VIVO on Rails

This is the code for the new VIVO front-end for Researchers@Brown. This is a Ruby on Rails application that searches directly against the VIVO's Solr index. See the General Architecture information below for more details.

Pre-requisites

You need to have Solr running. We use the Solr index that VIVO provides out of the box but we have added several fields needed by this project. See Solr Index section below.

We are currently using Ruby 2.7.1, Rails 6.0.3, and MySQL.

brew install ruby-install
brew install chruby
ruby-install ruby 2.7.1
source /usr/local/opt/chruby/share/chruby/chruby.sh
chruby 2.7.1
brew install mysql
gem install bundle

To get started

git https://github.com/Brown-University-Library/vivo-on-rails.git
cd vivo-on-rails
source .env_sample
bundle config set path 'vendor/bundle'
bundle install
bundle exec rake db:create
bundle exec rake db:migrate
bundle exec rails server

Update the values in .env_sample to match the URLs where Solr is running in your environment.

General Architecture

This is a Ruby on Rails web site that shows faculty information stored in Solr. All searches and retrieval of information are done against Solr.

In the future, access to the Fuseki triple store will be used to perform linked data queries and other visualizations.

A few diagrams on how the project is structure can be found here

Solr Index

The site assumes that the Solr index has Faculty information stored in a structure like the following:

  {
    id: "http://vivo.brown.edu/individual/jane_researcher",
    record_type: "PEOPLE",
    affiliations: ["dept1", "dept2", ... "deptN"],
    research_areas":["r1", "r2", "r3", ... "rN"],
    json_txt: "# string representation of a JSON object with the all the data (see below)"
  }

Below is a sample of the data stored in json_txt for a sample PEOPLE document. Notice that the field is a string with the JSON of the data for the given document.

"{
  \"record_type\":\"PEOPLE\",
  \"affiliations_text\":\"\",
  \"affiliations\":[
    {
      \"uri\":\"http://vivo.brown.edu/individual/org-brown-univ-dept57\",
      \"name\":\"Pathology and Laboratory Medicine\",
      \"id\":\"http://vivo.brown.edu/individual/org-brown-univ-dept57\"
    }
  ],
  \"awards\":\"\",
  \"collaborators\":[],
  \"contributor_to\":[],
  \"education\":[
    {
      \"school_uri\":\"http://vivo.brown.edu/individual/n37819\",
      \"date\":\"1980\",
      \"degree\":\"BA\",
      \"school_name\":\"Mount Holyoke College\"
    },
    {
      \"school_uri\":\"http://vivo.brown.edu/individual/n67738\",
      \"date\":\"1985\",
      \"degree\":\"MD\",
      \"school_name\":\"State University of New York\"
    }
  ],
  \"email\":\"[email protected]\",
  \"funded_research\":\"\",
  \"name\":\"Hansen, Katrine\",
  \"org_label\":\"Pathology and Laboratory Medicine\",
  \"overview\":\"\",
  \"research_overview\":\"\",
  \"research_statement\":\"\",
  \"scholarly_work\":\"\",
  \"teacher_for\":[],
  \"teaching_overview\":\"\",
  \"title\":\"Associate Professor of Pathology and Laboratory Medicine\",
  \"thumbnail\":null,
  \"research_areas\":[],
  \"uri\":\"http://vivo.brown.edu/individual/khansenm\",
  \"id\":\"http://vivo.brown.edu/individual/khansenm\"
}"

Organization records follow a similar pattern.

  {
    id: "http://vivo.brown.edu/individual/some-org-id",
    record_type: "ORGANIZATION",
    json_txt: "# string representation of a JSON object with the all the data (see below)"
  }

Below is a sample of the data stored in json_txt for a sample ORGANIZATION. Notice that the field is a string with the JSON of the data for the given document.

"{
  \"record_type\":\"ORGANIZATION\",
  \"uri\":\"http://vivo.brown.edu/individual/org-brown-univ-dept57\",
  \"name\":\"Pathology and Laboratory Medicine\",
  \"overview\":...",
  \"thumbnail\":\"http://vivo.brown.edu/individual/n5457\",
  \"people\":[
  {
    \"id\":\"http://vivo.brown.edu/individual/khansenm\",
    \"uri\":\"http://vivo.brown.edu/individual/khansenm\",
    \"label\":\"Hansen, Katrine\",
    \"specific_position\":\"Associate Professor of Pathology and Laboratory Medicine\",
    \"general_position\":\"Faculty Position\"
  },
  {
    \"id\":\"http://vivo.brown.edu/individual/elaposat\",
    \"uri\":\"http://vivo.brown.edu/individual/elaposat\",
    \"label\":\"\",
    \"specific_position\":\"Clinical Associate Professor of Pathology and Laboratory Medicine\",
    \"general_position\":\"Faculty Position\"
  }]
}

We use json_txt to reconstruct the data that will be displayed when the user wants to view the detail for a specific individual or organization. This is what allows us to run the system without interfacing with the Fuseki endpoint.

TODO: We still need to create individual fields in Solr to allow for a better search experience, for example, a separate field for name to give it a higher boost on search results and the ability to facet the data by other fields.

Caveats

The code at this point is hard-coded for very specific predicates that we use at Brown and might not work on your particular VIVO database.

The unit tests (bundle exec rake vivo:tests) are hard-coded to find very specific people/departments so they will very likely fail on your installation.