Dataset Version Manager

A Python module that manages versioned datasets with commit history tracking, rollback capabilities, and history visualization.

Capabilities

Commit dataset changes

• Creating a dataset, adding data, and committing with message "Initial data" should record a commit in history @test
• Modifying a dataset and committing with message "Updated data" should create a second commit in history @test
• Committing without any changes since the last commit should still create a new commit entry @test

Access commit history

• After creating three commits, the history should contain exactly three entries @test
• Each commit in history should have a timestamp and commit message @test

Rollback to previous version

• After committing twice, rolling back should revert the dataset to the first commit's state @test
• Rolling back multiple versions should restore the dataset to the target version's state @test

Implementation

@generates

API

def create_versioned_dataset(path: str) -> object:
    """
    Creates a new versioned dataset at the specified path.

    Args:
        path: Storage path for the dataset

    Returns:
        Dataset object with version control capabilities
    """
    pass

def commit_changes(dataset: object, message: str) -> None:
    """
    Commits current changes to the dataset with a descriptive message.

    Args:
        dataset: The dataset to commit
        message: Commit message describing the changes
    """
    pass

def get_history(dataset: object) -> list:
    """
    Retrieves the commit history for the dataset.

    Args:
        dataset: The dataset to get history from

    Returns:
        List of commit records with metadata
    """
    pass

def rollback_to_version(dataset: object, version_index: int) -> None:
    """
    Reverts the dataset to a previous version.

    Args:
        dataset: The dataset to rollback
        version_index: Index of the version to rollback to (0-based from oldest)
    """
    pass

Dependencies { .dependencies }

deeplake { .dependency }

Provides dataset management with built-in version control capabilities.