Skip to content

Latest commit

 

History

History
273 lines (188 loc) · 8 KB

for-developers.md

File metadata and controls

273 lines (188 loc) · 8 KB
Raw

For Developers

Prerequisites

This application can run on a single Docker host using docker-compose. (recommended for development environments). For production, see the cloudformation directory for AWS Elastic Container Service stack templates.

Running Concordia

Docker Compose

git clone https://github.com/LibraryOfCongress/concordia.git

If you're intending to edit static resources, templates, etc. and would like to enable Django's DEBUG mode ensure that your environment has DEBUG=true set before you run docker-compose up for the app container. The easiest way to do this permanently is to add it to the .env file:

echo DEBUG=true >> .env
Install the application virtual environment

These steps only need to be performed the first time you setup a fresh virtualenv environment:

  1. Ensure that you have the necessary C library dependencies available:

    • libmemcached
    • postgresql
    • node & npm for the front-end tools

  2. Ensure that you have Python 3.8 or later installed

  3. Install pipenv either using a tool like Homebrew (brew install pipenv) or using pip:

    pip3 install pipenv

  4. If you encounter errors installing psycopg, you may need to set LDFLAGS in your environment variables.

  5. Let Pipenv create the virtual environment and install all of the packages, including our developer tools:

    pipenv install --dev

    n.b. if libmemcached is installed using Homebrew you will need to set the CFLAGS long enough to build it:

    CFLAGS=$(pkg-config --cflags libmemcached) LDFLAGS=$(pkg-config --libs libmemcached) pipenv install --dev

    Once it has been installed you will not need to repeat this process unless you upgrade the version of libmemcached or Python installed on your system.

  6. Configure the Django settings module in the .env file which Pipenv will use to automatically populate the environment for every command it runs:

    echo DJANGO_SETTINGS_MODULE="concordia.settings_dev" >> .env

    You can use this to set any other values you want to customize, such as POSTGRESQL_PW or POSTGRESQL_HOST.

    n.b to allow a local server to connect to the dockerized db set POSTGRESQL_PORT=54323 - the db containers external postgres port.

  7. Make sure that redis is installed and running.

Local Development Environment

You will likely want to run the Django development server on your localhost instead of within a Docker container if you are working on the backend. This is best done using the same pipenv-based toolchain as the Docker deployments:

Python Dependencies

Python dependencies and virtual environment creation are handled by pipenv.

If you want to add a new Python package requirement to the application environment, it must be added to the Pipfile and the Pipfile.lock file. This can be done with the command:

pipenv install <package>

If the dependency you are installing is only of use for developers, mark it as such using --dev so it will not be deployed to servers — for example:

pipenv install --dev django-debug-toolbar

Both the Pipfile and the Pipfile.lock files must be committed to the source code repository any time you change them to ensure that all testing uses the same package versions which you used during development.

Launching the environnment

In order to successfully launch the environment, the environment variables POSTGRESQL_PW and DJANGO_SETTINGS_MODULE must be set. POSTGRESQL_PW may be set to any value (which will become the database password for the environment), but DJANGO_SETTINGS_MODULE should be set to concordia.settings_dev to use the development settings file.

export POSTGRESQL_PW=password
export DJANGO_SETTINGS_MODULE=concordia.settings_dev
cd concordia
docker-compose up

Browse to localhost

Setting up a local development server

See section - Ensuring your work follows the Library's coding standards in How We Work
Start the support services

Instead of doing docker-compose up as above, instead start everything except the app:

docker-compose up -d db redis importer celerybeat

This will run the database in a container to ensure that it always matches the expected version and configuration. If you want to reset the database, simply delete the local container so it will be rebuilt the next time you run docker-compose up: docker-compose rm --stop db.

Install front end

  1. Install Node 12. If you're on MacOS, you can install it using brew:

    brew install node@12

  2. Use NPM to install our development tools:

    npm install

  3. In another terminal, start Gulp to watch for changes to the SCSS files and compile them to CSS:

    npx gulp

    If you only want to compile them a single time without live updates:

    npx gulp build

  1. You may need to manually create a logs directory.

    mkdir logs

  2. Collect Django static files:

    pipenv run ./manage.py collectstatic
Start the application server

  1. Apply any database migrations:

    pipenv run ./manage.py migrate

  2. Start the development server:

    pipenv run ./manage.py runserver

Running the unit tests

Use the settings_local_test Django settings in your environment. Your .env file should look something like:

POSTGRESQL_PW=password
DJANGO_SETTINGS_MODULE=concordia.settings_local_test

Bring up the docker database and redis servers:

docker-compose up -d db redis

Then execute the tests:

pipenv run ./manage.py test

Import Data

Once the database, redis service, importer and the application are running, you're ready to import data. First, create a Django admin user and log in as that user. Then, go to the Admin area (under Account) and click "Bulk Import Items". Upload a spreadsheet populated according to the instructions. Once all the import jobs are complete, publish the Campaigns, Projects, Items and Assets that you wish to make available.

Data Model Graph

To generate a model graph, make sure that you have GraphViz installed (e.g. brew install graphviz or apt-get install graphviz) and use the django-extensions graph_models command:

dot -Tsvg <(pipenv run ./manage.py graph_models concordia importer) -o concordia.svg

Other Front-End Tools

Public-facing URLs

  1. If you need a list of public-facing URLs for testing, there's a management command which may be helpful:

    pipenv run ./manage.py print_frontend_test_urls

Accessibility testing using aXe

Automated tools such as aXe are useful for catching low-hanging fruit and regressions. You run aXe against a development server by giving it one or more URLs:

npx axe-cli --show-errors http://localhost:8000/
pipenv run ./manage.py print_frontend_test_urls | xargs npx axe-cli --show-errors

Static Image Compression

When you update any of the files under concordia/static/img, please use an optimizer such as ImageOptim or Caesium to losslessly compress JPEG, PNG, SVG, etc. files.

brew cask install imageoptim
open -a ImageOptim concordia/static/img/