rehla-dbx-tools 1.0.0

Production-focused Databricks API toolkit by Rehla Digital Inc for workspace and account automation on AWS.

Unified Databricks API

Single Python package to call Databricks Workspace and Account APIs and convert JSON responses into Pandas or PySpark DataFrames.

About Rehla Digital Inc

Rehla Digital Inc builds cloud and data engineering solutions that help teams standardize platform operations, accelerate delivery, and reduce integration risk. This package is maintained as part of that effort to provide a practical, production-oriented Databricks API toolkit.

Install

pip install rehla-dbx-tools

Import in Python with underscores:

from rehla_dbx_tools import DatabricksApiClient

Install Spark extras if needed:

pip install "rehla-dbx-tools[spark]"

Quick Start

from rehla_dbx_tools import DatabricksApiClient

client = DatabricksApiClient.from_env()
if client.workspace is not None:
    jobs = client.workspace.list_jobs()
    df = jobs.to_pandas()
    print(df.head())

# Force both workspace/account config to a target cloud
client = DatabricksApiClient.from_env_for_cloud("azure")

Notebook Context Bootstrap

Inside Databricks notebooks:

from rehla_dbx_tools import DatabricksApiClient

client = DatabricksApiClient.from_notebook_context()
if client.workspace is not None:
    clusters = client.workspace.list_clusters()
    spark_df = clusters.to_spark()
    display(spark_df)

Account API

account client is enabled when DATABRICKS_ACCOUNT_HOST and DATABRICKS_ACCOUNT_ID are set.

if client.account is not None:
    workspaces = client.account.list_workspaces()
    print(workspaces.to_pandas().head())

Version-Aware Generic Request

response = client.workspace.request_versioned(
    "GET",
    service="unity-catalog",
    endpoint="metastores",
    api_version="2.1",
)

Expanded Convenience Wrappers

import getpass

if client.workspace is not None:
    run = client.workspace.run_job_now(job_id=123)
    runs = client.workspace.list_job_runs(job_id=123, active_only=True, limit=10)
    run_export = client.workspace.export_job_run(run_id=987, views_to_export="CODE")
    run_output = client.workspace.get_job_run_output(run_id=987)
    run_submit = client.workspace.submit_job_run({"run_name": "ad-hoc-check"})
    run_delete = client.workspace.delete_job_run(run_id=987)
    repair = client.workspace.repair_job_run(run_id=987, rerun_all_failed_tasks=True)
    cancel_all = client.workspace.cancel_all_job_runs(job_id=123, all_queued_runs=True)
    cluster = client.workspace.get_cluster(cluster_id="0123-abc")
    catalogs = client.workspace.list_catalogs(max_results=25)
    warehouses = client.workspace.list_sql_warehouses()
    dbfs_files = client.workspace.list_dbfs("dbfs:/tmp")
    token = client.workspace.create_token(lifetime_seconds=3600, comment="ci-short-lived")
    rotated_token = client.workspace.rotate_token(
        token_id_to_revoke="old-token-id",
        lifetime_seconds=3600,
        comment="ci-rotation",
    )
    repos = client.workspace.list_repos(path_prefix="/Repos/team")
    repo = client.workspace.get_repo(repo_id=12345)
    client.workspace.put_secret(
        scope="app-prod",
        key="api-token",
        string_value=getpass.getpass("Secret value: "),
    )

if client.account is not None:
    ws = client.account.get_workspace(workspace_id=101)
    creds = client.account.list_credentials()
    storage_cfgs = client.account.list_storage_configurations()
    networks = client.account.list_networks()
    private_access = client.account.list_private_access_settings()
    vpc_endpoints = client.account.list_vpc_endpoints()
    cmks = client.account.list_customer_managed_keys()
    users = client.account.list_users()
    user = client.account.get_user("user-101")
    groups = client.account.list_groups()
    group = client.account.get_group("group-101")
    budgets = client.account.list_budget_policies()
    log_delivery_configs = client.account.list_log_delivery_configurations()

For detailed setup and examples, see docs/USAGE.md.