Branch

Each dataset is created with a default branch main. When getting a dataset through the SDK, its version is the latest commit of the default branch.

The following sections will introduce the operations and precautions related to branches in the SDK. Of course, it is necessary to get a dataset first:

from graviti import Workspace

ws = Workspace(f"{YOUR_ACCESSKEY}")
dataset = ws.datasets.get(f"{DATASET_NAME}")

Note

In most cases, a branch can be thought of as a named commit. But if there is a dataset that has just been created and has no commit history, the commit id of its default branch is None.

Create a Branch

SDK provides method create() to support creating a branch based on a revision. The revision can be one commit ID:

dataset.branches.create(f"{BRANCH_NAME}", f"{COMMIT_ID}")

The revision can also be the branch name. In this situation, the new branch will be created based on the latest commit of the source branch:

dataset.branches.create(f"{BRANCH_NAME}", f"{SOURCE_BRANCH_NAME}")

The revision can also be the tag name:

dataset.branches.create(f"{BRANCH_NAME}", f"{TAG_NAME}")

If no revision is specified, the created branch will be based on the current commit of the dataset:

dataset.branches.create(f"{BRANCH_NAME}")

List Branches

SDK provides method list() to support listing branches:

dataset.branches.list()

Get a Branch

SDK provides method get() to support getting a branch by name:

dataset.branches.get(f"{BRANCH_NAME}")

Delete a Branch

SDK provides method delete() to support deleting a branch by name:

dataset.branches.delete(f"{BRANCH_NAME}")

Checkout Branch

SDK provides method checkout() to support switching the version of the dataset by branches:

dataset.checkout(f"{BRANCH_NAME}")
dataset.HEAD  # Check whether the dataset version is correct.

In addition, this checkout method is often used to update the version of the local dataset without getting the dataset again, for example:

# Other users committed a draft on the Branch("main").
dataset.checkout("main")  # Update the version of the dataset.