From 560e6d24ba4de19d81cee2660c0d4cfb724c2308 Mon Sep 17 00:00:00 2001
From: Markus Neteler <neteler@osgeo.org>
Date: Wed, 5 Feb 2025 18:23:27 +0100
Subject: [PATCH] docs: script to convert HTML manual pages to markdown (#4620)

---
 utils/grass_html2md.sh     | 69 ++++++++++++++++++++++++++++++++++++++
 utils/pandoc_codeblock.lua |  8 +++++
 2 files changed, 77 insertions(+)
 create mode 100755 utils/grass_html2md.sh
 create mode 100644 utils/pandoc_codeblock.lua

diff --git a/utils/grass_html2md.sh b/utils/grass_html2md.sh
new file mode 100755
index 00000000000..9ac2e548ae1
--- /dev/null
+++ b/utils/grass_html2md.sh
@@ -0,0 +1,69 @@
+#!/bin/bash
+set -eu
+
+###############################################################################
+# Convert recursively all .html files to .md (GitHub flavoured Markdown)
+#
+# Dependencies:
+#    pandoc
+#    wget
+#
+# Author(s):
+#    Martin Landa, Markus Neteler
+#
+# Usage:
+#    If you have "pandoc" in PATH, execute for HTML file conversion in
+#    current directory and subdirectories:
+#      ./utils/grass_html2md.sh
+#
+# COPYRIGHT: (C) 2024 by the GRASS Development Team
+#
+#            This program is free software under the GNU General Public
+#            License (>=v2). Read the file COPYING that comes with GRASS
+#            for details.
+#
+###############################################################################
+
+# cleanup at user break
+cleanup()
+{
+   rm -f "${f%%.html}_tmp.html"
+}
+
+# what to do in case of user break:
+exitprocedure()
+{
+   echo "User break!"
+   cleanup
+   exit 1
+}
+# shell check for user break (signal list: trap -l)
+trap "exitprocedure" 2 3 15
+
+# path to LUA file (./utils/pandoc_codeblock.lua)
+UTILSPATH="utils"
+
+# run recursively: HTML to MD
+for f in $(find . -name *.html); do
+    echo "${f}"
+
+    # HTML: Process the tmp file to selectively replace .html with .md only in relative URLs
+    sed -E '
+  # Step 1: Preserve http/https links with .html (and optional anchors)
+  s|(<a href="https?://[^"]+\.html)(#[^"]*)?">|\1_KEEPHTML\2">|g;
+  # Step 2: Replace .html with .md for local links (with or without anchors)
+  s|(<a href=")([^"]+)\.html(#[^"]*)?">|\1\2.md\3">|g;
+  # Step 3: Restore preserved http/https links with .html
+  s|_KEEPHTML||g;
+' "${f%%.html}.html" > "${f%%.html}_tmp.html"
+
+    cat "${f%%.html}_tmp.html" | \
+        sed 's#<div class="code"><pre>#<pre><code>#g' | \
+        sed 's#</pre></div>#</code></pre>#g' | \
+        pandoc --from=html --to=markdown -t gfm \
+               --lua-filter "${UTILSPATH}/pandoc_codeblock.lua" | \
+        sed 's+ \\\$+ \$+g' | sed 's+%20+-+g' > "${f%%.html}.md"
+
+    rm -f "${f%%.html}_tmp.html"
+
+done
diff --git a/utils/pandoc_codeblock.lua b/utils/pandoc_codeblock.lua
new file mode 100644
index 00000000000..e2a0a54910f
--- /dev/null
+++ b/utils/pandoc_codeblock.lua
@@ -0,0 +1,8 @@
+-- Pandoc Lua filter to handle code blocks
+-- Test cases
+-- raster/r.sun/r.sun.html
+
+-- Function to convert code blocks to markdown
+function CodeBlock (cb)
+    return pandoc.RawBlock('markdown', '```shell\n' .. cb.text .. '\n```\n')
+end