Adapt regex detection to Pandoc 3 figures to still get numbering

NEWS.md

@@ -1,5 +1,7 @@
 # CHANGES IN bookdown VERSION 0.44
 
+- Figure numbering is now correctly working in HTML with Pandoc 3, when Markdown syntax is used to include images (thanks, @N0rbert, #1467).
+
 
 # CHANGES IN bookdown VERSION 0.43
 

R/html.R

@@ -740,7 +740,7 @@ parse_fig_labels = function(content, global = FALSE) {
   m = gregexpr(sprintf('\\(#((%s):[-/[:alnum:]]+)\\)', reg_label_types), content)
   labs = regmatches(content, m)
   cntr = new_counters(label_types, chaps)  # chapter counters
-  figs = grep('^<div class="figure', content)
+  figs = grep('^<div class="(figure|float)', content)
   eqns = grep('<span class="math display">', content)
 
   for (i in seq_along(labs)) {
@@ -764,7 +764,7 @@ parse_fig_labels = function(content, global = FALSE) {
     arry = c(arry, setNames(num, lab))
 
     switch(type, fig = {
-      if (length(grep('^<p class="caption', content[i - 0:1])) == 0) {
+      if (length(grep('^<p class="caption|^<div class="figcaption', content[i - 0:1])) == 0) {
         # remove these labels, because there must be a caption on this or
         # previous line (possible negative case: the label appears in the alt
         # text of <img>)

tests/testthat/resources/figures.Rmd

@@ -0,0 +1,23 @@
+---
+title: Figure cross-reference issue in HTML
+output:
+  bookdown::html_document2: default
+---
+
+# Zero 1 {-#zero1}
+
+Some text in 1st unnumbered section.
+
+# Practically long header about some theories {#theory}
+
+## Markdown Syntax {#markdown-syntax}
+
+Content
+
+### Images {#markdown-syntax-media}
+
+First figure of first section is below, it reference is Fig. \@ref(fig:md-logo).
+
+![(\#fig:md-logo) Markdown logo](md.png)
+
+...

tests/testthat/test-html.R

@@ -79,3 +79,27 @@ test_that("add_toc_class() works for all pandoc versions", {
   TOC <- xml2::xml_find_all(content, "//div[@id = 'TOC']/ul/li")
   expect_equal(xml2::xml_attr(TOC, "class"), c("has-sub", NA, "has-sub"))
 })
+
+
+test_that("Figure are numbered correctly", {
+  skip_on_cran()
+  skip_if_not_pandoc()
+  skip_if_not_installed("xml2")
+
+  dir <- withr::local_tempdir("bookdown-test")
+  rmd_file <- "figures.Rmd"
+  file.copy(test_path("resources", rmd_file), dir)
+  file.copy(test_path("resources", "md.png"), dir)
+  withr::local_dir(dir)
+  out <- rmarkdown::render(rmd_file, output_format = "bookdown::html_document2")
+
+  content <- xml2::read_html(out)
+  xpath <- if (rmarkdown::pandoc_version() >= "3") {
+    "//div[@class='figcaption']"
+  } else {
+    "//p[@class='caption']"
+  }
+  figure <- xml2::xml_find_first(content, xpath)
+  expect_match(xml2::xml_text(figure), "Figure 1.1", fixed = TRUE)
+
+})