STAR

The Search Tag & Analyze Resource for collaborative annotation and interpretation of disease from open digital samples from GEO.

Installing from source

1. Create an Ubuntu 16.04 instance.

2. Configure ssh connection by adding something like this to ~/.ssh/config:

   Host stargeo
     HostName 123.45.67.89
     User ubuntu
     IdentityFile /path/to/stargeo.pem
     IdentitiesOnly yes

3. Clone code locally:

   git clone [email protected]:HadleyLab/star-django.git stargeo
   cd stargeo

   Switch to branch:

   git checkout limited

4. Install deployment dependencies (unless you've already installed dev ones):

   We will use virtualenv and virtualenvwrapper to create isolated python environment, so start with:

   sudo pip install virtualenv virtualenvwrapper

   When create a virtualenv for our project and install dependencies:

   mkvirtualenv star
   pip install -r requirements-deploy.txt
   

5. Install and configure:

   fab install
   fab config  # Fill in all the keys here
   fab restart

   Config command will open ~/app/.env remote file in vim, alternatively you can ssh into instance and edit it directly. You will need to specify BIOPORTAL_API_KEY (get here), FROM_EMAIL and ADMIN.

6. Create a superuser for yourself (need one to log into admin):

   fab manage:createsuperuser

Making a working copy

Here are steps to make local deployment of this app in order to tinker it.

1. Get sources:

   git clone [email protected]:HadleyLab/star-django.git
   cd star-django

2. Install development dependencies (like above, different file):

   pip install -r requirements-dev.txt
   

3. Update settings:

   All settings that should vary by deployment go to .env file, so:

   cp .env.example .env
   <edit> .env

   Adjust settings in .env file, you will probably need to only set DATABASE_URL for your working copy.

4. Create or migrate database tables:

   ./manage.py migrate
   ./manage.py createsuperuser

5. Run it and have fun:

   ./manage.py runserver 5000

   Go to http://localhost:5000/ to see the app or to http://localhost:5000/admin/ to see admin panel.

   To debug background tasks you'll need to start celery:

   honcho start celery

   NOTE: it doesn't autorestart, you'll need to do that manually.

   To run both development web-server and celery in single terminal and autorestart both do:

   # .. install node.js and npm somehow
   npm install -g nodemon
   
   nodemon -x 'honcho start' -e py

Deploying

1. Configure ssh connection (see in install).

2. Install deployment dependencies (see in install).

3. Run locally to deploy latest commit:

   fab deploy

Restore backup

restore

docker compose -f <compose-config-file> up -d <db_service_name>
zcat < db_dump__arch_file.sql.gz | docker exec -i <db_container_name_or_id> psql -U <db_user> <db_name>