Beagle

Beagle is a backend service for managing files, pipelines and runs.

Beagle Responsibilities

• Users
  • Authentication using MSKCC LDAP
  • Every user will have same permissions
• Files
  • List files in Beagle DB
  • Search files (filename, metadata, file-type, file-group)
  • Create File in Beagle DB
• FileMetadata
  • Metadata is associated with file.
  • Metadata versioning. Changes are tracked, and can be reverted.
  • Metadata validation using JsonSchema.
• Pipelines
  • Using pipelines hosted on github
  • Creating RUNs from pipelines
• Run
  • Creating run (choosing pipeline, choosing inputs)
  • Submitting job to rabix executor
  • Receiving updates about job status from rabix
  • List outputs generated from run
• LIMS integration
  • Periodically fetch new samples from LIMS and create File objects in Beagle DB
  • Try to pair fails, and create runs
  • Notify if there are some errors with files or file metadata

beagle_cli.py

• Command line utility which helps handles authentication and accessing beagle endpoints.

Setup

• Requirements

  • PostgreSQL==11
  • RabbitMQ
  • python 3

• Instructions

  • virtualenv beagle

  • pip install -r requirements.txt

  • setup your environment using the environment page

  • python manage.py migrate

  • python manage.py runserver

• Async

  • Celery is used for scheduling tasks related to ETL from LIMS and submission to CWL Executor
  • celery -A beagle_etl beat -l info -f beat.log (starting the periodic task)
  • celery -A beagle_etl worker -l info -Q -f beagle-worker.log (starting the worker)
  • celery -A beagle_etl worker --concurrency 1 -l info -Q -f scheduler-worker.log
  • celery -A beagle_etl worker -l info -Q -f beagle-runner.log

Read more detailed specification on wiki page.

Development Instance

A development instance can be easily set up using conda with the following commands:

• Clone this repo:

git clone https://github.com/mskcc/beagle.git
cd beagle

• Install dependencies in the current directory with conda:

make install

• If using a m1 mac, install with:

  make install-m1

and activate the conda environment:

conda activate beagle

• Initialize the PostgreSQL database:

make db-init

• Initialize the Django database and set an admin ('superuser') account:

make django-init

• Start Postgres, RabbitMQ, and Celery servers:

make start-services

• Start the main Django development server:

make runserver

The included Makefile will pre-populate most required environment variables needed for Beagle to run, using default settings. These settings can be changed when you invoke make on the command line by including them as keyword args, for example:

make db-init BEAGLE_DB_NAME=db-dev

Some environment variables needed for full functionality are not included; you should save these separately and source them before running the Makefile. These variables are:

BEAGLE_LIMS_USERNAME
BEAGLE_LIMS_PASSWORD
BEAGLE_LIMS_URL
BEAGLE_AUTH_LDAP_SERVER_URI

Beagle can run without these, but it will not be able to access IGO LIMS and LDAP server for authentication.