From 9526f30c8396293d2c6f60bc10b141a0d6d44d8d Mon Sep 17 00:00:00 2001
From: Hugo <hugo@notainutilis.fr>
Date: Thu, 22 Feb 2024 23:40:01 +0100
Subject: [PATCH] support for regexes with ublacklist

---
 scripts/ublacklist.py              | 9 +++++++++
 scripts/update.sh                  | 4 ++++
 sources/regex_addresses/README.md  | 4 ++++
 sources/regex_titles/Clickbait.txt | 2 ++
 sources/regex_titles/README.md     | 1 +
 5 files changed, 20 insertions(+)
 create mode 100644 sources/regex_addresses/README.md
 create mode 100644 sources/regex_titles/Clickbait.txt
 create mode 100644 sources/regex_titles/README.md

diff --git a/scripts/ublacklist.py b/scripts/ublacklist.py
index d3820339..622f2935 100644
--- a/scripts/ublacklist.py
+++ b/scripts/ublacklist.py
@@ -15,9 +15,18 @@
 	linesurls = urls.readlines()
 blocklist = linesdomains + linestlds + linesurls
 
+with open("sources/regex_addresses.txt", "r") as addresses:
+	linesaddresses = titles.readlines()
+with open("sources/regex_titles.txt", "r") as titles:
+	linestitles = titles.readlines()
+
 # Print blocklist
 for line in linesheader:
 	print(line.strip())
 print()
 for line in blocklist:
 	print('*://*.' + line.strip() + '/*')
+for line in linesaddresses:
+	print(line.strip())
+for line in linestitles:
+	print('title/' + line.strip() + '/i')
diff --git a/scripts/update.sh b/scripts/update.sh
index d55ed9c1..268a3618 100755
--- a/scripts/update.sh
+++ b/scripts/update.sh
@@ -31,6 +31,10 @@ find ./sources/domains -type f -iname "*fediverse*.txt" -exec cat {} \; > ./sour
 find ./sources/urls -type f -iname "*.txt" -exec cat {} \; > ./sources/urls.txt
 ### TLDs
 find ./sources/tlds -type f -iname "*.txt" -exec cat {} \; > ./sources/tlds.txt
+### Regex addresses
+find ./sources/regex_addresses -type f -iname "*.txt" -exec cat {} \; > ./sources/regex_addresses.txt
+### Regex titles
+find ./sources/regex_titles -type f -iname "*.txt" -exec cat {} \; > ./sources/regex_titles.txt
 
 ## Cleanup the lists
 ### Remove comments, inline comments, spaces and empty lines
diff --git a/sources/regex_addresses/README.md b/sources/regex_addresses/README.md
new file mode 100644
index 00000000..2428f5ae
--- /dev/null
+++ b/sources/regex_addresses/README.md
@@ -0,0 +1,4 @@
+https://help.adblockplus.org/hc/en-us/articles/360062733293-How-to-write-filters#regexps
+https://github.com/gorhill/uBlock/wiki/Regular-expression-based-filters
+https://iorate.github.io/ublacklist/docs/advanced-features#regular-expressions
+https://developer.mozilla.org/en-US/docs/Web/JavaScript/Guide/Regular_Expressions
\ No newline at end of file
diff --git a/sources/regex_titles/Clickbait.txt b/sources/regex_titles/Clickbait.txt
new file mode 100644
index 00000000..c159c1f5
--- /dev/null
+++ b/sources/regex_titles/Clickbait.txt
@@ -0,0 +1,2 @@
+# Spoilers!
+/ending explained/i
\ No newline at end of file
diff --git a/sources/regex_titles/README.md b/sources/regex_titles/README.md
new file mode 100644
index 00000000..c94dc5e4
--- /dev/null
+++ b/sources/regex_titles/README.md
@@ -0,0 +1 @@
+https://iorate.github.io/ublacklist/docs/advanced-features#regular-expressions-for-page-titles
\ No newline at end of file