AKakkio
integrations

Controller Handle Preview Predict

Represents a user providing connection details and a table of interest, and us giving them a preview of what the data transform looks like. The user is expected to follow this up with a "Predict Commit" request representing a user asking for an ad-hoc import and persistent prediction schedule.

post/integrations/preview/predict

Request

  • Base URL
    https://api.akkio.com (the document declares no server; this is the origin it was published from)
  • URL
    https://api.akkio.com/integrations/preview/predict
  • Auth
    none declared

Request body

flow_idstring required

Flow ID these predictions will be run off of. Important to specify because this is how we track down the exact model we'll use for predictions.

connection_idstring nullable

Connection ID to use for credentials. If not provided, we use the connection attached to the flow (older meltano deployments)] = None.

specifierobject required

The table specifier specifying the table we'd like a preview of. We pull it down and run N predictions on it, then return the results.

mappingsobject

Field mappings, useful incase you are trying to run on a table with the same information but different column names. Keys of the object are keys on the table we'll be running predictions on, Values of the object are keys on our training table.

step_titlestring nullable

Which deployment user-friendly name (Snowflake, Postgres, etc) is running

step_keystring nullable

Which deployment key (postgre-sql-output, etc) is running

Response

Successful Response

dataset_idstring required

Dataset created during the preview process.

input_table_specifier_idstring required

The table specifier of where the source data for predictions is.

target_table_specifier_idstring required

The table specifier of where we'll be placing predictions.

Changes

No recorded changes to this operation across all 4 revisions of this API.