API Basics
The Flow Graph Engine APIs are used to run compute jobs in the Flow Platform, for example, simulation.
To use the APIs, upload data to cloud storage using the resource API. Then, use the job API to run job(s) that operates on that data. Afterward, monitor the state of the job(s) using the job API. Finally, download the results using the resource API or use them as input to other jobs.
A job contain one or more tasks, with each task specifying work to be performed by an executor. The executor performs the work and produces outputs that can be retrieved. Outputs can also form inputs to other tasks in the current job or to tasks in subsequent jobs.
The most common use cases for the Flow Graph Engine APIs can be found here.
The following table defines important terms for the Flow Graph Engine APIs:
Term | Definition |
---|---|
Bifrost | Bifrost is a technology that provides a node-based framework for building custom effects, including smoke, fire, explosions, sand, snow, instancing, and more.
Bifrost graph(s) can be built and edited in Maya using the Bifrost plugin.
More information on Bifrost can be found here.
|
executor | An executor indicate which compute capability should perform a given task. Currently the only supported executor is “bifrost”. |
storage provider | A storage provider is responsible for storing the data and enforcing access permissions. Currently only 2 storage providers, internal to the Flow Graph Engine APIs, are supported. |