apache_beam.ml.gcp.recommendations_ai module

A connector for sending API requests to the GCP Recommendations AI API (https://cloud.google.com/recommendations).

class apache_beam.ml.gcp.recommendations_ai.CreateCatalogItem(project: str | None = None, retry: Retry | None = None, timeout: float = 120, metadata: Sequence[Tuple[str, str]] = (), catalog_name: str = 'default_catalog')[source]

Bases: PTransform

Creates catalog item records.

The PTransform returns a PCollectionTuple of successfully created catalog items (created_catalog_items) and failures (failed_catalog_items).

Example usage:

result = (
    pipeline
    | CreateCatalogItem(
        project='example-gcp-project', catalog_name='my-catalog'))
created = result.created_catalog_items

Initializes a CreateCatalogItem transform.

Parameters:

project (str) – Optional. GCP project name in which the catalog data will be imported.
retry – Optional. Designation of what errors, if any, should be retried.
timeout (float) – Optional. The amount of time, in seconds, to wait for the request to complete.
metadata – Optional. Strings which should be sent along with the request as metadata.
catalog_name (str) – Optional. Name of the catalog. Default: ‘default_catalog’

expand(pcoll)[source]

class apache_beam.ml.gcp.recommendations_ai.WriteUserEvent(project: str | None = None, retry: Retry | None = None, timeout: float = 120, metadata: Sequence[Tuple[str, str]] = (), catalog_name: str = 'default_catalog', event_store: str = 'default_event_store')[source]

Bases: PTransform

Write user event information. The PTransform returns a PCollectionTuple with PCollections of successfully and failed written UserEvents.

Example usage:

pipeline
| WriteUserEvent(
    project='example-gcp-project',
    catalog_name='my-catalog',
    event_store='my_event_store')

Initializes a WriteUserEvent transform.

Parameters:

project (str) – Optional. GCP project name in which the catalog data will be imported.
retry – Optional. Designation of what errors, if any, should be retried.
timeout (float) – Optional. The amount of time, in seconds, to wait for the request to complete.
metadata – Optional. Strings which should be sent along with the request as metadata.
catalog_name (str) – Optional. Name of the catalog. Default: ‘default_catalog’
event_store (str) – Optional. Name of the event store. Default: ‘default_event_store’

expand(pcoll)[source]

class apache_beam.ml.gcp.recommendations_ai.ImportCatalogItems(max_batch_size: int = 5000, project: str | None = None, retry: Retry | None = None, timeout: float = 120, metadata: Sequence[Tuple[str, str]] = (), catalog_name: str = 'default_catalog')[source]

Bases: PTransform

Imports catalogitems in bulk. The PTransform returns a PCollectionTuple with PCollections of successfully and failed imported CatalogItems.

Example usage:

pipeline
| ImportCatalogItems(
    project='example-gcp-project',
    catalog_name='my-catalog')

Initializes a ImportCatalogItems transform

Parameters:

batch_size (int) – Required. Maximum number of catalogitems per request.
project (str) – Optional. GCP project name in which the catalog data will be imported.
retry – Optional. Designation of what errors, if any, should be retried.
timeout (float) – Optional. The amount of time, in seconds, to wait for the request to complete.
metadata – Optional. Strings which should be sent along with the request as metadata.
catalog_name (str) – Optional. Name of the catalog. Default: ‘default_catalog’

expand(pcoll)[source]

class apache_beam.ml.gcp.recommendations_ai.ImportUserEvents(max_batch_size: int = 5000, project: str | None = None, retry: Retry | None = None, timeout: float = 120, metadata: Sequence[Tuple[str, str]] = (), catalog_name: str = 'default_catalog', event_store: str = 'default_event_store')[source]

Bases: PTransform

Imports userevents in bulk. The PTransform returns a PCollectionTuple with PCollections of successfully and failed imported UserEvents.

Example usage:

pipeline
| ImportUserEvents(
    project='example-gcp-project',
    catalog_name='my-catalog',
    event_store='my_event_store')

Initializes a WriteUserEvent transform.

Parameters:

batch_size (int) – Required. Maximum number of catalogitems per request.
project (str) – Optional. GCP project name in which the catalog data will be imported.
retry – Optional. Designation of what errors, if any, should be retried.
timeout (float) – Optional. The amount of time, in seconds, to wait for the request to complete.
metadata – Optional. Strings which should be sent along with the request as metadata.
catalog_name (str) – Optional. Name of the catalog. Default: ‘default_catalog’
event_store (str) – Optional. Name of the event store. Default: ‘default_event_store’

expand(pcoll)[source]

class apache_beam.ml.gcp.recommendations_ai.PredictUserEvent(project: str | None = None, retry: Retry | None = None, timeout: float = 120, metadata: Sequence[Tuple[str, str]] = (), catalog_name: str = 'default_catalog', event_store: str = 'default_event_store', placement_id: str | None = None)[source]

Bases: PTransform

Make a recommendation prediction. The PTransform returns a PCollection

Example usage:

pipeline
| PredictUserEvent(
    project='example-gcp-project',
    catalog_name='my-catalog',
    event_store='my_event_store',
    placement_id='recently_viewed_default')

Initializes a PredictUserEvent transform.

Parameters:

project (str) – Optional. GCP project name in which the catalog data will be imported.
retry – Optional. Designation of what errors, if any, should be retried.
timeout (float) – Optional. The amount of time, in seconds, to wait for the request to complete.
metadata – Optional. Strings which should be sent along with the request as metadata.
catalog_name (str) – Optional. Name of the catalog. Default: ‘default_catalog’
event_store (str) – Optional. Name of the event store. Default: ‘default_event_store’
placement_id (str) – Required. ID of the recommendation engine placement. This id is used to identify the set of models that will be used to make the prediction.

expand(pcoll)[source]