Skip to main content

Introduction

OpenK9 is a powerful modular Open Source Cognitive Enterprise Search System, enhanced through AI and Machine Learning. Using state-of-art Information Retrieval and Machine Learning algorithms, OpenK9 is able to enrich and index data from different data sources.

How It Works#

The system is composed by several modules that make up a data processing pipeline.

  • Data Source plugins are called from a scheduler, and they are responsible for fetching new data and put it into the Queue;
  • A MQTT Queue handles ingestion backpressure, splitting the processing workload into a pool of workers for enrichment;
  • Several Enrich plugins process the data from the queue and extract info using AI;
  • The enriched document are then saved and indexed into a Search Index;
  • The entities found by the enrich pipeline are also persisted into a Knowledge Graph;
  • A Query Parser is able to convert complex queries into the language used to query the Search Index and the Knowledge Graph.

img

OpenK9 is extremely modular and easily extensible. A new data source can be connected quickly and efficiently.

Furthermore, OpenK9 is easily manageable: through the administration panel, it is very easy to create and configure a new search environment for a new data source.

Usage#

Once the data has been indexed we provide several ways to search into it.

  • Standalone Search App: we provide a standalone ready to use frontend to search into data;
  • Reusable UI Components: we provide several React components to be embedded into your application, that allow interacting with OpenK9 Search;
  • Headless API: OpenK9 exposes several REST endpoints that allow you to to search into entities and data using HTTP.