feat: extract language for code blocks correctly

README.md

@@ -33,7 +33,7 @@ Written in Go.
 1. [x] Custom font - defaults to Lexend
 1. [x] Use draft date as a fallback date for draft posts
 1. [x] Maintain the draft status for draft and pending posts
-1. [ ] Migrate code blocks correctly - syntax highlighting is not working right now
+1. [x] Migrate code blocks correctly - migrate existing code class information if available
 1. [ ] Featured images - I tried this [WordPress plugin](https://wordpress.org/plugins/export-media-with-selected-content/) but featured images are simply not exported
 
 ## Why existing tools don't work

src/wp2hugo/internal/hugogenerator/hugopage/hugo_page.go

@@ -36,7 +36,13 @@ const _WordPressMoreTag = "<!--more-->"
 const _customMoreTag = "{{< more >}}"
 const _wordPressTocTag = "[toc]"
 
-var _markdownImageLinks = regexp.MustCompile(`!\[.*?]\((.+?)\)`)
+var (
+	_markdownImageLinks = regexp.MustCompile(`!\[.*?]\((.+?)\)`)
+	// E.g. <pre class="EnlighterJSRAW" data-enlighter-language="golang">
+	_preTagExtractor1 = regexp.MustCompile(`<pre class="EnlighterJSRAW" data-enlighter-language="([^"]+?)".*?>([\s\S]*?)</pre>`)
+	// E.g. <pre class="lang:bash" nums="false">
+	_preTagExtractor2 = regexp.MustCompile(`<pre class=".*?lang:([^" ]+).*?>([\s\S]*?)</pre>`)
+)
 
 // Extracts "src" from Hugo figure shortcode
 // {{< figure align=aligncenter width=905 src="/wp-content/uploads/2023/01/Stollemeyer-castle-1024x768.jpg" alt="" >}}
@@ -127,6 +133,7 @@ func (page *Page) getMarkdown(provider ImageURLProvider, htmlContent string) (*s
 		return nil, fmt.Errorf("empty HTML content")
 	}
 	converter := getMarkdownConverter()
+	htmlContent = improvePreTagsWithCode(htmlContent)
 	htmlContent = replaceCaptionWithFigure(htmlContent)
 	htmlContent = replaceAWBWithParallaxBlur(provider, htmlContent)
 
@@ -164,6 +171,26 @@ func (page *Page) getMarkdown(provider ImageURLProvider, htmlContent string) (*s
 	return &markdown, nil
 }
 
+func improvePreTagsWithCode(htmlContent string) string {
+	// Replace all occurrences of "data-enlighter-language" with "language"
+	// Ref: https://github.com/JohannesKaufmann/html-to-markdown/blob/master/commonmark.go#L334
+	if strings.Contains(htmlContent, "data-enlighter-language") {
+		htmlContent = strings.ReplaceAll(htmlContent, `data-enlighter-language="golang"`, `data-enlighter-language="go"`)
+		htmlContent = strings.ReplaceAll(htmlContent, `data-enlighter-language="shell"`, `data-enlighter-language="bash"`)
+		htmlContent = strings.ReplaceAll(htmlContent, `data-enlighter-language="sh"`, `data-enlighter-language="bash"`)
+		htmlContent = strings.ReplaceAll(htmlContent, `data-enlighter-language="lang:`, `data-enlighter-language="`)
+		htmlContent = strings.ReplaceAll(htmlContent, `data-enlighter-language="language-`, `data-enlighter-language="`)
+		htmlContent = strings.ReplaceAll(htmlContent, `data-enlighter-language="raw"`, "")
+		htmlContent = strings.ReplaceAll(htmlContent, `data-enlighter-language="generic"`, "")
+		htmlContent = _preTagExtractor1.ReplaceAllString(htmlContent, `<pre><code class="$1">$2</code></pre>`)
+		htmlContent = strings.ReplaceAll(htmlContent, `class="EnlighterJSRAW"`, "")
+	}
+	if strings.Contains(htmlContent, "pre class=") {
+		htmlContent = _preTagExtractor2.ReplaceAllString(htmlContent, `<pre><code class="$1">$2</code></pre>`)
+	}
+	return htmlContent
+}
+
 // Mark code blocks with auto-detected language
 // Note: https://github.com/alecthomas/chroma is fairly inaccurate in detecting languages
 func highlightCode(markdown string) string {

src/wp2hugo/internal/hugogenerator/hugopage/hugo_page_test.go

@@ -0,0 +1,20 @@
+package hugopage
+
+import (
+	"github.com/stretchr/testify/assert"
+	"testing"
+)
+
+func TestPreTagExtractor2(t *testing.T) {
+	const example1 = `<pre class="lang:js decode:true">document.querySelector("video").playbackRate = 2.0;   // For 2X speed-up</pre>`
+	const example2 = `<pre class="theme:solarized-dark lang:sh decode:true">echo "whatever"</pre>`
+	const example3 = `<pre class="lang:sh decode:true"># Sample invocation:\n</pre>`
+	assert.True(t, _preTagExtractor2.MatchString(example1))
+	assert.True(t, _preTagExtractor2.MatchString(example2))
+	assert.True(t, _preTagExtractor2.MatchString(example3))
+
+	result3 := _preTagExtractor2.FindAllStringSubmatch(example3, -1)
+	assert.Equal(t, 1, len(result3))
+	assert.Equal(t, 3, len(result3[0]))
+	assert.Equal(t, "sh", result3[0][1])
+}