h2o-drive

Python Client for H2O Drive.

Installation

pip install h2o-drive

Usage

Connecting to H2O Drive

import h2o_drive

drive = h2o_drive.connect()

When used within the H2O AI Cloud environment or locally with the H2O AI Cloud CLI configured, no further configuration is needed.

h2o_drive.connect() can be configured via optional parameters:

• discovery: The h2o_discovery.Discovery object through which to discover services for determining connection details.
• userdrive_url: URL to Drive's userdrive service.
• workspacedrive_url: URL to Drive's workspacedrive service.
• sts_url: URL to the STS service.
• token_provider: A token provider generating authentication credentials.

Object Operations

To perform object operations, first open any Drive bucket. For example:

bucket = drive.user_bucket()
# or
bucket = drive.workspace_bucket("default")

Buckets have the following methods:

• upload_file() uploads a local file, resulting in a new object at the specified key in the bucket. It takes two arguments:

  • filename: The file to upload. The contents of this file will become an object in the Drive bucket.
  • key: The key at which to store the resulting object. In other words, the name attached to the object.

• list_objects() returns the list of objects in the bucket. It takes one optional argument:

  • prefix: When specified, only objects at keys starting with the specified value are listed.

• download_file() downloads the object at the specified key and writes it to the specified local file. It takes two arguments:

  • key: The key of the object to download.
  • filename: The file, on the local filesystem, the object is written to.

• delete_object() deletes the object at the specified key. It takes a single argument:

  • key: The key of the object to delete.

• generate_presigned_url() generates a presigned URL for accessing the object at the specified key in the bucket. It takes two arguments:

  • key: The key to generate a presigned URL for.
  • duration_seconds: (Optional) The duration, in seconds, for which the URL should be valid. Defaults to 1 day.

• with_prefix() returns a prefixed view of this bucket. All the keys in each of its object operations are automatically prefixed with the specified value. It takes a single argument:

  • prefix: The key prefix to automatically apply to all object operations.

Examples

Example: Open the storage bucket for your personal H2O workspace, write a local file, and list its contents

import h2o_drive

# Prepare a local test file.
with open("test-file.txt", "w") as f:
    f.write("Hello, world!")

drive = h2o_drive.connect()
bucket = drive.workspace_bucket("default")

await bucket.upload_file("test-file.txt", "drive-test/uploaded-test-object.txt")

objects = await bucket.list_objects()
print(objects)