From 771ada0184ae9c0b13c363f94e06d3ca9e10b362 Mon Sep 17 00:00:00 2001
From: Max Patiiuk <max@patii.uk>
Date: Mon, 12 Feb 2024 06:44:50 -0800
Subject: [PATCH] Clipper: handle tables with captions (#9910)

---
 .../tests/html_to_md/table_with_caption.html  | 15 +++++++++++++++
 .../tests/html_to_md/table_with_caption.md    |  6 ++++++
 .../tests/html_to_md/table_with_colgroup.html | 19 +++++++++++++++++++
 .../tests/html_to_md/table_with_colgroup.md   |  4 ++++
 packages/turndown-plugin-gfm/src/tables.js    | 15 ++++++++++++++-
 5 files changed, 58 insertions(+), 1 deletion(-)
 create mode 100644 packages/app-cli/tests/html_to_md/table_with_caption.html
 create mode 100644 packages/app-cli/tests/html_to_md/table_with_caption.md
 create mode 100644 packages/app-cli/tests/html_to_md/table_with_colgroup.html
 create mode 100644 packages/app-cli/tests/html_to_md/table_with_colgroup.md

diff --git a/packages/app-cli/tests/html_to_md/table_with_caption.html b/packages/app-cli/tests/html_to_md/table_with_caption.html
new file mode 100644
index 000000000..0047f9549
--- /dev/null
+++ b/packages/app-cli/tests/html_to_md/table_with_caption.html
@@ -0,0 +1,15 @@
+<table>
+	<caption>Caption</caption>
+	<tbody>
+		<tr>
+			<th>Row 1, Column 1</th>
+			<th>Row 1, Column 2</th>
+			<th>Row 1, Column 3</th>
+		</tr>
+		<tr>
+			<th>Row 2, Column 1</th>
+			<td>Row 2, Column 2</td>
+			<td>Row 2, Column 3</td>
+		</tr>
+	</tbody>
+</table>
\ No newline at end of file
diff --git a/packages/app-cli/tests/html_to_md/table_with_caption.md b/packages/app-cli/tests/html_to_md/table_with_caption.md
new file mode 100644
index 000000000..99e085ccf
--- /dev/null
+++ b/packages/app-cli/tests/html_to_md/table_with_caption.md
@@ -0,0 +1,6 @@
+Caption
+
+|     |     |     |
+| --- | --- | --- |
+| Row 1, Column 1 | Row 1, Column 2 | Row 1, Column 3 |
+| Row 2, Column 1 | Row 2, Column 2 | Row 2, Column 3 |
\ No newline at end of file
diff --git a/packages/app-cli/tests/html_to_md/table_with_colgroup.html b/packages/app-cli/tests/html_to_md/table_with_colgroup.html
new file mode 100644
index 000000000..2f0ff5d91
--- /dev/null
+++ b/packages/app-cli/tests/html_to_md/table_with_colgroup.html
@@ -0,0 +1,19 @@
+<table>
+	<colgroup>
+    <col />
+    <col span="2" class="test" />
+    <col span="2" class="test2" />
+  </colgroup>
+	<tbody>
+		<tr>
+			<th>Row 1, Column 1</th>
+			<th>Row 1, Column 2</th>
+			<th>Row 1, Column 3</th>
+		</tr>
+		<tr>
+			<th>Row 2, Column 1</th>
+			<td>Row 2, Column 2</td>
+			<td>Row 2, Column 3</td>
+		</tr>
+	</tbody>
+</table>
\ No newline at end of file
diff --git a/packages/app-cli/tests/html_to_md/table_with_colgroup.md b/packages/app-cli/tests/html_to_md/table_with_colgroup.md
new file mode 100644
index 000000000..befad14cf
--- /dev/null
+++ b/packages/app-cli/tests/html_to_md/table_with_colgroup.md
@@ -0,0 +1,4 @@
+|     |     |     |
+| --- | --- | --- | 
+| Row 1, Column 1 | Row 1, Column 2 | Row 1, Column 3 |
+| Row 2, Column 1 | Row 2, Column 2 | Row 2, Column 3 |
\ No newline at end of file
diff --git a/packages/turndown-plugin-gfm/src/tables.js b/packages/turndown-plugin-gfm/src/tables.js
index 0df0116d1..1492211f5 100644
--- a/packages/turndown-plugin-gfm/src/tables.js
+++ b/packages/turndown-plugin-gfm/src/tables.js
@@ -98,10 +98,23 @@ rules.table = {
       }
     }
 
-    return '\n\n' + emptyHeader + content + '\n\n'
+    const captionContent = node.caption ? node.caption.textContent || '' : '';
+    const caption = captionContent ? `${captionContent}\n\n` : '';
+    const tableContent = `${emptyHeader}${content}`.trimStart();
+    return `\n\n${caption}${tableContent}\n\n`;
   }
 }
 
+rules.tableCaption = {
+  filter: ['caption'],
+  replacement: () => '',
+};
+
+rules.tableColgroup = {
+  filter: ['colgroup', 'col'],
+  replacement: () => '',
+};
+
 rules.tableSection = {
   filter: ['thead', 'tbody', 'tfoot'],
   replacement: function (content) {