Details
-
Backlog Task
-
Resolution: Fixed
-
Minor
-
None
-
None
-
None
-
None
-
Small
Description
Goal of this task is to ensure we are compatible with the studio. Few pointers:
- data stream doesn't need a design API since it relies on dynamic data sampling to fulfill that "lack"
edit: checked out with [stream] team and schema handling can arrive but is not yet planned or needed.
- data preparation doesn't really need a design API since it can use the sampling as well
edit: [preparation] can probably migrate to use dynamic schema evaluation (TDP-4521)
- data catalog can need some design API but it will likely be more linked with the UI completion and data navigation than a dedicated API
edit: [datalog] can infer the schema from the sampling as well so no need of a dedicated API
- the studio requires a design schema definition (input/output)
edit: the [stream] logic sounds doable to start
High level idea: try to do the same as for data stream into the studio and avoid to require any schema definition for component developers.
Note: there is an optional @DiscoverSchema action in the framework which can allow to define explicitly a schema, it requires a better Schema API definition (don't use avro API but create our own API more centered around our usage)
Outcome of this task is a component family (source and processor) fully working into the studio without a lot of new API - note: only relevant if creating design API.