Toloka documentation

create_tasks

toloka.client.TolokaClient.create_tasks

Creates several tasks in Toloka using a single request.

Tasks can be added to different pools. You can add together regular tasks and control tasks. You can send a maximum of 100,000 requests of this kind per minute and a maximum of 2,000,000 requests per day.

By default, create_tasks starts asynchronous operation internally and waits for the completion of it. Do not change async_mode to False, if you do not understand clearly why you need it.

Parameters Description

Parameters Type Description
tasks List[Task]

List of tasks to be created.
allow_defaults Optional[bool]

Overlap setting:

  • True — Use the overlap value that is set in the defaults.default_overlap_for_new_task_suites pool parameter.
  • False — Use the overlap value that is set in the overlap task suite parameter.
open_pool Optional[bool]

Open the pool immediately after creating a task suite, if the pool is closed.
skip_invalid_items Optional[bool]

Task validation option:

  • True — All valid tasks are added. If a task does not pass validation, then it is not added to Toloka. All such tasks are listed in the response.
  • False — If any task does not pass validation, then operation is cancelled and no tasks are added to Toloka.
async_mode Optional[bool]

Request processing mode:

  • True — Asynchronous operation is started internally and create_tasks waits for the completion of it. It is recommended to create no more than 10,000 tasks per request in this mode.
  • False — The request is processed synchronously. A maximum of 5000 tasks can be added in a single request in this mode.

  • Returns:

    An object with created tasks in items and invalid tasks in validation_errors.

  • Return type:

    TaskBatchCreateResult

Examples:

The first example shows how to create regular tasks using a TSV file.

dataset = pandas.read_csv('dataset.tsv', sep='  ')
tasks = [
    toloka.task.Task(input_values={'image': url}, pool_id=existing_pool_id)
    for url in dataset['image'].values[:50]
]
created_result = toloka_client.create_tasks(tasks, allow_defaults=True)
print(len(created_result.items))

The second example shows how to create control tasks.

dataset = pd.read_csv('dateset.tsv', sep=';')
golden_tasks = []
for _, row in dataset.iterrows():
    golden_tasks.append(
            toloka.task.Task(
                input_values={'image': row['image']},
                known_solutions = [toloka.task.BaseTask.KnownSolution(output_values={'animal': row['label']})],
                pool_id = existing_pool_id,
            )
        )
created_result = toloka_client.create_tasks(golden_tasks, allow_defaults=True)
print(len(created_result.items))