Missed TensorFlow Dev Summit? Check out the video playlist. Watch recordings

tf.contrib.cloud.BigtableClient

View source on GitHub

Class BigtableClient

BigtableClient is the entrypoint for interacting with Cloud Bigtable in TF.

BigtableClient encapsulates a connection to Cloud Bigtable, and exposes the table method to open a Bigtable table.

__init__

View source

__init__(
    project_id,
    instance_id,
    connection_pool_size=None,
    max_receive_message_size=None
)

Creates a BigtableClient that can be used to open connections to tables.

Args:

  • project_id: A string representing the GCP project id to connect to.
  • instance_id: A string representing the Bigtable instance to connect to.
  • connection_pool_size: (Optional.) A number representing the number of concurrent connections to the Cloud Bigtable service to make.
  • max_receive_message_size: (Optional.) The maximum bytes received in a single gRPC response.

Raises:

  • ValueError: if the arguments are invalid (e.g. wrong type, or out of expected ranges (e.g. negative).)

Methods

table

View source

table(
    name,
    snapshot=None
)

Opens a table and returns a tf.contrib.bigtable.BigtableTable object.

Args:

  • name: A tf.string tf.Tensor name of the table to open.
  • snapshot: Either a tf.string tf.Tensor snapshot id, or True to request the creation of a snapshot. (Note: currently unimplemented.)

Returns:

A tf.contrib.bigtable.BigtableTable Python object representing the operations available on the table.