From 675b17700a26cbf05c4aeffb12ba17cd4f228d5b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ma=C3=ABlle=20Salmon?= <maelle.salmon@yahoo.se>
Date: Fri, 5 Apr 2024 10:53:20 +0200
Subject: [PATCH] feat: footnote spacing&colon protection

---
 R/translate.R                                 | 32 +++++++++++++++----
 inst/example-equations-footnote.md            |  3 ++
 tests/testthat/_snaps/translate.md            | 11 +++++++
 .../api/translate-746e3b-POST.json            |  8 +++++
 tests/testthat/test-translate.R               | 18 ++++++++++-
 5 files changed, 64 insertions(+), 8 deletions(-)
 create mode 100644 inst/example-equations-footnote.md
 create mode 100644 tests/testthat/fixtures/example-equations-footnote/api/translate-746e3b-POST.json

diff --git a/R/translate.R b/R/translate.R
index 6f8e449..20ec117 100644
--- a/R/translate.R
+++ b/R/translate.R
@@ -386,11 +386,13 @@ unprotect_math <- function(math) {
 }
 protect_squaries <- function(node) {
   text <- xml2::xml_text(node)
-  text <- gsub("\\[", "</text><squary>", text)
-  text <- gsub("\\]", "</squary><text>", text)
-  text <- sprintf("<text>%s</text>", text)
-  text <- gsub("<text><\\/text>", "", text)
-  text <- sprintf("<text>%s</text>", text)
+  text <- gsub('\\[', '</text><squary>', text)
+  text <- gsub('\\]', '</squary><text>', text)
+  text <- sprintf('<text xml:space="preserve">%s</text>', text)
+  text <- gsub('<text xml:space="preserve"><\\/text>', '', text)
+  text <- sprintf('<text xml:space="preserve">%s</text>', text)
+  # hack to preserve colon that DeepL API kills
+  text <- gsub("</squary><text>:", ":</squary><text>", text)
 
   at_things <- regmatches(text, gregexpr("@[[:alnum:]]*", text))[[1]]
   footnote_things <- regmatches(text, gregexpr("\\^[[:alnum:]]*", text))[[1]]
@@ -407,7 +409,13 @@ protect_squaries <- function(node) {
 
 unprotect_squary <- function(node) {
   xml2::xml_name(node) <- "text"
-  xml2::xml_text(node) <- sprintf("[%s]", trimws(xml2::xml_text(node)))
+  node_text <- trimws(xml2::xml_text(node))
+  if (!grepl(":$", node_text)) {
+    xml2::xml_text(node) <- sprintf("[%s]", node_text)
+  } else {
+    node_text <- sub(":$", "", node_text)
+    xml2::xml_text(node) <- sprintf("[%s]:", node_text)
+  }
 }
 
 unprotect_notranslate <- function(node) {
@@ -423,11 +431,21 @@ unprotect_non_code_block <- function(non_code_block) {
 }
 
 untangle_text <- function(node) {
-  text <- trimws(xml2::xml_text(node))
+  text <- xml2::xml_text(node)
+  text <- gsub("\\s+", " ", text) # like str_squish w/o str_trim
+  # trying to only leave space where needed
+  no_left_sibling <- (length(xml2::xml_find_first(node, "preceding-sibling::*")) == 0)
+  which <- if (no_left_sibling) {
+    "both"
+  } else {
+    "right"
+  }
+  text <- trimws(text, which = which)
   xml2::xml_remove(xml2::xml_children(node))
   xml2::xml_replace(
     node,
     xml2::xml_name(node),
+    `xml:space`="preserve",
     asis = 'true',
     gsub("\\\n", "", text)
   )
diff --git a/inst/example-equations-footnote.md b/inst/example-equations-footnote.md
new file mode 100644
index 0000000..1fe953a
--- /dev/null
+++ b/inst/example-equations-footnote.md
@@ -0,0 +1,3 @@
+Voyez $a$ une équation avec une note ensuite [^1].
+
+[^1]: une note de bas de page.
diff --git a/tests/testthat/_snaps/translate.md b/tests/testthat/_snaps/translate.md
index efd456f..8379f04 100644
--- a/tests/testthat/_snaps/translate.md
+++ b/tests/testthat/_snaps/translate.md
@@ -46,3 +46,14 @@
     Output
       [1] "$i_t = j_t$"
 
+# deepl_translate() handles equations+footnote well
+
+    Code
+      foot_math_lines
+    Output
+      [1] "See $a$ an equation with a note afterwards [^1] ."
+      [2] ""                                                 
+      [3] "[^1]: a footnote."                                
+      [4] ""                                                 
+      [5] ""                                                 
+
diff --git a/tests/testthat/fixtures/example-equations-footnote/api/translate-746e3b-POST.json b/tests/testthat/fixtures/example-equations-footnote/api/translate-746e3b-POST.json
new file mode 100644
index 0000000..4f8bf0a
--- /dev/null
+++ b/tests/testthat/fixtures/example-equations-footnote/api/translate-746e3b-POST.json
@@ -0,0 +1,8 @@
+{
+    "translations": [
+        {
+            "detected_source_language": "FR",
+            "text": "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<!DOCTYPE document SYSTEM \"CommonMark.dtd\">\n<document xmlns=\"http://commonmark.org/xml/1.0\">\n  <paragraph>\n   <text xml:space=\"preserve\">See <\/text>\n   <math asis=\"true\">$a$<\/math>\n   <text xml:space=\"preserve\">\n      <text xml:space=\"preserve\"> an equation with a note afterwards <\/text>\n     <squary>\n        <notranslate>^1<\/notranslate>\n      <\/squary>\n     <text>.<\/text>\n   <\/text>\n  <\/paragraph>\n  <paragraph>\n   <text xml:space=\"preserve\">\n     <squary><notranslate>^1<\/notranslate>:<\/squary>\n      <text> a footnote.<\/text>\n   <\/text>\n  <\/paragraph>\n<\/document>\n"
+        }
+    ]
+}
diff --git a/tests/testthat/test-translate.R b/tests/testthat/test-translate.R
index 9e6fd11..23631e4 100644
--- a/tests/testthat/test-translate.R
+++ b/tests/testthat/test-translate.R
@@ -205,7 +205,6 @@ test_that("deepl_translate() handles square brackets stuff well", {
   })
 })
 
-
 test_that("deepl_translate() handles equations well", {
   to_translate <- system.file("example-equations.md", package = "babeldown")
   out_path <- withr::local_tempfile()
@@ -224,3 +223,20 @@ test_that("deepl_translate() handles equations well", {
   expect_snapshot(sub(".*que ", "", math_lines[7]))
   expect_snapshot(math_lines[9])
 })
+
+test_that("deepl_translate() handles equations+footnote well", {
+  to_translate <- system.file("example-equations-footnote.md", package = "babeldown")
+  out_path <- withr::local_tempfile()
+  with_mock_dir("example-equations-footnote", {
+    deepl_translate(
+      path = to_translate,
+      out_path = out_path,
+      source_lang = "FR",
+      target_lang = "EN-US",
+      yaml_fields = NULL
+    )
+  })
+  foot_math_lines <- brio::read_lines(out_path)
+  expect_snapshot(foot_math_lines)
+})
+