Projects

Project

A Kognic project must first be set in order to create inputs. Projects are usually configured by the Kognic Professional Services team, during the Guideline Agreement Process (GAP) of a new client engagement.

List Projects

All existing projects within your organization can be listed by using the Python API KognicIOClient. Make sure the authentication are set (see Authentication).

from kognic.io.client import KognicIOClient
client = KognicIOClient()
projects = client.project.get_projects()

Alternatively, projects can be listed with the kognicutil command line interface (CLI)

kognicutil projects

Batch

Input batches allow further grouping of inputs into smaller batches within a project. Specifying batch during the input creation is optional, and will otherwise be the latest open batch by default.

Batch Status

Status	Description
pending	Batch has been created but is still not fully configured by Kognic. Either during project setup or requested changes
open	Batch is open for new inputs
ready	Batch has been published and no longer open for new inputs.
in-progress	Kognic has started annotation of inputs within the batch.
completed	Annotations has been completed.

Listing Batches

All existing batches withint a project can be listed by using the Python API KognicIOClient.

from kognic.io.client import KognicIOClient
client = KognicIOClient()
project_batches = client.project.get_project_batches(project="<project_external_id>")

Alternatively, batches can be listed with the kognicutil CLI

kognicutil projects <project_external_id> --get-batches

Creating Batches

To create a new batch in the open state within a project

from kognic.io.client import KognicIOClient
client = KognicIOClient()
project_batch = client.project.create_batch(
  project="<project_external_id>",
  batch="<batch_external_id>",
)

The newly created batch will contain the same Annotation Types as the latest batch by default.

This method has an optional flag publish_previous_batches which defaults to False. By setting this flag to True, as shown in the example below, all previous batches in the "open" state would be published and you would no longer be able to upload new inputs to those batches. You should therefore be certain that you no longer need to upload more inputs to the previous batches if you use this flag.

from kognic.io.client import KognicIOClient
client = KognicIOClient()
project_batch = client.project.create_batch(
  project="<project_external_id>",
  batch="<batch_external_id>",
  publish_previous_batches=True,
)

Contact Kognic before use

Kognic usually helps with creating batches before a client becomes autonomous, in order to avoid any confusion please contact Kognic before you start using this feature.

Publish Batch

from kognic.io.client import KognicIOClient
client = KognicIOClient()
project_batch = client.project.publish_batch(
  project="<project_external_id>",
  batch="<batch_external_id>",
)

When the input batch is published, the status of the batch will be set to "ready". Published batches are not open for new inputs any longer. A project with multiple open batches will require you to specify which open batch to target when creating new inputs, whereas a project with a single open batch will allow you omit the batch parameter when creating inputs.