1. Add “core_elasticsearch_app” to your INSTALLED_APPS setting

INSTALLED_APPS = [
  ...
  'core_elasticsearch_app',
]

The package should be placed above core_explore_keyword_app in the INSTALLED_APPS to enable autocomplete powered by elasticsearch.

2. Include the core_elasticsearch_app URLconf in your project urls.py

re_path(r'^elasticsearch/', include('core_elasticsearch_app.urls')),

3. Deploy an instance of Elasticsearch

Recommended version: 7.14.1 See https://github.com/usnistgov/cdcs-docker#elasticsearch for help with Elasticsearch deployment.

4. Configure what should be indexed in Elasticsearch

It is recommended to configure the mapping between the CDCS templates and the Elasticsearch indices before starting to insert data, so that data can be indexed as soon as they are published in the CDCS. For that, set the following setting:

ELASTICSEARCH_AUTO_INDEX = True

Then, configure the mapping between the XSD templates and the Elasticsearch indices. Documents stored in Elasticsearch have a common structure. They are composed of a data_id, a title and a description field. A mapping can be defined to tell where in the XML data the values for the title and for the description of a resource can be found. Below is an example to set this mapping via REST, for a given template:

import requests
payload = {
    "template": "5f43ba192b8fd5c092e30e62",
    "title_path": "Resource/identity/title",
    "description_paths": ["Resource/content/description", "Resource/content/subject"],
}

requests.post(
    SERVER_URI + "/elasticsearch/rest/elasticsearch_template/", data=payload, auth=(USER, PASSWORD)
)