v0.5.0
As always, we’re out to squash bugs and enhance the developer experience.
Key changes/features:
- Triggers and flows have now been decoupled. Meaning that multiple triggers can be mapped to a single flow enabling a single flow to process logic from multiple sources (HTTP, Kafka, MQTT, an as an example).
- Flows as Functions: A flow exposes function-like constructs. A flow can define input and output parameters. This allows the flow to operate without any dependency on the trigger(s).
- Shim-style triggers. These are triggers that override the entry point of the framework.
- First-class support for AWS Lambda. A Flow can now be built to target running as a function within AWS Lambda.
- Native Google TensorFlow activity used to inference data against a tf.estimator protobuf model.
- Revised evaluation syntax. A simplified and revised syntax for evaluation mapping expressions.
- Object-type mapping. This enables a flow developer to construct a JSON response object with values from the flow context.