PrankWeb deploy with Docker

What are you going to learn here:

• how to run PrankWeb locally using Docker compose

Prerequisities

• Install Docker You can check using command:

  docker info

• WARNING: Note that this build will most likely not work with M1 Apple processors as the HMMer software does not support this processor architecture. See this issue.

Steps

• Clone git repository to empty directory.

  git clone https://github.com/cusbg/prankweb.git .

• Create Docker mounts. Please update the paths to the tmp directory to reflect your setup (this setup means that for example the predictions will be saved to the /tmp/predictions directory on your computer).

  docker volume create --name prankweb_rabbitmq --opt type=none --opt device=/tmp/rabbitmq --opt o=bind
  docker volume create --name prankweb_conservation --opt type=none --opt device=/tmp/conservation --opt o=bind
  docker volume create --name prankweb_predictions --opt type=none --opt device=/tmp/predictions --opt o=bind
  docker volume create --name prankweb_services --opt type=none --opt device=/tmp/services --opt o=bind
  docker volume create --name prankweb_docking --opt type=none --opt device=/tmp/docking --opt o=bind

  If you are running using Linux system, keep in mind that the directories must be writable for user 5988:5988. Alternatively variables UID and GID to specify user before building the images.
• Optional step. Create .env file in the directory and set variables:
  • WEB_SERVICE_USER
  • WEB_SERVICE_PASSWORD
  • RABBITMQ_DEFAULT_USER
  • RABBITMQ_DEFAULT_PASS
  • UID
  • GID
• Build images.

  docker compose build

• Optional step. Download conservation file(s) using following command this may take a while. Be aware that you need about 30GB of space. Note that this is optional, but when not installed, conservation scores are not available.

  docker compose run --rm executor-p2rank python3 /opt/hmm-based-conservation/download_database.py

• Finally, you can start the application using.

  docker compose up

• You should be able to see the frontend at localhost:8020.