From c0e388f09bfde68764f3ca757c2607f3ac2b144d Mon Sep 17 00:00:00 2001
From: Benjamin Renard <brenard@easter-eggs.com>
Date: Sun, 8 Oct 2023 23:19:07 +0200
Subject: [PATCH] doc: clean all-in-one HTML file

---
 doc/Makefile                      |  4 +++-
 doc/clean-all-in-one-html-file.py | 40 +++++++++++++++++++++++++++++++
 doc/requirements.txt              |  1 +
 3 files changed, 44 insertions(+), 1 deletion(-)
 create mode 100644 doc/clean-all-in-one-html-file.py

diff --git a/doc/Makefile b/doc/Makefile
index 61567e8c..2b03d305 100644
--- a/doc/Makefile
+++ b/doc/Makefile
@@ -4,7 +4,9 @@ public_html: venv
 	venv/bin/mkdocs build -s
 
 LdapSaisie.html: venv public_html
-	venv/bin/htmlark public_html/print_page/index.html -o LdapSaisie.html
+	venv/bin/htmlark public_html/print_page/index.html -o LdapSaisie.tmp.html
+	venv/bin/python clean-all-in-one-html-file.py LdapSaisie.tmp.html LdapSaisie.html
+	rm -f LdapSaisie.tmp.html
 
 LdapSaisie.pdf: LdapSaisie.html
 	docker run -v $(CURDIR):/workspace pink33n/html-to-pdf --url file:///workspace/LdapSaisie.html --pdf LdapSaisie.pdf
diff --git a/doc/clean-all-in-one-html-file.py b/doc/clean-all-in-one-html-file.py
new file mode 100644
index 00000000..22dd3155
--- /dev/null
+++ b/doc/clean-all-in-one-html-file.py
@@ -0,0 +1,40 @@
+import sys
+import os.path
+
+from bs4 import BeautifulSoup
+
+# Check & handle arguments
+if len(sys.argv) < 2:
+    print(f'Usage: {sys.argv[0]} [/path/to/input.html] [/path/to/output.html]')
+    sys.exit(1)
+
+input_path = sys.argv[1]
+
+if not os.path.exists(input_path):
+    print(f'{input_path} not found')
+    sys.exit(1)
+
+output_path = sys.argv[2] if len(sys.argv) > 2 else input_path
+
+# Open & parse HTML input file
+with open(input_path, 'r') as fp:
+    soup = BeautifulSoup(fp, 'html.parser')
+
+# Delete some useless elements
+to_delete = [
+    (['div'], {'class': 'md-sidebar'}),
+    (['div'], {'class': 'md-search'}),
+    (['label'], {'for': '__search'}),
+    (['div'], {'id': 'print-site-banner'}),
+    (['div'], {'class': 'md-header__source'}),
+]
+for args, kwargs in to_delete:
+    for el in soup.find_all(*args, **kwargs):
+        el.decompose()
+
+# Change LdapSaisie logo header link to JS scroll top action
+soup.find('a', attrs={'class': 'md-logo'})['href'] = 'javascript:window.scrollTo(0,0)'
+
+# Store resulting HTML document in output file
+with open(output_path, 'w') as fp:
+    fp.write(str(soup))
diff --git a/doc/requirements.txt b/doc/requirements.txt
index 00364688..5da1695e 100644
--- a/doc/requirements.txt
+++ b/doc/requirements.txt
@@ -8,3 +8,4 @@ requests
 htmlark
 mike
 git+https://github.com/jimporter/mike.git@master
+beautifulsoup4