You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Right now, the PDF versions and HTML distributions of papers/reports we write are manually-generated. This is absolutely not scalable and means any change has to be made meticulously by hand (and there are a lot of manual tweaks/hacks that had to be added to make everything work). We would much prefer an automated process, roughly something like this:
Every paper is added to a papers.yml config file which specifies the name of each of the papers, which we write in TeXmacs out of convenience (because LaTeX is incredibly arduous to set up locally, requires installing a lot of packages even with a modern distribution unless you use an experimental LaTeX engine like Tectonic, and also has so many commands and such a...unique syntax that it's easy to forget, making writing papers incredibly slow)
This also allows us to specify that some papers be drafts that won't be built by the build scripts, giving us more flexibility in authoring papers/technical reports.
The PDF is generated by running TeXmacs in headless mode to produce a PDF
The web version is generated by (again) running TeXmacs in headless mode, then applying our manual customizations to the output HTML via a Python script/bash script
We can also optionally generate LaTeX files for each paper by (again) running headless TeXmacs and we can save a zip archive of all of the LaTeX outputs of every paper by using a GitHub action to push that zip archive to github releases
This might also be helpful for submissions to journals
The whole process is done by GitHub actions to automatically build the web and PDF papers on every commit, which avoids needing to add those paper HTML/PDF versions into the repo itself (which would mean cluttering the repo with heavy HTMLs and PDFs on every commit, which Git is not designed to diff efficiently)
The text was updated successfully, but these errors were encountered:
Right now, the PDF versions and HTML distributions of papers/reports we write are manually-generated. This is absolutely not scalable and means any change has to be made meticulously by hand (and there are a lot of manual tweaks/hacks that had to be added to make everything work). We would much prefer an automated process, roughly something like this:
papers.yml
config file which specifies the name of each of the papers, which we write in TeXmacs out of convenience (because LaTeX is incredibly arduous to set up locally, requires installing a lot of packages even with a modern distribution unless you use an experimental LaTeX engine like Tectonic, and also has so many commands and such a...unique syntax that it's easy to forget, making writing papers incredibly slow)The text was updated successfully, but these errors were encountered: