TheHaloMod

The web-app for calculating Halo Model quantities.

This is the source code for TheHaloMod, which uses the hmf and halomod Python package to calculate halo model quantities.

The web-app is powered by Django and uses Bootstrap 4 as a CSS framework.

To acknowledge use of this app, cite Murray, Power & Robotham (2013).

This app began as HMFcalc back in 2013, and has been thoroughly reworked and expanded to provide full halo model capabilities.

Contributing

Clearly, this repository is intended for developers (users shouldn't need to see this code, or I've done something wrong!). I would love contributions to this app from the community -- whether in the form of bug reports, feature requests, or new code. Please consider getting involved!

If you'd like to get involved but have never done any web-development before (like me when I started this project), I'd suggest taking the Django Beginner's Tutorial to wrap your head around things a bit.

After that, here's a bit of a primer of how this repo is laid out:

The top-level is reserved for interesting project-level stuff like this README, and scripts for managing the website itself (things you call manually when you're on the server, outside the context of the web-app itself).

The TheHaloMod directory contains the website-level settings and structure (eg. where urls point to).

The templates directory contains the HTML files (really, they're just kinda-sorta HTML, they're actually templates) that define how each page of the website will be structured (and often, the text on the website).

Finally, the most important directory: halomod_app. This contains the logic of the web-application, as a bunch of python files. Most important here are the views.py, which defines the various things that happen when different URLs are accessed (not always a web-page opening -- sometimes a download of a file, or a form displaying etc.), and the forms.py which defines the (fairly large) form into which inputs for the HMF are given.

Running the Server Locally

To set up locally, you should install poetry, then run poetry install in the top-level directory. This will install all dependencies. To add another dependency, run poetry add <dep> then make sure you do poetry export -f requirements.txt -o requirements.txt and commit the output file.

To run the dev server, you should simply have to do docker-compose -f compose-dev.yml up --build. This should open a browser at the correct address.

There are .envs/base and .envs/local files that are read, and do not contain any sensitive information (in production, there are other files here that do have sensitive info, and are not stored in VCS). You may modify these if need be.

Deployment

I'm gonna use this space to remind myself how I go about deployment for TheHaloMod. For local dev, just use the above instructions.

To deploy:

1. Make a branch
2. Make some changes
3. Commit
4. Push to github
5. Ensure everything's working, then merge to master
6. Log in to the server (galileo.sese.asu.edu)
7. Do a git pull in the repo
8. Run ENV=env/test PORT=8010 docker-compose -f compose-prod.yml up --build to build the test server.
9. (Manually!) go to the test server and make sure things are OK.
10. Run ENV=env/production PORT=8000 docker-compose -f compose-prod.yml up --build to build the production server.

Secret settings are set in the .envs/ directory, and not kept in VCS for obvious reasons. This .env/ directory is in the repo on the server itself, and my local copy on my own computer. In order to keep it safe, it's also backed up. The different compose files essentially just take different .envs/ files in, and set a couple of different env variables.

Some things to do if the actual server has to be changed:

1. Fix up the references to SSL certificates in nginx/default.conf
2. Need to copy the local versions of .envs/test and .envs/production onto the server.