This is the main application for the ArtsAPI project.

NB: Every time $ is seen in a code snippet, you should type this into a terminal.

1. Install Fuseki.
2. Copy the fuseki_config.ttl file from the /config folder.
3. In the file, change the value next to tdb:location to the location of your tdb folder, for example /Users/<your-user>/tdb_data/artsapi-dev.
4. Add alias fuseki_artsapi="cd $HOME/jena-fuseki-1.0.1 ; ./fuseki-server --config=fuseki_config.ttl" to your shell to start Fuseki using $ fuseki-artsapi. Change jena-fuseki-1.0.1 in the previous command to the version and folder of your installed version of Fuseki.
5. Install redis and run it.
6. Use bundle exec sidekiq -c 1 to bring up sidekiq for background processing.
7. Fill in the correct details in /db/seeds.example, rename the file to seeds.rb and use rake db:seed to generate a user.

NB: You must be running Fuseki on port 3030 if you do not want to modify the multiple tenancy code. If you want to use the Dropbox upload features, you will also need a copy of the ArtsAPI Grafter project on your hard drive so that the Rails app can call leiningen to run pipelines. Put the absolute path to this directory (e.g. /Users/jeff/artsapi-email-processing-tool) in the grafter_config.example file and then rename it grafter_config.rb.

In production, you will use env vars to declare dropbox credentials. In development, you will need to find the file in /config/initializers called dropbox.example - add your credentials for Dropbox in here and rename it dropbox.rb. Make sure you do not commit this, as you will have to change production.rb to not expect env vars with the locations provided (see the Docker/Deployment) section below.

If you wish to use Foreman to manage development (we encourage this), then add the locations into the Procfile.example file, and rename it Procfile. Then, change out of this directory and $ gem install foreman; you will now be able to use foreman start to bring up development dependencies. Make sure you have development unicorn configuration at ./config/unicorn_development.rb

Although effort has been made to try and make the application thread-safe, we cannot guarantee thread safety; hence you will need to pass the -c 1 option to Sidekiq when running.

Deploy using Docker. Setup a Redis, Mongo and Fuseki instance before linking them to this container.

In order to run the application and create uploads, the env vars DROPBOX_APP_KEY and DROPBOX_APP_SECRET need to be set. Pass these in when running the container using the -e Docker flag. You will need to make sure you have registered for the Dropbox API in order to get these.

1. Use the Docker docs' advice to set up swap, or running the other containers will not work:

   1. Log into Ubuntu as a user with sudo privileges.

   2. Edit the /etc/default/grub file.

   3. Set the GRUB_CMDLINE_LINUX value as follows:

      GRUB_CMDLINE_LINUX="cgroup_enable=memory swapaccount=1"

   4. Save and close the file.

   5. Update GRUB.

      $ sudo update-grub

   6. Reboot your system.

2. Use the iptables rules in the /docker folder in this repo to set up the firewall.

3. Install Nginx on the host

4. Install Docker edge (for other installation options, see the Docker docs):

   $ curl -s https://get.docker.io/ubuntu/ | sudo sh

1. Start a Fuseki instance. If you have any TDB data to mount, mount it internally so it sits in the /tdb_data/artsapi folder inside the Fuseki Docker container. You will also need to mount the config file found in this repo at config/production_config.ttl. Some examples:

   • Assuming the config file has been saved to /var/lib/fuseki-config/config.ttl and no TDB data exists:

     $ sudo docker run -d --name artsapi-fuseki -v /var/lib/fuseki-config:/opt/fuseki/config thefrey/fuseki:latest

   • Assuming the config file has been saved to /var/lib/fuseki-config/config.ttl and TDB data exists at /var/lib/artsapi-data/database

     $ sudo docker run -d --name artsapi-fuseki -v /var/lib/fuseki-config:/opt/fuseki/config -v /var/lib/artsapi-data/database:/data/artsapi thefrey/fuseki:latest

2. Start a MongoDB instance.

   $ sudo docker run --name artsapi-mongo -d mongo

3. Start the application. Note that you can also set ADMIN_MAILER_PASSWORD in production.rb and SENTRY_KEY in sentry.rb and thus avoid having to pass them through.

   $ sudo docker run -d --name artsapi-test \
       --link artsapi-mongo:mongodb \
       --link artsapi-fuseki:artsapi-fuseki \
       -v /var/lib/artsapi-assets:/artsapi-assets \
       -p 127.0.0.1:1955:8080 \
       -e DROPBOX_APP_KEY=<dropbox-app-key> \
       -e DROPBOX_APP_SECRET=<dropbox-app-secret> \
       -e ADMIN_MAILER_PASSWORD=<mandrill-password-here> \
       -e SENTRY_KEY=<sentry-key> \
       <artsapi-image>
   

4. Make sure you have secured your server! In the example above, you would want Nginx on the host to proxy to http://localhost:1955.

   • The nginx.conf in the /docker directory contains some good defaults
   • To use this, first generate a stronger DHE parameter using $ cd /etc/ssl/certs && openssl dhparam -out dhparam.pem 2048
   • You can then copy the file over the existing nginx.conf in /etc/nginx and $ sudo service nginx restart

NB: Remember, to inspect running containers, you can use sudo docker exec -i -t <container-name> /bin/bash -l.

• Fuseki
• MongoDB

Accounts needed:

1. Dropbox
2. Mandrill
3. Sentry

In development, rake db:seed will create an example user; however, the data bootstrap task will try to infer the logged-in user's organisation in the data using the registered email, so if the organisation whose data is in the system is called WidgetCorp, with a URL of http://www.widgetcorp.org and a mailserver at the same host, you will want to create a user on the command line (using an email @widgetcorp.org) rather than using the rake task, as this will allow you to use the data bootstrapping/background processing class methods on Organisation.

In production, you will need to make sure you have correctly set up your Fuseki config and then ensure that the user's email matches the host that their data will use in the system. See the development instructions above for more details on that.

In production, so long as you have seeded your User correctly for each tenant on the system, the bootstrap data tasks will seed the database with the concept schemes and resources you require.

In development, you will need to upload the concept scheme for keywords; this is in the ArtsAPI Grafter Project, in the doc folder. Upload these to the named graph of the concept scheme using Fuseki's s-put tool.

1. cd into your Fuseki installation directory
2. Load the concept scheme resources ./s-put http://localhost:3030/artsapi-dev/data 'http://data.artsapi.com/graph/keywords' '/path/to/artsapi-graft/doc/keywords_concept_scheme.ttl'
3. Load the keyword resources ./s-put http://localhost:3030/artsapi-dev/data 'http://data.artsapi.com/graph/keywords' '/path/to/artsapi-graft/doc/keywords_resources.ttl'
4. Load the SIC resources and extensions ./s-put http://localhost:3030/artsapi-dev/data 'http://data.artsapi.com/graph/sic' '/path/to/artsapi/lib/sic2007.ttl'

MIT - see the LICENCE file.