Staying on top of the latest research is a task which is hard in itself. Although arXiv supports some RSS-feeds, I read all my RSS feeds through newsboat, and given that it is (of course) a CLI application, it cannot display the math tags of the abstracts.

Luckily someone else also found this to be tedious and created the arXiv sanity preserver, some kind of news aggregator for the latest papers on arXiv in predefined categories. And, as he is no bad person, he open sourced the whole thing, to my pleasure. The application is easy to use and intuitive, with the ability to save papers to one's personal library, which requires a really simple account setup (you just type in a username and password once, that's all).

Setup was straightforward, following the instructions on github, using virtualenv for python, which I did not know about until now. However, I did not want to fiddle around with ports or so to make flask serve on :80, since my nginx is already listening there, so I put up a reverse proxy:

   server {
      listen 80;
      server_name sanity.marcelwunderlich.com;
      access_log /var/log/nginx/arxiv-sanity.access_log main;
      error_log /var/log/nginx/arxiv-sanity.error_log info;
      location / {
         proxy_pass http://localhost:5000;
      }
   }

The flask based server runs without the production flag in some screen of my user account. The only modifications I had to made to the code were the categories I want to display (which are math.AP, math.DG, math.SP and math.MG at the moment), as well as replacing app.debug = True to app.debug = False for non production mode, for obvious reasons. As one can guess, now there is my own sanity preserver running on my own domain. ¹ Feel free to use it, although I do not guarantee for any availability or regular updates, as updating is at the present stage highly non automatic. The only thing I dislike is that the application does not work without javascript, but at least I can trust javascript served from my own machines.