Read me!

Installation

1. Install Docker and Virtualbox

2. Create a local virtual machine named 'local' with Virtualbox

$ docker-machine create --driver=virtualbox local

3. Switch to the virtual machine's docker:

$ eval $(docker-machine env local)  

To later get back to your computer's docker :

$ eval $(docker-machine env -u)  

To check which docker is active :

$ docker-machine ls  

4. Get the ip of your virtual machine

$ docker-machine ip local

5. Init the Swarm with this ip.

docker swarm init --advertise-addr 192.168.99.100

6. Configure port forwarding on your virtual machine to enable sockets and debugger for development https://stackoverflow.com/questions/35372399/connect-to-docker-machine-using-localhost

Databases

Allow sparql update through endpoint in Virtuoso :

1. Grant update rights Access Virtuoso 192.168.99.100:8890, login to the conductor (local access codes should be dba/dba. Go to System Admin > User Accounts, edit SPARQL user, and select SPARQL_UPDATE as Primary Role. Save changes.

2. Enable CORS Go to Web Application Server > Virtual Domains and Directories, click on the first directory pic in the left column to unfold content, edit /sparql line, fill Cross-Origin Resource Sharing field with a *, and save changes.

Load data in Virtuoso :

Simple way

Access Virtuoso 192.168.99.100:8890, login to the conductor (local access codes should be dba/dba) and use ISQL pop up in the left menu to create graphs

> SPARQL CREATE GRAPH <http://nobel.ilda.fr>;
> SPARQL CREATE GRAPH <http://dbpedianobel.ilda.fr>;

Then use the top menu Linked Data > Quad Store Upload to upload files (available in our data repo)

Advanced procedure (for big files and / or bulk upload)

MAC

Once the containers are up and running (see below), copy files to upload in the dump folder, it is mapped to the containers' /usr/local/virtuoso-opensource/share/virtuoso/vad folder in the docker-compose file.

- ./dump:/opt/virtuoso-opensource/vad

LINUX

Once the containers are up and running (see below), connect to the VM

- docker-machine ssh local

and copy files to upload in the /home/docker/dumps folder, it is mapped to the containers' /usr/local/virtuoso-opensource/share/virtuoso/vad folder in the docker-compose file.

- /home/docker/dumps:/opt/virtuoso-opensource/vad

BOTH

Then get the id of your virtuoso container

$ eval $(docker-machine env local) 
$ docker container ls

And use it to start an isql console (replace 93211349efc4 with your container id)

$ docker exec -it 93211349efc4 bash
$ /opt/virtuoso-opensource/bin/isql 1111 dba dba

You can now use bulk upload (leave the full directory path as below, it is mapped to your dump folder)

> SPARQL CREATE GRAPH <http://geonames.ilda.fr>;
> ld_dir_all('/opt/virtuoso-opensource/database', '%.nt', 'http://geonames.ilda.fr');
> select * from DB.DBA.load_list;
> rdf_loader_run();
> DELETE FROM DB.DBA.load_list;

Interact with mongo

Once the containers are up and running (see below), get the id of your mongo container

$ eval $(docker-machine env local) 
$ docker container ls

And use it to start a mongo console (replace 61a76db18ade with your container id)

$ docker exec -it 61a76db18ade bash
$ mongo

You can now interact with mongo

> use discover
> db.resources.find()

Development

Switch to you local machine's docker (when needed) Note: If you don't the containers will run and be accessible on localhost, but the swarm won't be available, so you might have network problems, since the swarm is needed to have a local network config similar to the production's one.

$ docker-compose -f docker-compose-dev.yml up

This will run 3 containers :

• the app on port 80 : custom image derived from node:9.11-jessie
• mongo on port 27017 : jessie 3.6
• virtuoso on port 8890 : tenforce/virtuoso:1.3.1-virtuoso7.2.2

It will also mount a local volume that can be used for bulk uploads.

Simulate deployment on your local machine

First commit the version to deploy to git master branch. When tests are passed (step 1 in CI) and the new image is built and published to Gitlab registry (step 2 in CI)

$ docker stack deploy semanticStack --compose-file docker-compose.yml --with-registry-auth

Deploy

Commit to master (if you have not already done it to simulate deployment). When tests are passed (step 1 in CI) and the new image is built and published to Gitlab registry (step 2 in CI), activate Step 3 (deployment) manually.

Tests

$ npm test

Reducers

• config: computed possible configurations of the views, derived from the stats in the dataset + the views configuration
• data: json-ld final data when loaded
• dataset: definition of the data to visualize: entrypoint config, options and stats
• display: settings + calculated scaled values
• selections: elements selected by the user on a screen

Authors

• Marie Destandau
• Emmanuel Pietriga
• Caroline Appert

Many thanks to

• Silvio and Dylan
• Logilab: Adrien, Tanguy, Luc & co
• BnF: Aude, Raphaëlle and Sébastien & co
• Staff = technical support team @ LRI: Laurent & co
• Browserstack, for supporting opensource and thus granting us a free open source licence