eibio

Biographies made easy, with neo4j and node express.

(person)-[:employed_as]->(activity)
(institution)-[:appears_in]->(activity)
(role)-[:appears_in]->(activity)

##installation on UNIX download and extract neo4j database server in a proper folder. Follow the neo4j installation documentation, then configure the the ~/neo4j/conf/neo4j-server.properties file: fill the neo4j org.neo4j.server.database.location with your database folder path, adjust the server port org.neo4j.server.webserver.port. In order to change the password of the neo4j default user via the neo4j browser, modify the org.neo4j.server.webserver.address if you do not have access to localhost. Setup a new password for the default user neo4j and you're done.

Run neo4j with

~/neo4j/bin/neo4j start

Once neo4j is functional, clone the eibio package, then install the required npm dependencies.

git clone https://github.com/CVCEeu-dh/eibio.git
cd eibio
npm install

Copy settings.js.example to settings.js, then modify neo4j section according to your configuration.

Add the following to your neo4j.properties file

# Autoindexing

# Enable auto-indexing for nodes, default is false
node_auto_indexing=true

# The node property keys to be auto-indexed, if enabled
node_keys_indexable=name_search

Open neo4j shell in order to configure node_auto_index as full_text index

  ~/neo4j/bin/neo4j-shell
...
$ index --create node_auto_index -t Node
  $ index --set-config node_auto_index type fulltext
...
$ index --get-config node_auto_index

Execute the setup script that flushes all the constraints to the neo4j db

npm run-script setup

Then test it

npm test

then if everything is fine, you can start the eibio app:

npm start

to run eibio from another port than 8000, change the value of the port variable in settings.js

Eibio API service should now be available in localhost http://localhost:8000/api

If you want to user forever to serve the application

forever start --uid "eibio" --sourceDir /path/to/eibio server.js >>  /path/to/eibio/logs/eibio.log 2>&1

External sources

Eurovoc basename (name disambiguation in english) e.g european parliament -> http://eurovoc.europa.eu/2243

DBPEDIA api

EIBIO makes use of dbpedia API, via the discover method of each models/<model>.js. First of all and if no wiki_id has been provided, we look for the autosuggest which responds with a list of porobable matches in dbpedia

<ArrayOfResult xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns="http://lookup.dbpedia.org/">
<Result>
<Label>European Parliament</Label>
<URI>http://dbpedia.org/resource/European_Parliament</URI>
<Description>
The European Parliament (abbreviated as Europarl or the EP) is the directly elected parliamentary institution of the European Union (EU). Together with the Council of the European Union (the Council) and the European Commission, it exercises the legislative function of the EU and it has been described as one of the most powerful legislatures in the world.
</Description>

Once the URI has been loaded, it can be enriched with the corresponding dbpedia resource, e.g http://dbpedia.org/page/European_Parliament

Some interesting queries

1. Get the number of job activities per role: MATCH (n:`role`)--(t:activity) WITH n, count(*) as total RETURN n.name, total ORDER BY total DESC

2. Get the number of job activities per institution: ``` MATCH (n:institution)--(t:activity) WITH n, count(*) as total RETURN n.name, total ORDER BY total DESC

3. Get the network of activities and person (direct "colleagueness")

     MATCH (t:activity)-[r2]-(p:person) RETURN t,r2,p LIMIT 10000;
   

4. Get the network of closest institutions

   MATCH (ins:institution {slug: 'european-commission-bel'})-[r1:appears_in]->(act:activity)-[r3]-(per:person)-[r4]-(act2:activity)<-[r2:appears_in]-(ins2:institution)
   

RETURN ins, r2, act, per, act2, r3, r4, ins2 ```

1. Find some direct collegue of a specific person MATCH (per:person {slug:"wim-duisenberg"}) OPTIONAL MATCH (per)-[r1:employed_as]->(act:activity)-[r2:employed_as]-(per2:person) WHERE id(per) <> id(per2) WITH r1, act, per, { slug: per2.slug, colleagueness: count(DISTINCT act), dt: min(abs(r1.start_time - r2.start_time)) } as colleague RETURN colleague ORDER BY colleague.colleagueness DESC

2. Manually attach an institution to an activity

     MATCH(act:activity {slug:'professor-at-yale-university'}), (ins:institution {slug:'yale-university-usa'})
     WITH act, ins
     MERGE (ins)-[r:appears_in]->(act)
     RETURN act, ins