optimize deployment protocol

[andrewsu]

The original deployment protocol to https://sulab.github.io/DrugMechDB/ was based entirely on github actions. But then the data got so big that we overran the time limits for github actions in the jekyll build step, so we had to do the jekyll build locally and commit/push the rendered HTML files. That process is described here. But now, the data is even so big that we are right at the edge of what Github pages will serve without complaining (1GB total site size). (EDIT: now at 1.6 GB, and the deployment is failing.)

Thankfully I think there is a really obvious solution. It looks like >90% of that size is due to the HTML code for the side navigation bar, and that is repetitively duplicated for every HTML page on our site. So each of our 5000 HTML files is 2.8MB, and all but ~15kb of that is the navigation bar. So that could definitely be optimized (either with a jquery load or a server-side include?). This issue tracks the exploration and implementation of an alternate deployment protocol.

[andrewsu]

Perhaps web components would be helpful, as described in https://daverupert.com/2017/07/jekyll-includes-are-cool/ ?