Merge pull request #7 from TheGroundZero/main

mms-gianni · web-flow · commit 347f3cef0bbc · 2025-01-07T09:09:25.000+01:00
Add Belgian sites
diff --git a/.github/workflows/concat.yml b/.github/workflows/concat.yml
@@ -1,19 +1,21 @@
 name: Concatenate
 
 on:
+  workflow_dispatch:
   push:
     branches:
       - main
 
 jobs:
   build:
+    if: "!contains(github.event.*.labels.*.name, 'automated-pr')"
     runs-on: ubuntu-latest
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
       - id: listup-targets
         run: |
-          files=$(find rulesets -type f -print)
+          files=$(find rulesets/ -name "*.yaml" -type f -print)
           escaped_files="${files//$'\n'/\ }"
           echo "files=${escaped_files}" >> $GITHUB_OUTPUT
       - name: Clear concat file
@@ -24,13 +26,24 @@ jobs:
         with:
           sources: ${{ steps.listup-targets.outputs.files }}
           destination: ruleset.yaml
-      - name: Commit files
-        run: |
-          git config --local user.email "2406013+github-actions[bot]@users.noreply.github.com"
-          git config --local user.name "github-actions[bot]"
-          git commit -a -m "Add changes"
-      - name: Push changes
-        uses: ad-m/github-push-action@master
+      - name: Create Pull Request
+        uses: peter-evans/create-pull-request@v6
+        id: cpr
         with:
-          github_token: ${{ secrets.GITHUB_TOKEN }}
-          branch: ${{ github.ref }}
+          commit-message: |
+            Concatenate ruleset
+            [skip ci]
+          labels: |
+            automated-pr
+          delete-branch: true
+      - name: Check outputs
+        if: ${{ steps.cpr.outputs.pull-request-number }}
+        run: |
+          echo "Pull Request Number - ${{ steps.cpr.outputs.pull-request-number }}"
+          echo "Pull Request URL - ${{ steps.cpr.outputs.pull-request-url }}"
+      - name: Enable Pull Request Automerge
+        if: ${{ steps.cpr.outputs.pull-request-number }}
+        run: |
+          gh pr merge --merge --auto "${{ steps.cpr.outputs.pull-request-number }}"
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
diff --git a/.gitignore b/.gitignore
@@ -0,0 +1 @@
+/ruleset.yaml
diff --git a/ruleset.yaml b/ruleset.yaml
@@ -1,53 +1,55 @@
-- domains: 
-  - www.thestar.com
-  - www.niagarafallsreview.ca
-  - www.stcatharinesstandard.ca
-  - www.thepeterboroughexaminer.com
-  - www.therecord.com
-  - www.thespec.com
-  - www.wellandtribune.ca
+- domain: theathletic.com
   injections:
     - position: head
       append: |
         <script>
-          window.localStorage.clear();
           document.addEventListener("DOMContentLoaded", () => {
-            const paywall = document.querySelectorAll('div.subscriber-offers');
-            paywall.forEach(el => { el.remove(); });
-            const subscriber_only = document.querySelectorAll('div.subscriber-only');
-            for (const elem of subscriber_only) {
-              if (elem.classList.contains('encrypted-content') && dompurify_loaded) {
-                const parser = new DOMParser();
-                const doc = parser.parseFromString('<div>' + DOMPurify.sanitize(unscramble(elem.innerText)) + '</div>', 'text/html');
-                const content_new = doc.querySelector('div');
-                elem.parentNode.replaceChild(content_new, elem);
-              }
-              elem.removeAttribute('style');
-              elem.removeAttribute('class');
-            }
-            const banners = document.querySelectorAll('div.subscription-required, div.redacted-overlay, div.subscriber-hide, div.tnt-ads-container');
-            banners.forEach(el => { el.remove(); });
-            const ads = document.querySelectorAll('div.tnt-ads-container, div[class*="adLabelWrapper"]');
-            ads.forEach(el => { el.remove(); });
-            const recommendations = document.querySelectorAll('div[id^="tncms-region-article"]');
-            recommendations.forEach(el => { el.remove(); });
+            let overlays = document.querySelectorAll('div[id*="overlay"], div:empty:not([data-rjs])');
+            overlays.forEach(el => el.remove());
+
+            let slideup = document.querySelectorAll('div[id^=slideup-]');
+            slideup.forEach(el => el.remove());
+
+            let apron = document.getElementById('free-apron-cta');
+            apron.remove();
+            
+            let body = document.querySelector('body');
+            body.style.overflow = 'visible';
+            body.style.position = 'relative';
           });
         </script>
-
+  tests:
+    - url: https://theathletic.com/5097900/2023/11/29/corey-perry-blackhawks-nhl/
+      test: document.querySelector('div[id*="overlay"], div:empty:not([data-rjs])') === null 
+    - url: https://theathletic.com/5097851/2023/11/29/mandel-mailbag-cfp-texas-alabama/
+      test: document.getElementById('free-apron-cta') === null
 - domains: 
-  - www.nytimes.com
-  - www.time.com
-  headers:
-    ueser-agent: Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)
-    cookie: nyt-a=; nyt-gdpr=0; nyt-geo=DE; nyt-privacy=1
-    referer: https://www.google.com/ 
+  - www.foxbusiness.com
+  - www.foxnews.com
   injections:
     - position: head
       append: |
         <script>
-          window.localStorage.clear();
           document.addEventListener("DOMContentLoaded", () => {
-            const banners = document.querySelectorAll('div[data-testid="inline-message"], div[id^="ad-"], div[id^="leaderboard-"], div.expanded-dock, div.pz-ad-box, div[id="top-wrapper"], div[id="bottom-wrapper"]');
+            let featuredVideo = document.querySelector('.featured-video');
+            featuredVideo.remove();
+
+            let relatedArticles = document.querySelectorAll('p a strong');
+            relatedArticles.forEach(el => el.remove());
+          });
+        </script>
+  tests:
+    - url: https://www.foxbusiness.com/fox-news-tech/elon-musk-tells-advertisers-who-left-x-go-f-yourself
+      test: document.querySelector('video') === null && document.querySelector('strong') === null
+    - url: https://www.foxnews.com/media/mayim-bialik-calls-feminists-silence-hamas-rape-torture-october-7-terror-attacks
+      test: document.querySelector('video') === null && document.querySelector('strong') === null
+- domain: www.usatoday.com
+  injections:
+    - position: head
+      append: |
+        <script>
+          document.addEventListener("DOMContentLoaded", () => {
+            const banners = document.querySelectorAll('div.roadblock-container, .gnt_nb, [aria-label="advertisement"], div[id="main-frame-error"]');
             banners.forEach(el => { el.remove(); });
           });
         </script>
@@ -88,6 +90,37 @@
             }
           });
         </script>
+  tests:
+    - url: https://www.americanbanker.com/news/how-foster-youth-become-financially-forgotten
+      test: document.querySelector('.inline-gate') === null
+- domains: 
+  - www.nytimes.com
+  - www.time.com
+  headers:
+    ueser-agent: Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)
+    cookie: nyt-a=; nyt-gdpr=0; nyt-geo=DE; nyt-privacy=1
+    referer: https://www.google.com/ 
+    content-security-policy: "default-src * 'unsafe-inline' 'unsafe-eval' data: blob:;"
+  injections:
+    - position: head
+      append: |
+        <script>
+          window.localStorage.clear();
+          document.addEventListener("DOMContentLoaded", () => {
+            const banners = document.querySelectorAll('div[data-testid="inline-message"], div[id^="ad-"], div[id^="leaderboard-"], div.expanded-dock, div.pz-ad-box, div[id="top-wrapper"], div[id="bottom-wrapper"]');
+            banners.forEach(el => { el.remove(); });
+          });
+        </script>
+  tests:
+    - url: https://www.nytimes.com/2023/11/24/business/retail-workers-black-friday-shopping.html
+      test: document.querySelector('div[data-testid="inline-message"]') === null
+- domain: medium.com
+  headers:
+    referer: https://t.co/x?amp=1
+    x-forwarded-for: none
+    user-agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/119.0.0.0 Safari/537.36
+    content-security-policy: script-src 'self';
+    cookie: 
 
 - domain: www.washingtonpost.com
   injections:
@@ -103,25 +136,47 @@
             headimage.forEach(image => { image.style.filter = ''; });
           });
         </script>
-
-- domain: www.usatoday.com
+  tests:
+    - url: https://www.washingtonpost.com/world/2023/11/24/russia-crimea-tunnel-china/
+      test: document.querySelector('div[data-qa="subscribe-promo"]') === null
+- domains: 
+  - www.thestar.com
+  - www.niagarafallsreview.ca
+  - www.stcatharinesstandard.ca
+  - www.thepeterboroughexaminer.com
+  - www.therecord.com
+  - www.thespec.com
+  - www.wellandtribune.ca
   injections:
     - position: head
       append: |
         <script>
+          window.localStorage.clear();
           document.addEventListener("DOMContentLoaded", () => {
-            const banners = document.querySelectorAll('div.roadblock-container, .gnt_nb, [aria-label="advertisement"], div[id="main-frame-error"]');
+            const paywall = document.querySelectorAll('div.subscriber-offers');
+            paywall.forEach(el => { el.remove(); });
+            const subscriber_only = document.querySelectorAll('div.subscriber-only');
+            for (const elem of subscriber_only) {
+              if (elem.classList.contains('encrypted-content') && dompurify_loaded) {
+                const parser = new DOMParser();
+                const doc = parser.parseFromString('<div>' + DOMPurify.sanitize(unscramble(elem.innerText)) + '</div>', 'text/html');
+                const content_new = doc.querySelector('div');
+                elem.parentNode.replaceChild(content_new, elem);
+              }
+              elem.removeAttribute('style');
+              elem.removeAttribute('class');
+            }
+            const banners = document.querySelectorAll('div.subscription-required, div.redacted-overlay, div.subscriber-hide, div.tnt-ads-container');
             banners.forEach(el => { el.remove(); });
+            const ads = document.querySelectorAll('div.tnt-ads-container, div[class*="adLabelWrapper"]');
+            ads.forEach(el => { el.remove(); });
+            const recommendations = document.querySelectorAll('div[id^="tncms-region-article"]');
+            recommendations.forEach(el => { el.remove(); });
           });
         </script>
-
-- domain: medium.com
-  headers:
-    referer: https://t.co/x?amp=1
-    x-forwarded-for: none
-    user-agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/119.0.0.0 Safari/537.36
-    content-security-policy: script-src 'self';
-    cookie: 
+  tests:
+  - url: https://www.wellandtribune.ca/news/niagara-region/niagara-transit-commission-rejects-council-request-to-reduce-its-budget-increase/article_e9fb424c-8df5-58ae-a6c3-3648e2a9df66.html
+    test: document.querySelector('div.subscriber-offers') === null
 
 # loads amp version of page
 - domain: tagesspiegel.de
@@ -132,6 +187,9 @@
     query:
       - key: amp
         value: 1
+  tests:
+    url: https://www.tagesspiegel.de/politik/impfungen-in-deutschland-und-der-welt-der-aktuelle-stand-der-impfkampagne/26809888.html
+    test: document.querySelector('html').classList.contains('amp-mode')
 
 - domain: www.nzz.ch
   paths:
@@ -157,7 +215,34 @@
             removeDOMElement(paywall)
           });
         </script>
+  tests:
+    - url: https://www.nzz.ch/feuilleton/edward-w-said-der-ideologe-des-palaestinensischen-terrors-ld.1766658
+      test: document.querySelector('.dynamic-regwall') === null
 
+- domain: ft.com
+  headers:
+    referer: https://t.co/x?amp=1
+    content-security-policy: "default-src * 'unsafe-inline' 'unsafe-eval' data: blob:;"
+  injections:
+    - position: head
+      append: |
+        <script>
+          document.addEventListener("DOMContentLoaded", () => {
+            const styleTags = document.querySelectorAll('link[rel="stylesheet"]');
+            styleTags.forEach(el => { 
+              const href = el.getAttribute('href').substring(1);
+              const updatedHref = href.replace(/(https?:\/\/.+?)\/{2,}/, '$1/');
+              el.setAttribute('href', updatedHref);
+            });
+            setTimeout(() => {
+              const cookie = document.querySelectorAll('.o-cookie-message, .js-article-ribbon, .o-ads, .o-banner, .o-message, .article__content-sign-up');
+              cookie.forEach(el => { el.remove(); });
+            }, 1000);
+          })
+        </script>
+  tests:
+    - url: https://www.ft.com/content/5348ec64-010e-40f4-a27e-6d1252a0c537
+      test: document.querySelector('.o-cookie-message') === null
 - domains:
   - myprivacy.dpgmedia.be
   - myprivacy.dpgmedia.nl
@@ -172,6 +257,7 @@
     user-agent: Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Googlebot/2.1; Googlebot-News; +http://www.google.com/bot.html) Chrome/121.0.6140.0 Safari/537.36
     x-forwarded-for: none
     referer: https://news.google.com
+    content-security-policy: "default-src * 'unsafe-inline' 'unsafe-eval' data: blob:;"
   regexRules:
     - match: (\s&&\s)?window\.temptation(\s&&\s)?(\.init)?(\s&&\s)?(\(([^)]\s*)*\);)?
       replace: true
@@ -187,25 +273,56 @@
             document.querySelector('div[data-advert-placeholder-collapses]').remove();
           });
         </script>
+  tests:
+    - url: https://www.demorgen.be/nieuws/verkiezingen-2024-zullen-uitdraaien-op-een-strijd-tussen-vivaldi-en-bourgeois-blok~b0f7b8a9/
+      test: document.querySelector('div[data-temptation-position="ARTICLE_BOTTOM"]') === null
 
-- domain: www.ft.com
+- domains:
+  - apache.be
   headers:
-    referer: https://t.co/x?amp=1
+    x-forwarded-for: none
+    cookie: cookiesjsr=%7B%22base%22%3Afalse%2C%22functional%22%3Afalse%2C%22video%22%3Afalse%7D
   injections:
     - position: head
       append: |
         <script>
           document.addEventListener("DOMContentLoaded", () => {
-            const styleTags = document.querySelectorAll('link[rel="stylesheet"]');
-            styleTags.forEach(el => { 
-              const href = el.getAttribute('href').substring(1);
-              const updatedHref = href.replace(/(https?:\/\/.+?)\/{2,}/, '$1/');
-              el.setAttribute('href', updatedHref);
-            });
-            setTimeout(() => {
-              const cookie = document.querySelectorAll('.o-cookie-message, .js-article-ribbon, .o-ads, .o-banner, .o-message, .article__content-sign-up');
-              cookie.forEach(el => { el.remove(); });
-            }, 1000);
-          })
+            // remove paywall
+            document.querySelector("#spb-block-apachepopupblock").remove();
+            // restore scrolling
+            document.body.style.overflow = "auto";
+          });
+        </script>
+  tests:
+    - url: https://apache.be/2024/03/25/chinese-spionkop-van-vlaams-belang
+      test: document.querySelector('#spb-block-apachepopupblock') === null
+
+- domains:
+  - kw.be
+  headers:
+    x-forwarded-for: none
+    referer: https://news.google.com
+    user-agent: Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Googlebot/2.1; Googlebot-News; +http://www.google.com/bot.html) Chrome/121.0.6140.0 Safari/537.36
+  injections:
+    - position: head
+      append: |
+        <script>
+          document.addEventListener("DOMContentLoaded", () => {
+            // remove paywall
+            document.querySelector("#paywall-modal").remove();
+          });
+        </script>
+    - position: head
+      append: |
+        <script>
+          document.addEventListener("scroll", () => {
+            // reset overflow
+            let htmlObject = document.getElementsByTagName("html")[0];
+            htmlObject.classList.remove("is-locked", "has-paywall");
+            htmlObject.style.overflow = "auto";
+          });
         </script>
+  tests:
+    - url: https://kw.be/nieuws/criminaliteit/geweld/in-40-jaar-nog-nooit-meegemaakt-chauffeur-meense-touroperator-krijgt-slag-in-gezicht-na-aanrijding-in-frankrijk/
+      test: document.querySelector('#paywall-modal') === null
 
diff --git a/rulesets/be/apache.yaml b/rulesets/be/apache.yaml
@@ -0,0 +1,19 @@
+- domains:
+  - apache.be
+  headers:
+    x-forwarded-for: none
+    cookie: cookiesjsr=%7B%22base%22%3Afalse%2C%22functional%22%3Afalse%2C%22video%22%3Afalse%7D
+  injections:
+    - position: head
+      append: |
+        <script>
+          document.addEventListener("DOMContentLoaded", () => {
+            // remove paywall
+            document.querySelector("#spb-block-apachepopupblock").remove();
+            // restore scrolling
+            document.body.style.overflow = "auto";
+          });
+        </script>
+  tests:
+    - url: https://apache.be/2024/03/25/chinese-spionkop-van-vlaams-belang
+      test: document.querySelector('#spb-block-apachepopupblock') === null
diff --git a/rulesets/be/roularta.yaml b/rulesets/be/roularta.yaml
