Cookbook

Remove gifs or ad elements

filters:
  - remove_element:
      - 'img[src$=".gif"]'
      - ".ad, .banner"

Hacker News top links

- path: /hn-best.xml
  source: https://news.ycombinator.com/best
  note: Hacker News Top Links
  filters:
    - split:
        title_selector: ".titleline > a"
        link_selector: ".titleline > a"
        description_selector: "tr.athing + tr"

Inspect all fields in a post

  - path: /inspect.xml
    note: Utility for inspecting RSS feeds
    filters:
      - js: |
          function update_post(feed, post) {
            post.content = `<pre>${JSON.stringify(post, null, 2)}</pre>`;
            return post;
          }

Note you need to supply ?source=http://your-domain.com/feed.xml in the query string if you omit the source field in the endpoint definition.

Scan for magnet links and put into enclosure

filters:
  - js: |
      function find_magnet(text) {
        const regex = /\b(?<info_hash>[0-9a-fA-F]{40})\b/g;
        const match = regex.exec(text);
        if (match === null) {
          return null;
        }
        return `magnet:?xt=urn:btih:${match.groups.info_hash}`;
      }

      function update_post(feed, post) {
        if (post.enclosures) {
          return post;
        }
        if (post.description === null) {
          return post;
        }
        const magnet = find_magnet(post.description);
        if (magnet === null) {
          return post;
        }
        post.enclosure = {
           url: magnet,
           mime_type: "application/x-bittorrent",
           length: '1337'
        };
        return post;
      }

Subreddits post listing with comments

  - path: /reddit-comments.xml
    source: https://www.reddit.com/r/selfhosted.rss
    filters:
      - js: |
          function modify_post(feed, post) {
            const comment_regex = /<a\b[^>]*href="(?<url>[^"]+)"[^>]*>\[comments\]<\/a>/;
            const match = comment_regex.exec(post.content.value);
            if (match === null) {
              return post;
            }
            // redirect to old reddit
            const url = match.groups.url.replace(/www\.reddit\.com/, "old.reddit.com");
            post.links[0].href = url;
            return post;
          }
      - full_text:
          keep_element: .nestedlisting
          append_mode: true

Note that you can leave out the source field and supply any reddit's official rss via the ?source= parameter. e.g. http://example.com/reddit-comments.xml?source=https://old.reddit.com/r/selfhosted/top/.rss.

DeArrow YouTube feed

https://github.com/ajayyy/DeArrow is a crowdsourced database providing de-clickbait-ed YouTube title and thumbnail. The following config demonstrate the use of fetch API in modify_post to replace the YouTube feed's title and thumbnail.

  - path: /kurtzgesagt-dearrow.xml
    source: https://www.youtube.com/feeds/videos.xml?channel_id=UCsXVk37bltHxD1rDPwtNM8Q
    filters:
      - modify_post: |
          const video_id = post.id.split(':').pop();
          let resp = await fetch(`https://sponsor.ajay.app/api/branding?videoID=${video_id}`);
          if (resp.status == 200) {
            let [{title}] = resp.json().titles;
            post.title.value = title;
          }
          let thumbnail = post.extensions.media.group.filter(g => g.name == 'media:group')[0].children.thumbnail[0];
          thumbnail.attrs.url = `https://dearrow-thumb.ajay.app/api/v1/getThumbnail?videoID=${video_id}`;

Remove YouTube shorts

  - path: /kurtzgesagt-dearrow.xml
    source: https://www.youtube.com/feeds/videos.xml?channel_id=UCsXVk37bltHxD1rDPwtNM8Q
    filters:
      - modify_post: |
          if (post.title.value.includes('#shorts')) {
            return null;
          }