This repository enables the deployment of the noctua editor (which includes minerva, barista, golr and noctua) locally, using self-generated GO-CAM models (in the form of .ttl files).

Note: For production deployment see this readme file

• important files:
  • build_images.yaml
    • builds all the docker images necessary to spin up a local instance.
  • push_images.yaml
    • pushes images to dockerhub if planning on staging to a remote machine.
  • stage.yaml
    • Tasks are executed on staging machine which can be local or remote.
    • The stages directory is created on build.

• blazegraph journal
• Solr Index
• Cloned repositories including:
  • noctua-form, noctua-landing-page, noctua-models, go-site.
• docker-compose and configuration files are generated from templates.

• The docker images created by this repository have been successfully tested using:

  • MacOs (10.15.3)
  • Docker (19.03.5)
  • Docker Compose (1.25.2)
  • Ansible (2.10.3), Python (3.8.5), docker (4.3.1)

• Resources:

  • Docker should be given at least 3 CPUs and 8G RAM. (on mac see Docker Preferences | Resources)
  • python 2.7 should work as well.

There are two ways to install the requirements necessary for building this application stack:

To install miniconda go to url

This will install ansible, docker ansible module, docker-compose, and terraform. See dependencies/bare-metal.yml

conda env create --file dependencies/bare-metal.yml
conda activate noctua_app_stack

# when done
conda deactivate

# delete a conda environment:
conda env remove  -n noctua_app_stack

The ansible docker plugin is used to buid docker images.

pip install ansible
pip install docker 

The playbook build_images.yaml clones minerva, noctua and amigo git repositories and builds corresponding docker images. The default branch used is master.

Note: To change the branch being cloned, see repo_map in docker-vars.yaml

ansible-playbook -e docker_hub_user=xxxx build_images.yaml
docker image list | egrep 'minerva|noctua|golr'

• These can also be set on command line using the -e flag.
  • Barista:
    • uri
    • username
    • password

# on Mac:
export HOST=`ipconfig getifaddr en0`
ansible-playbook -e "host=$HOST" -i "localhost," --connection=local stage.yaml

Two docker-compose files are staged:

• docker-compose-golr.yaml
  • Uses a lightweight solr image for golr
• docker-compose-amigo.yaml
  • Uses the official geneontology/amigo-standalone for golr

# assuming stage_dir is in current directory and docker-compose-golr.yaml is used:
docker-compose -f stage_dir/docker-compose-golr.yaml up -d

# minerva takes a long time to start up the first time
# Tail minerva logs to see its progress
docker-compose -f stage_dir/docker-compose-golr.yaml logs -f minerva
# Or tail all logs
docker-compose -f stage_dir/docker-compose-golr.yaml logs -f

# When minerva is ready all other services should be up
docker-compose -f stage_dir/docker-compose-golr.yaml ps

• Use http://localhost:8080 if default noctua_proxy_port was used

docker-compose -f stage_dir/docker-compose-golr.yaml down
# kill works faster ...
docker-compose -f stage_dir/docker-compose-golr.yaml kill
#delete containers:
docker-compose -f stage_dir/docker-compose-golr.yaml rm -f

If staging the stack to a remote machine:

1. Create an account on dockerhub if you do not have one.
2. Add a public dockerhub repository named minerva, noctua and golr.
3. set docker_hub_user in docker-vars.yaml or simply
4. use the -e option when using ansible-playbook command.
5. Refer to this document on provisionning an AWS.

• Skip this step if planning on staging locally.
• Refer to this document on provisionning an instance on AWS.

• Skip this step if planning on staging locally.

ansible-playbook -e docker_hub_user=xxxx push_images.yaml