NSIDC Open Search (yet another OpenSearch)

The service currently exposes four endpoints:

• /OpenSearchDescription - OpenSearch Description Document (accessing this endpoint provides descriptions for each other endpoint)
• /OpenSearch - dataset search, used by NSIDC Search and the ADE
• /Facets - facets to go with the dataset search
• /suggest - retrieve suggested phrase completions for auto-suggest/auto-complete

For the most part, these endpoints support queries following the OpenSearch and relevant OpenSearch Extension specifications. However, they each take an additional optional URL parameter that is not in the standard, source. This service was originally built to support both NSIDC Search and the Arctic Data Explorer (ADE) in a single instance of the service, using the same Solr store for metadata. The source parameter provided a simple way to distinguish between the two client applications. ADE is now decommissioned, but the source value of NSIDC may still be used by the NSIDC data search application.

Installation & Usage

Requirements

• Ruby (see .ruby-version for the required version)
• Bundler
• Ubuntu 14.04 (only required if setting up the Upstart service)

Configuration

config/app_config.yaml contains settings for different deployment environments. Each environment other than development merges in the common settings. You may need to modify some settings in the appropriate section in this file:

• relative_url_root determines the path to access the service, e.g., with a value of /api/dataset/2, you can access the running service at http://localhost:3000/api/dataset/2
• solr_url is the URL of the Solr core containing your data
• solr_auto_suggest is the URL of the Solr core containing autocomplete data

Simple Setup

1. Clone the source code - git clone git@github.com:nsidc/dataset-search-services.git
2. cd dataset-search-services
3. Install the dependencies - bundle install
4. Modify the development section in config/app_config.yaml so that it has the values you need (if RACK_ENV is set to something other than development, the section matching its value will be used)
5. bundle exec rake run starts the service on port 3000
6. The service can be queried via curl or in the browser with an OpenSearch request, e.g., http://localhost:3000/OpenSearchDescription, which will provide a list of endpoints provided by the service and their required parameters

As a Systemd service (requires Ubuntu 16.04 or higher)

tasks/deploy.rake contains rake tasks to set up and start the service on a Vagrant VM hosted on NSIDC's internal network. To set up the service outside of this network, follow these steps:

1. Clone the source code - git clone git@github.com:nsidc/dataset-search-services.git
2. cd dataset-search-services
3. Set APP_PATH in config/app_config.rb to where you want the code to be run
   • the following steps assume the environment variable APP_PATH is set to the same value used in config/app_config.rb
4. mkdir -p $APP_PATH
5. cp -R . $APP_PATH
6. cd $APP_PATH; bundle install
7. mkdir -p $APP_PATH/run/log
8. Set ENV to your desired environment (development, integration, qa, etc.)
9. Configure the appropriate section in $APP_PATH/config/app_config.yaml
10. echo $ENV > $APP_PATH/config/environment
11. Create a search_services.service file in /etc/systemd/system. See the example file below, which assumes the $APP_PATH is /opt/search_services
    • The User and Group should be changed to whatever user/group will own the service
12. Create a puma run file:
    • sudo mkdir /etc/search_services
    • Create a file /etc/search_services/puma.rb file; see below for an example. As above, this example assumes the $APP_PATH is /opt/search_services
13. sudo systemctl start search_services.service
    • the service must be able to write to /var/log and /var/run/puma (the full paths of the relevant files can be seen in config/app_config.yaml)
14. The service can be queried at using the port and relative_url_root set in $APP_PATH/config/app_config.yaml
    • For example, if port is 10680 and relative_url_root is /api/dataset/2, you can access the OpenSearch Description Document provided by the service at http://localhost:10680/api/dataset/2/OpenSearchDescription.

Example search_services.service file

[Unit]
Description=search_services - puma application

[Service]
Type=forking
WorkingDirectory=/opt/search_services
RuntimeDirectory=search_services
PIDFile=/var/run/search_services/puma.pid
User=www
Group=www

# The command to start puma
ExecStart=/usr/local/bin/bundle exec puma -C /etc/search_services/puma.rb
Restart=on-failure

[Install]
WantedBy=multi-user.target

Example /etc/search_services/puma.rb file

Configure environment, port, etc, according to your needs.

directory '/opt/search_services
environment 'development'
daemonize
pidfile "/var/run/search_services/puma.pid"
stdout_redirect "/var/log/search_services.puma.stderr.log", "/var/log/search_services.puma.stdout.log"
threads 1, 1
bind 'tcp://0.0.0.0:10680'
plugin 'tmp_restart'
restart_command "bundle exec puma"
workers 2
preload_app!

Changes

Notes on changes can be found in CHANGELOG.md.

Developer Info

Instructions and notes for developing this project are in DEVELOPMENT.md.

How to contact NSIDC

User Services and general information: Support: http://support.nsidc.org Email: nsidc@nsidc.org

Phone: +1 303.492.6199 Fax: +1 303.492.2468

Mailing address: National Snow and Ice Data Center CIRES, 449 UCB University of Colorado Boulder, CO 80309-0449 USA

License

Every file in this repository is covered by the GNU GPL Version 3; a copy of the license is included in the file COPYING.