+#' @param source_encoding Encoding of the source file 
+#' @param new_path Where to save the modified file
 #' @return The return value, if any, from executing the utility.
-#'
+#' @importFrom dplyr mutate_all
 #' @export
-check_translation_csv <- function(path) {
-    if (!file.exists(path)) {
-        stop("File doesn't  exists")
-    }
-    if (tools::file_ext(path) != "csv") {
-        stop("File should be a csv file")
-    }
-    fls <- read.csv(path)
-    if (
-        names(fls)[1] != "entry"
-    ) {
-        stop("The first column should be named 'entry'")
-    }
-    if (
-        !"en" %in% names(fls)
-    ) {
-        stop("Their is no 'en' column")
-    }
-    if (
-        !"fr" %in% names(fls)
-    ) {
-        stop("Their is no 'fr' column")
-    }
-    cli::cli_alert_success("CSV seems correct")
+check_translation_csv <- function(path, source_encoding = "latin1", new_path = path) {
+  if (!file.exists(path)) {
+    stop("File doesn't  exists")
+  }
+  if (tools::file_ext(path) != "csv") {
+    stop("path should be a csv file")
+  }
+  if (tools::file_ext(new_path) != "csv") {
+    stop("new_path should be a csv file")
+  }
+  fls <- read.csv(path, encoding = source_encoding)
+  if (
+    names(fls)[1] != "entry"
+  ) {
+    stop("The first column should be named 'entry'")
+  }
+  if (
+    !"en" %in% names(fls)
+  ) {
+    stop("Their is no 'en' column")
+  }
+  if (
+    !"fr" %in% names(fls)
+  ) {
+    stop("Their is no 'fr' column")
+  }
+  if (source_encoding != "UTF-8") {
+    fls <- fls %>% mutate_all(enc2utf8)
+    write.csv(fls, file = new_path, 
+              fileEncoding = "UTF-8",
+              # col.names = TRUE, 
+              row.names = FALSE)
+    cli::cli_alert_success("CSV was re-encoded to UTF-8")
+  } else {
+    file.copy(path, new_path)
+  }
+  cli::cli_alert_success("CSV seems correct")
+
+  return(fls)
 }

dev/translation.Rmd

@@ -103,17 +103,24 @@ DBI::dbGetQuery(
 + `inst/translation.csv` est le fichier à compléter à la main. 
 Il peut être ouvert avec Excel et __doit__ être sauvegardé en `.csv`.
 
-Vous pouvez vérifier l'intégriter du csv via : 
+Vous devez vérifier l'intégrité du csv via et indiquer avec quel encodage vous avez créé le fichier : 
+
+- `source_encoding = "UTF-8"` a priori, si créé sous Linux/MacOS
+- `source_encoding = "latin1"` a priori, si créé sur Excel/LibreOffice avec Windows
+
+Ainsi, si vous rencontrez des problèmes d'encodage sur l'application, vous pouvez modifier de nouveau le csv et changer le paramètre ci-dessous.
+Vous pouvez aussi le vérifier en regardant le fichier lu avec `output` ci-dessous.
 
 ```{r}
-check_translation_csv("inst/translation.csv")
+output <- check_translation_csv("inst/translation.csv", source_encoding = "latin1")
+head(output)
 ```
 
 ## Notes pour les dévelopeurs 
 
 Comment créer un nouvel élément "traductible" dans l'UI?
 
-+ La traduction est assurée par le module JS `i18n` et la fonction `with_i18` dans l'app
++ La traduction est assurée par le module JS `i18n` et la fonction `with_i18()` dans l'app
 
 ```{r}
 with_i18(

inst/reprex/trad_latin1.csv

@@ -0,0 +1,2 @@
+entry,en,fr
+toto,t�t�,t�t�

inst/reprex/trad_utf8.csv

@@ -0,0 +1,2 @@
+entry,en,fr
+toto,tùtù,tàtà

tests/testthat/test-check_translation_csv.R

@@ -0,0 +1,42 @@
+trad_latin <- system.file("reprex", "trad_latin1.csv", package = "diades.atlas")
+
+new_output_test <- tempfile(fileext = ".csv")
+
+test_that("check_translation_csv works with encoding Latin1", {
+  output <- check_translation_csv(
+    path = trad_latin,
+    source_encoding = "latin1", new_path = new_output_test)
+  expect_equal(names(output), c("entry", "en", "fr"))
+  expect_equal(as.character(output[1,]), c("toto", "t\u00f9t\u00f9", "t\u00e0t\u00e0"))
+  expect_equal(
+    readLines(new_output_test)[1],
+    c("\"entry\",\"en\",\"fr\"")
+  )
+
+  # Read re-encoded one: new_output_test
+  output <- check_translation_csv(
+    path = new_output_test,
+    source_encoding = "UTF-8", new_path = new_output_test)
+  expect_equal(names(output), c("entry", "en", "fr"))
+  expect_equal(as.character(output[1,]), c("toto", "t\u00f9t\u00f9", "t\u00e0t\u00e0"))
+  expect_equal(
+    readLines(new_output_test)[1],
+    c("\"entry\",\"en\",\"fr\"") # Copied with write.csv
+  )
+})
+
+# with uf8 ----
+trad_utf8 <- system.file("reprex", "trad_utf8.csv", package = "diades.atlas")
+new_output_test <- tempfile(fileext = ".csv")
+
+test_that("check_translation_csv works with encoding UTF8 directly", {
+  output <- check_translation_csv(
+    path = trad_utf8,
+    source_encoding = "UTF-8", new_path = new_output_test)
+  expect_equal(names(output), c("entry", "en", "fr"))
+  expect_equal(as.character(output[1,]), c("toto", "t\u00f9t\u00f9", "t\u00e0t\u00e0"))
+  expect_equal(
+    readLines(new_output_test)[1],
+    c("entry,en,fr") # Copied as is
+  )
+})