1.37 Release Notes
Server Enhancements
Large Language Model/AI Support
The Vantiq platform now supports interactions with Large Language Models (LLM) to develop AI powered applications. It also offers a Semantic Search capability which combines the use of LLMs and a vector database to support asking questions about application specific content (this is used to implement the new AI Documentation Search feature outlined below). This feature introduces two new system resources:
- llms – represents a specific LLM which can be used to perform actions based on its type:
- embedding – used to create vector embeddings for Semantic Search.
- generative – used to generate responses based on user input.
- semanticindexes – represents a collection of user provided content over which Semantic Searches can be performed.
Visual Service Event Handlers
- The Notify Activity Pattern has a new option called “firstResponseFilter” which is a WHERE clause for the firstResponse event. If specified, the firstResponse event will only be raised if the response passes the filter.
- All reservoirs for the Analytics pattern now accept Reals. Existing apps may need to cast the results to Integers.
- Escalations have been reintroduced with the Escalate pattern.
- The Submit Prompt and Answer Question activity patterns were added to support use of the new AI features.
- The Assign Activity Pattern has a new roleType called
result
which supports assigning a value to an arbitrary result property in the current collaboration.
VAIL
Added the following built-in services:
Edge Nodes
- Nodes that are defined using the WebSocket scheme along with the
vqs
option for back-communication (typically from cloud to edge) now support the reconnect
property.
- Catalogs can now support edge nodes and VQS connections. See here for information on enabling this for catalogs. See here for information on connecting as an edge node.
- The CLI can now be downloaded from the Edge edition, using the Help -> Developer Resources menu/link.
- Added Edge Vision server setup documentation.
Kubernetes Management and Deployment
- Configurations may now include items of type
hardAffinity
. The annotation
and label
items may now include the targetResource
property. See Deploying a K8s Installation for details.
- The Vantiq system may enhance K8sInstallations deployed to the
self
cluster with hardAffinity
and/or label
specifications. See Using the self Cluster for details.
- Service connectors deployed via Kubernetes support additional properties (see link for details).
Sources
REMOTE
- Support for sending multipart messages has been added. See REMOTE Source.
- Support for
asFullExchange
option in SELECT WITH clause to help troubleshoot HTTP request/response exchange.
AMQP
- The AMQP source now allows offset to be considered when reconnecting to Azure service bus via AMQP, using the selector AMQP configuration option. See documentation at Azure Event Hubs partition offset.
User Authorization
CLI
- New command “run” and “stop” have been added for running tests and testsuites. You can also run procedures using the run command. See CLI Reference Guide.
The “execute” command for invoking procedures is deprecated and may be removed in a future release.
- New commands to support loading content into a semantic index and support backup and restore of index content. See CLI Reference Guide.
UI Enhancements
Modelo
1.37 Release Notes
Server Enhancements
Large Language Model/AI Support
The Vantiq platform now supports interactions with Large Language Models (LLM) to develop AI powered applications. It also offers a Semantic Search capability which combines the use of LLMs and a vector database to support asking questions about application specific content (this is used to implement the new AI Documentation Search feature outlined below). This feature introduces two new system resources:
Visual Service Event Handlers
result
which supports assigning a value to an arbitrary result property in the current collaboration.VAIL
Added the following built-in services:
io.vantiq.text.Template
– generates text from templates and substitution values.io.vantiq.ai.LLM
– facilitates LLM interactions.io.vantiq.ai.SemanticSearch
– perform semantic searches and manage optional conversation state.Edge Nodes
vqs
option for back-communication (typically from cloud to edge) now support thereconnect
property.Kubernetes Management and Deployment
hardAffinity
. Theannotation
andlabel
items may now include thetargetResource
property. See Deploying a K8s Installation for details.self
cluster withhardAffinity
and/orlabel
specifications. See Using the self Cluster for details.Sources
REMOTE
asFullExchange
option in SELECT WITH clause to help troubleshoot HTTP request/response exchange.AMQP
User Authorization
CLI
UI Enhancements
Modelo
// change widget property
// upload file