Skip to main content
Join the official 2019 Python Developers SurveyStart the survey!

Looker API 3.1

Project description

The Looker SDK for Python provides a convenient way to communicate with the Looker API available on your Looker server. The library requires python3.7+ and is annotated using the typing module.

DISCLAIMER: This is a beta version of the Looker SDK, using a completely new code generator developed by Looker. Implementations are still subject to change, but we expect most SDK method calls to work correctly. If you run into problems with the SDK, please feel free to report an issue, and please indicate which language SDK you’re using in the report.

Sample project setup

Install python 3.7. We highly recommend using pyenv to install different versions of python. Mac users should use homebrew to install pyenv:

brew install pyenv

Follow the remaining steps 3 - 5 of otherwise your python3.7 installation may break.

Now you’re ready to install python 3.7:

pyenv install 3.7.4

Next we’ll use pipenv as an awesome enhancement to pip to manage project dependencies.

brew install pipenv

Create a project directory

mkdir looker-sdk-example

Install python3.7 and use it for this directory

cd looker-sdk-example/
pyenv local 3.7.4

Install looker_sdk using pipenv

pipenv install --python 3.7 --pre looker_sdk

Configuring the SDK

The SDK supports configuration through a .ini file on disk as well as setting environment variables (the latter override the former).

Note: The .ini configuration for the Looker SDK is a sample implementation intended to speed up the initial development of python applications using the Looker API. See this note on Securing your SDK Credentials for warnings about using .ini files that contain your API credentials in a source code repository or production environment.

In order to configure the SDK client, create a “looker.ini” file to reference during client.setup()

example file:

# API version is required
# Base URL for API. Do not include /api/* in the url
# API 3 client id
# API 3 client secret
# Set to false if testing locally against self-signed certs. Otherwise leave True

Note: If the application using the Looker SDK is going to be committed to a version control system, be sure to ignore the looker.ini file so the API credentials aren’t unintentionally published.

For any .ini setting you can use an environment variable instead. It takes the form of LOOKERSDK_<UPPERCASE-SETTING-FROM-INI> e.g. LOOKERSDK_CLIENT_SECRET

Code example

Copy the following code block into Note: it’s helpful to launch your code editor with your virtual environment loaded so that it can find the looker_sdk library and give you a nice code completion experience. Run pipenv shell to start load the virtual environment and then run your editor command (e.g. for VSCode - code

from looker_sdk import client, models, error

# client calls will now automatically authenticate using the
# api3credentials specified in 'looker.ini'
sdk = client.setup("looker.ini")
looker_api_user =

# models can be passed named parameters to the constructor
new_user = models.WriteUser(first_name="John", last_name="Doe")

# as well as have fields set on the instance
new_user.is_disabled = True
new_user.locale = "fr"

# create the user with the client
created_user = sdk.create_user(new_user)
    f"Created user({}): "
    f"{created_user.display_name} "

# Updating the user: change first_name and explicitly nullify
# locale so that it defaults to looker system locale
update_user = models.WriteUser(
    first_name="Jane", locale=models.EXPLICIT_NULL  # do not use None

# update the user with the client
user_id =
updated_user = sdk.update_user(user_id, body=update_user)
    f"Updated user({user_id}): {updated_user.display_name} "

# perform API calls on behalf of the user: "sudo"
    print(f"Sudo as {user_id}")
except error.SDKError:
    print(f"Oops, we need to enable user({user_id}) first")
    sdk.update_user(user_id, body=models.WriteUser(is_disabled=False))

sudo_user =
assert == user_id
assert !=

# logout to switch back to authenticating per 'looker.ini'
print(f"Ending sudo({user_id}) session")
assert ==

# "sudo" using a context manager
with sdk.login_user(user_id):
    assert == user_id

# exiting context manager is the same as
# calling sdk.logout()
assert ==

# cleanup
print(f"Removed user({user_id})")

You can run the example code above but be aware it will actually create and delete a user in your looker instance.

pipenv run python

If you see a lot of InsecureRequestWarning errors because you’re running against an instance with a self-signed cert, this will clean up the output:

PYTHONWARNINGS=ignore pipenv run python

A note on static type checking

All client calls are annotated with with basic types as well as model types. Many client calls accept a fields argument which limits the JSON response from the API to the specified fields. For this reason, the all properties on the model are all typed as Optional[]. The effect is that static code analysis (mypy for example) will complain if you try to use a field from a model instance in a place that requires the value not be Optional. From the example above

created_user = sdk.create_user(new_user)
user_id =

# mypy error: Argument "user_id" to "update_user" of "LookerSDK"
# has incompatible type "Optional[int]"; expected "int"
sdk.update_user(user_id, ...)

This is because has type Optional[int] but we need to use it in the update_user() call which is annotated like this:

def update_user(
    user_id: int,  # note: not Optional[int]
    body: models.WriteUser,
    fields: Optional[str] = None,
) -> models.User:

We know that is an int (we didn’t pass in a fields argument to create_user() excluding id from the response). However, mypy does not so we must guide it in one of the following ways

# assert about the type
assert isinstance(user_id, int)

# or cast
from typing import cast
user_id = cast(, int)

Now mypy is happy with update_user(user_id, ...)

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for looker-sdk, version 0.1.3b4
Filename, size File type Python version Upload date Hashes
Filename, size looker_sdk-0.1.3b4-py3-none-any.whl (100.0 kB) File type Wheel Python version py3 Upload date Hashes View hashes
Filename, size looker_sdk-0.1.3b4.tar.gz (89.5 kB) File type Source Python version None Upload date Hashes View hashes

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page