This session takes you on a journey from zero to a fully functional hybrid search engine. We will walk through implementing Elasticsearch in Drupal, starting with the basics and advancing to a sophisticated solution that combines semantic understanding with rule-based precision for a healthcare platform.
We will start from the basics of Elasticsearch integration and build up to the advanced hybrid implementation, so no prior experience with Elasticsearch or semantic search is necessary.
The session follows a "Zero to Hero" narrative arc:
- The Foundation: We start with a fresh Drupal instance and connect it to Elasticsearch using the module elasticsearch_connector, explaining the basic architecture.
- The Semantic Layer: We teach the search engine to understand medical intent in Dutch by integrating a trained model through Elasticsearch ML/AI. This model is exposed to our Drupal instance and can be selected by the user who has the permission to access the form.
- The Rule-Based Layer: We implement strict logic to filter content based on specific medical contexts.
- The Hybrid Result: We combine these approaches into a single, high-performance API endpoint that powers the frontend.
In this session, we will learn how to design a search solution that layers semantic understanding (synonyms, medical classifications) on top of strict rule-based filtering to ensure both relevance and safety. At the end of the session, you would learn the basics of how to build a hybrid search in Elasticsearch and expose its form for custom configuration.
