Skip to content

Latest commit

 

History

History
183 lines (145 loc) · 8.53 KB

DEVELOPMENT.md

File metadata and controls

183 lines (145 loc) · 8.53 KB
Raw

Development Guidelines

Develop with Docker (recommended quickstart)

This is the simplest configuration for developers to start with.

Initial Setup

  1. Run docker compose run --rm django ./manage.py migrate
  2. Run docker compose run --rm django ./manage.py createcachetable
  3. Run docker compose run --rm django ./manage.py createsuperuser --email $(git config user.email) and follow the prompts to create your own user. This sets your username to your git email to ensure parity with how GitHub logins work. You can also replace the command substitution expression with a literal email address, or omit the --email option entirely to run the command in interactive mode.
  4. Run docker compose run --rm django ./manage.py create_dev_dandiset --owner $(git config user.email) to create a dummy dandiset to start working with.

Run Application

  1. Run docker compose up
  2. Access the site, starting at http://localhost:8000/admin/
  3. When finished, use Ctrl+C

Application Maintenance

Occasionally, new package dependencies or schema changes will necessitate maintenance. To non-destructively update your development stack at any time:

  1. Run docker compose pull
  2. Run docker compose build --pull --no-cache
  3. Run docker compose run --rm django ./manage.py migrate

Develop Natively (advanced)

This configuration still uses Docker to run attached services in the background, but allows developers to run Python code on their native system.

Initial Setup

  1. Install Docker and Docker Compose
  2. Run docker compose -f ./docker-compose.yml up -d
  3. Install Python 3.11
  4. Install psycopg2 build prerequisites. Example psycopg2 installation on Ubuntu 20.04:
sudo apt install libpq-dev
export PATH=/usr/lib/postgresql/X.Y/bin/:$PATH
pip install psycopg2
  1. Create and activate a new Python virtualenv
  2. Run pip install -e .[dev]
  3. Run source ./dev/export-env.sh
  4. Run ./manage.py migrate
  5. Run ./manage.py createcachetable
  6. Run ./manage.py createsuperuser --email $(git config user.email) and follow the prompts.
  7. Run ./manage.py create_dev_dandiset --owner $(git config user.email) to create a dummy dandiset to start working with.

Run Application

  1. Ensure docker compose -f ./docker-compose.yml up -d is still active
  2. Run:
    1. source ./dev/export-env.sh
    2. ./manage.py runserver
  3. Run in a separate terminal:
    1. source ./dev/export-env.sh
    2. celery --app dandiapi.celery worker --loglevel INFO --without-heartbeat -Q celery,calculate_sha256,ingest_zarr_archive -B
  4. When finished, run docker compose stop

Remap Service Ports (optional)

Attached services may be exposed to the host system via alternative ports. Developers who work on multiple software projects concurrently may find this helpful to avoid port conflicts.

To do so, before running any docker compose commands, set any of the environment variables:

  • DOCKER_POSTGRES_PORT
  • DOCKER_RABBITMQ_PORT
  • DOCKER_MINIO_PORT

The Django server must be informed about the changes:

  • When running the "Develop with Docker" configuration, override the environment variables:
    • DJANGO_MINIO_STORAGE_MEDIA_URL, using the port from DOCKER_MINIO_PORT.
  • When running the "Develop Natively" configuration, override the environment variables:
    • DJANGO_DATABASE_URL, using the port from DOCKER_POSTGRES_PORT
    • DJANGO_CELERY_BROKER_URL, using the port from DOCKER_RABBITMQ_PORT
    • DJANGO_MINIO_STORAGE_ENDPOINT, using the port from DOCKER_MINIO_PORT

Since most of Django's environment variables contain additional content, use the values from the appropriate dev/.env.docker-compose* file as a baseline for overrides.

Testing

Initial Setup

tox is used to execute all tests. tox is installed automatically with the dev package extra. To install the tox pytest dependencies into your environment, run pip install -e .[test]. These are useful for IDE autocompletion or if you want to run pytest directly (not recommended).

When running the "Develop with Docker" configuration, all tox commands must be run as docker compose run --rm django tox; extra arguments may also be appended to this form.

Running Tests

Run tox to launch the full test suite.

Individual test environments may be selectively run. This also allows additional options to be be added. Useful sub-commands include:

  • tox -e lint: Run only the style checks
  • tox -e type: Run only the type checks
  • tox -e test: Run only the pytest-driven tests
  • tox -e test -- -k test_file: Run the pytest-driven tests that match the regular expression test_file
  • tox -e test -- --cov=dandiapi: Generate a test coverage report
  • tox -e test -- --cov=dandiapi --cov-report=html: Generate an HTML test coverage report in the htmlcov directory

To automatically reformat all code to comply with some (but not all) of the style checks, run tox -e format.

Profiling with Memray

To include a memory profile with your tests, add --memray at the end of your test command invocation. For example, to run a memory profile with all tests, you would run tox -e test -- --memray. This can be used in conjunction with other pytest CLI flags (like -k) as well. See the pytest-memray docs for more invocation details.

NOTE: If you have an existing dandi-archive installation in which you have previously run tox, you may need to recreate the tox environment (by adding -r to your tox invocation) the first time you attempt to use memray. If your attempt to use the --memray flag fails with pytest: error: unrecognized arguments: --memray, this is likely why.

dandiarchive.org WEB Interface

This repository now also contains sources for the web interface under web/ folder. If you would like to develop it locally, please see web/README.md file for instructions.

API Authentication

Read-only API endpoints (i.e. GET, HEAD) do not require any authentication. All other endpoints require token authentication to call.

Creating a Token

Visit the URL /admin with a web browser, logging in with the credentials entered during the createsuperuser setup step. Then go to /swagger and use GET /auth/token end-point.

Supplying the Token

In API endpoint calls, add the Authorization HTTP header with a value of Token <token_value>. For example, for a token 1234, include the header: Authorization: Token 1234.

Scripts

For frequent deployment administration tasks, django-extensions provides a convenient way to write and run scripts that execute in the Django context.

create_dev_dandiset

python manage.py create_dev_dandiset --owner $(git config user.email) --name My Dummy Dandiset

This creates a dummy dandiset with valid metadata and a single dummy asset. The dandiset should be valid and publishable out of the box. This script is a simple way to get test data into your DB without having to use dandi-cli.

import_dandisets

python manage.py import_dandisets [API_URL] --all

This imports all dandisets (versions + metadata only, no assets) from the dandi-api deployment living at API_URL. For example, to import all dandisets from the production server into your local dev environment, run python manage.py import_dandisets https://api.dandiarchive.org from your local terminal. Note that if a dandiset with the same identifier as the one being imported already exists, that dandiset will not be imported.

python manage.py import_dandisets [API_URL] --all --replace

Same as the previous example, except if a dandiset with the same identifier as the one being imported already exists, the existing dandiset will be replaced with the one being imported.

python manage.py import_dandisets [API_URL] --all --offset 100000

This imports all dandisets (versions + metadata only, no assets) from the dandi-api deployment living at API_URL and offsets their identifiers by 100000. This is helpful if you want to import a dandiset that has the same identifier as one already in your database.

python manage.py import_dandisets [API_URL] --identifier 000005

This imports dandiset 000005 from API_URL into your local dev environment. Note that if there is already a dandiset with an identifier of 000005, nothing will happen. Use the --replace flag to have the script overwrite it instead if desired.

Abbreviations