What’s new in Hume 2.14?
by Danica Stankovic & Dimitrios Stamatis
· 4 min read
We are proud to announce the release of Hume version 2.14. This release introduces the first iteration of the new federated search functionality and the Hume API. We also made various improvements to Hume Orchestra and Visualisations.
In Hume 2.14 you can:
- Store attributes of a single node in Elasticsearch and Neo4j with the first iteration of Federated Search.
- Test our REST API (in Beta).
- Customise your visualisation to your preferences with the Keep Positions on Expand option and adjustable Node Compactness.
- Use String Interpolation in Cypher queries.
- And much more.
Store Large Text Attributes in Elasticsearch
With Hume 2.14, we are introducing the first iteration of our ongoing federated search initiative. In this release, attributes can be stored in Elasticsearch, and then accessed and searched for in the visualisation canvas. This feature is particularly suited to storing large text attributes in Elasticsearch while the rest of the node is maintained in Neo4j, allowing for better performance.
Hume API Beta
2.14 introduces the Hume API beta. The Hume API replaces the experimental Configuration-as-Code feature and serves as a way to configure Hume programmatically. The API can be used in workflows and for further integration. Currently, the Hume API supports Resources, Roles, Permission Groups, Knowledge Graphs, Schema, Perspectives, Actions and Workflows.
In this release we are introducing several entirely new options that allow you to customise the appearance and behaviour of nodes in your visualisations. Keep Position on Expand is a new functionality which allows users to maintain their node position after reloading data using FORCE LAYOUT. The option also controls the automatic adjustment of nodes after the time filter is enabled in temporal analysis. This helps you not to lose track of your analysis during data exploration.
Node Compactness is another new feature which makes exploration easier. You can now set the distance between nodes on your canvas to your preference, and adjust it throughout your analysis.
- Finding your visualisation is even easier with the filter option in the visualisation list. Search by visualisation name, or filter by the perspective used, the user who created the visualisation, the date/time the visualisation was created, or the date/time when it was last used.
- The Export PNG option allows you to choose the resolution of your PNG image. You can also choose whether to export the entire visualisation or only the selection visible on your screen.
In this video you can see how you can use a filter to find your visualisation, adjust node compactness, and Keep the Position of your nodes.
Increased Python Support
The Orchestra Docker image has been updated to allow users to mount a volume with their own Python scripts which can be imported in message transformers or filters. The image can also be extended to add Python packages.
New Parallel Multicast Component
Hume 2.14. introduces a new component to Orchestra - the Parallel Multicast Component. This component enables the parallel processing of outgoing branches from a component, meaning that all outgoing components will be processed at the same time rather than one by one.
Use String Interpolation in Cypher Queries
String interpolation can be used in queries where part of the query needs to be dynamically replaced based on the data in the incoming message, or when parameters are not accepted in a query.
- The Orchestra workflow scheduler has been improved to include timezone support. You can now set the timezone according to which the scheduler will work, as well as whether it should take daylight savings changes into account.
- Metrics such as message counts, message and failure contents as well as event logs can now be persisted when using the db storage engine for Orchestra allowing users to inspect the information even after restarting Orchestra.
- The Generic Enricher (HTTP) component now has the ability to set an incoming message field value as the full content of the body and accepts empty http responses.
In this video you can see how messages are processed faster when Parallel Multicast is implemented, and how String Interpolation allows you to dynamically set any part of the query.
Further sources of knowledge
If you’re interested in learning more about how graphs and Hume deliver high value solutions…
Read our blog to find out how you can Optimise Logistic Chains with Hume.
Learn more about Data-driven Orchestration with Graphs.
Or read this introductory blog to Understand the Power of Graph Visualisation.