Databricks Labs Blueprint

0.12.0 · active · verified Sun Mar 29

Databricks Labs Blueprint is a Python library that provides common building blocks and utilities for Databricks Labs projects. It offers Python-native pathlib-like interfaces for Databricks Workspace paths, tools for trivial terminal user interfaces (TUI), and utilities for managing application and installation state. The current version is 0.12.0, with a regular release cadence, typically monthly or bi-monthly, reflecting active development and maintenance. [1, 3]

Warnings

breaking In `v0.11.3`, the unmarshalling of JSON floating-point values was fixed. Previously, JSON floats might have been silently truncated to integers. The updated functionality now raises a `SerdeError` when precision would be lost during conversion from float to integer. [release notes]
Fix: Review code that reads JSON configuration files, especially if it expects integer values from floating-point inputs. Handle `SerdeError` for explicit type conversions.
deprecated Starting with `v0.11.0`, using `Any` and `object` as type annotations on data classes for marshalling is deprecated and will issue a `DeprecationWarning`. [release notes]
Fix: Refactor data classes to use more specific type annotations instead of `Any` or `object` to avoid deprecation warnings and ensure type safety.
gotcha In versions prior to `v0.9.3`, there was an issue where `databricks-sdk` config objects could be unintentionally overridden when creating installation config files. This could lead to unexpected behavior or incorrect workspace configurations. [release notes]
Fix: Upgrade to `v0.9.3` or newer. If upgrading is not immediately possible, carefully review any custom logic that modifies or saves Databricks SDK configuration objects to ensure they are handled correctly.

Install

pip install databricks-labs-blueprint Install latest version

Imports

WorkspacePath
```
from databricks.labs.blueprint.paths import WorkspacePath
```
Provides pathlib-like interface for Databricks Workspace paths.
Prompts
```
from databricks.labs.blueprint.tui import Prompts
```
Offers utilities for interactive terminal user interfaces (TUI).
ProductInfo
```
from databricks.labs.blueprint.wheels import ProductInfo
```
Used for managing wheel installations and dependencies.

Quickstart

This quickstart demonstrates how to initialize a `WorkspaceClient` and use `WorkspacePath` to create and manage directories within your Databricks workspace. It expands a relative path to the user's home directory, creates the specified folder structure, verifies its existence, and then cleans it up. Ensure your Databricks SDK authentication (environment variables or CLI profile) is configured before running. [1]

import os
from databricks.sdk import WorkspaceClient
from databricks.labs.blueprint.paths import WorkspacePath

# Ensure DATABRICKS_HOST and DATABRICKS_TOKEN environment variables are set,
# or a Databricks CLI profile is configured.
# For local testing, you might run:
# export DATABRICKS_HOST='https://<your-databricks-instance>.cloud.databricks.com'
# export DATABRICKS_TOKEN='dapi...'

if not os.environ.get('DATABRICKS_HOST') or not os.environ.get('DATABRICKS_TOKEN'):
    print("Please set DATABRICKS_HOST and DATABRICKS_TOKEN environment variables or configure Databricks CLI.")
    # In a real quickstart, you might exit or raise an error here.
    # For demonstration, we'll use placeholder values that will likely fail.
    ws = WorkspaceClient(host=os.environ.get('DATABRICKS_HOST', 'https://example.cloud.databricks.com'), 
                         token=os.environ.get('DATABRICKS_TOKEN', 'dapi-fake-token'))
else:
    ws = WorkspaceClient()

print(f"Initialized WorkspaceClient for host: {ws.host}")

try:
    user_name = ws.current_user.me().user_name
    print(f"Current user: {user_name}")

    # Example: Working with user home folders
    folder_name = 'blueprint-test-folder'
    wsp = WorkspacePath(ws, f"~/{{folder_name}}/sub/dir")

    # Expand the user path and create the directory
    with_user = wsp.expanduser()
    print(f"Expanded path: {with_user}")
    with_user.mkdir()
    print(f"Directory '{with_user}' created.")

    # Verify existence
    assert with_user.is_dir()
    print(f"Directory '{with_user}' exists.")

    # Clean up (recursive rmdir)
    with_user.parent.parent.rmdir(recursive=True)
    assert not with_user.parent.parent.exists()
    print(f"Directory '{with_user.parent.parent}' and its contents removed.")

except Exception as e:
    print(f"An error occurred: {e}")
    print("Please ensure your Databricks environment is correctly configured (host, token, permissions).")

view raw JSON →