fish-databricks-jobs

fish-databricks-jobs is cli and python sdk to manage Jobs for Databricks. e.g assign permissions to multiple jobs. User can filter jobs by job name or tags.

The functionality of current databricks-cli(v0.17.4) is limited on the jobs api. e.g it can not assign job permission.

With fish-databricks-jobs, you can assign job permission, e.g:

to assign group mygroup with permission can_manage to job by filter 843966944901662(job_id)

$ fish-databricks-jobs permission-assign mygroup --type group --level can_manage --filter 843966944901662

Installation

Python Version >= 3.7

$ pip install --upgrade fish-databricks-jobs

$ fish-databricks-jobs --version
Version: 0.6.8

Usage

permission-assign

$ fish-databricks-jobs permission-assign -h

 Usage: fish-databricks-jobs permission-assign [OPTIONS] NAME

 Assign permission to user

╭─ Arguments ─────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
│ *    name      TEXT  User name, group name or serive principal id. Who will receive the permisssion. [default: None]    │
│                      [required]                                                                                         │
╰─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯
╭─ Options ───────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
│    --type     -t      [user|principal|group]                Permission receiver type. [default: user]                   │
│ *  --level    -l      [can_view|can_manage|can_manage_run]  Permission level. [default: None] [required]                │
│    --filter   -f      TEXT                                  filter jobs, case insensitively. [default: None]            │
│    --profile  -p      TEXT                                  profile name in ~/.databrickscfg [default: DEFAULT]         │
│    --force                                                  Attempt to assign permission without prompting for          │
│                                                             confirmation. **Use this flag with caution**                │
│    --help     -h                                            Show this message and exit.                                 │
╰─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯

use as sdk to list jobs

in databricks' notebook

%pip install fish-databricks-jobs

from fish_databricks_jobs.services.jobs import JobsService, Job
host = f'https://{spark.conf.get("spark.databricks.workspaceUrl")}'
token = 'exampletokenc0e27d8b91fd8c0144f0a23'

job_list = JobsService(host, token).list()
df = spark.createDataFrame(job_list)

display(df)

Config authentication

fish-databricks-jobs uses same config file as databricks-cli. e.g.~/.databrickscfg for macOS. Similar for Windows.

[DEFAULT]
host = https://example.cloud.databricks.com
token = exampletokenc0e27d8b91fd8c0144f0a23

Developer

Setup

1. Checkout code
2. install poetry, e.g: in macOS or Windows-cmder

curl -sSL https://install.python-poetry.org | python -

2. config poetry

poetry config virtualenvs.in-project true

3. install dependencies python libraries in the venv

poetry install
poetry shell 

4. in pycharm, Add New Interpreter -> Poetry Environment ...
5. Verifiy

(fish-databricks-jobs-py3.8) (base) username@macbook:~/projects/fish-databricks-jobs [main] 
$ python ./fish_databricks_jobs/cli.py --version
Version: 0.7.6