Skip to content

Latest commit

 

History

History
73 lines (44 loc) · 3.56 KB

README.md

File metadata and controls

73 lines (44 loc) · 3.56 KB
Raw

A Django-powered REST API for powering the data transformation and tagging pipeline for Print & Probability, a joint project between Carnegie Mellon University and University of California San Diego to identify typesets in early modern printed books.

The frontend is powered with VueJS, and styled via vue-bootstrap.

Images are served as IIIF, currently using go-iiif-vips as it can serve both JPEG as well as TIFF images on demand.

Architecture

  1. PostgreSQL database
  2. Django REST API
  3. IIIF image server
  4. Vue frontend
  5. nginx reverse proxy

Local development

DB and API

docker-compose.yml defines a local development environment running Postgres as well as the Django API.

You must have a .env file in the same directory to supply configuration variables. Run cp .template.env .env and edit the file.

The Makefile defines almost all tasks needed for setting up and tearing down the development environment.

make will start up the service, which includes building the docker images if needed, creating a blank database pp in Postgres if it does not yet exist, and then starting the Django application to listen on http://localhost:80/api/. Most other make tasks require the service to be running in order to work. make down will shut down the containers and remove the Docker network (but data in the associated volumes will persist.)

When building the service for the first time in your local development environment, also run make redeploy once the system is running, in order to copy all the static files (html, css, and js) that Django uses to the correct Docker volume. This is only necessary on first startup - as long as you keep the docker volumes, the static files will stay in place in between invocations of make and make down

make loadtest/make dumptest loads and outputs test data defined in app/pp/fixtures/test.json. This test data has fake values for most fields, so note that the tif paths for images in the test data are not valid URLs and will display locally as broken.

To access the API, log in to the test account at http://localhost/api/auth/login:

username: root password: print&probability

make wipe will drop the current database and recreate it, re-running migrations to create the necessary tables.

make redeploy will run all migrations against the current DB, and will refresh all static assets.

To run Django other management commands such as creating and running migrations, make attach will shell you into the Django container. make db will shell you into Postgres.

Frontend

Currently using Node 16.14.0 and npm 8.3.1

cd vue
npm install
npm run serve

The local frontend will be available at http://localhost:4000. You must first log in to the test account at http://localhost/api/auth/login to test the frontend locally.

For building for deployment, you must create vue/.env and define the envvars

VUE_APP_PP_ENDPOINT=https://printprobdb.psc.edu/api
VUE_HOME=https://printprobdb.psc.edu

These will properly set the API endpoint and the home page location for the frontend when you run npm run build.

Updating dependencies

We use poetry to pin python package versions and calculate all dependencies for local development, however the docker container itself (and production deployment) only reads requirements.txt. Use poetry to update this file automatically:

poetry export -f requirements.txt -o requirements.txt --without-hashes

Contact

Maintained by Matthew Lincoln