Export new WMF P/T jobs to a public rss feed

Author: brouberol

ansible/roles/gallifrey/page/files/rssgen.py

@@ -0,0 +1,35 @@
+#!/usr/bin/env python3
+import argparse
+import sys
+
+"""Generate an RSS feed template"""
+
+RSS_TPL = """
+<?xml version="1.0" encoding="UTF-8" ?>
+<rss version="2.0">
+
+<channel>
+  <title>{title}</title>
+  <link>{link}</link>
+  <description>{description}</description>
+  {items}
+</channel>
+
+</rss>
+"""
+
+def parse_args():
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument('--title', help='Feed title')
+    parser.add_argument('--link', help='Feed link')
+    parser.add_argument('--description', help='Feed description')
+    return parser.parse_args()
+
+
+def main():
+    args = parse_args()
+    print(RSS_TPL.format(title=args.title, link=args.link, description=args.description, items=sys.stdin.read()))
+
+
+if __name__ == "__main__":
+    main()

ansible/roles/gallifrey/page/files/wmfeed.py

@@ -0,0 +1,27 @@
+#!/usr/bin/env python3
+import sys
+
+from bs4 import BeautifulSoup
+
+ITEM_TPL = """
+<item>
+  <title>{title}</title>
+  <link>https://boards.greenhouse.io/{link}</link>
+  <description>{description}</description>
+</item>
+"""
+
+soup = BeautifulSoup(sys.stdin, features='html.parser')
+items = []
+
+
+for link in soup.find_all('a'):
+    items.append(
+        ITEM_TPL.format(
+            title=link.text,
+            link=link.attrs['href'],
+            description=link.text
+        )
+    )
+
+print('\n'.join(items))

ansible/roles/gallifrey/page/tasks/main.yml

@@ -72,3 +72,35 @@
       "Redeploy personal page after each birthday"
       "{{ page_repo }}/hooks/post-receive"
     user: "{{ git_user }}"
+
+- name: "Install htmlq"
+  unarchive:
+    src: https://github.com/mgdm/htmlq/releases/download/v0.4.0/htmlq-x86_64-linux.tar.gz
+    dest: /usr/local/bin
+    remote_src: yes
+
+- name: "Install rssgen"
+  copy:
+    src: rssgen.py
+    dest: /usr/local/bin/rssgen
+    group: git
+    mode: 0775
+
+- name: "Install wmfeed"
+  copy:
+    src: wmfeed.py
+    dest: /usr/local/bin/wmfeed
+    group: git
+    mode: 0775
+
+- name: "Schedule WMF jobs rss export"
+  cron:
+    name: "wmf jobs rss"
+    minute: "0"
+    hour: "4"
+    job: >
+      curl -s https://boards.greenhouse.io/wikimedia/ |
+      htmlq 'div.opening[data-department-91046=true] > a' |
+      wmfeed |
+      rssgen --title 'WMF Jobs' --link 'https://boards.greenhouse.io/wikimedia' --description 'P/T jobs at the WMF' > {{ page_web_dir }}/public/wmfjobs.rss
+    user: git