Designing solutions that leverage the Quantexa Entity Store often requires a flexible approach, particularly when you need to test or retrieve Entities without fully ingesting a Document into the platform. In such scenarios, Quantexa’s Search Related Entities REST API provides an efficient means to query Entity Store, returning insights on how a Document would bind to Entities and Compounds if it were to be ingested.
Introduction
This article explores a case study that demonstrates how to search for Entities without using Quantexa Kafka Data Ingest services. It illustrates a situation where the client already holds the required payload to use Search Related Entities API.
The Search Related Entities service enables you to see which Entities and Compounds would be bound if the Document were ingested. This can be crucial for workflows that need to confirm relevance before proceeding with ingestion, reducing overhead and accelerating decision-making.
Key consideration
The process of Entities being resolved and ingested into the Entity Store is beyond the purpose of this article. For more information on how the Entity Store works, please refer to the Entity Store on Quantexa's Documentation Site.
Case study - Search Without Ingest
In this scenario, the upstream system is capable of producing a Document payload that directly matches the Search Related Entities API schema. This negates the need for any intermediate transformations or lookups.
Process flow
- Entity Search request: The upstream system sends a REST request containing a fully formed payload to the Search Related Entities endpoint. Quantexa’s Search Related Entities API analyses the payload, extracting the relevant fields to identify potential Entity matches.
- Extract compounds and keys: The platform identifies any compounds or keys that would form part of the resolved Entity structure if the Document were ingested.
- Search Entity Store: The Quantexa Platform queries Elasticsearch/OpenSearch (via the Entity Store indices) to locate matching Entities and relationships.
- Return Entities: The API returns a response indicating which Entities (and associated Compounds) would bind to the submitted Document payload.
- Decision point: The upstream system retrieves Entities from the API response.
Benefits
- Minimal overhead: No need for custom transformation or Kafka ingestion if the payload already aligns with the required schema.
- Fast feedback loop: Enables a quick determination of whether a Document has meaningful matches in the Entity Store.
- Reduced infrastructure complexity: Bypasses the Kafka Data Ingest tier when ingestion is not immediately required.
Design considerations and best practices
- Security and authentication
- Use a Quantexa API key for each REST request to ensure secure access.
- Scalability
- Evaluate how frequently you will call the Search Related Entities API and whether your system can handle the throughput.
- Monitor resource usage on both the Quantexa Platform and Elasticsearch/OpenSearch for potential bottlenecks.
- Decision-making workflow
- Clearly define the criteria that determine if a Document should be ingested based on the Entities returned by the Search Related Entities API.
- Document your process for updating or discarding Documents following the search results.
Conclusion
The Search Related Entities REST API expands Quantexa’s flexibility by allowing you to query Entity Store relationships without fully ingesting Documents. The API offers a streamlined path to evaluating Entity matches.
