Signed cookie manager for communication between multiple trusted services.

Signs, verifies, and manages multiple cookies from trusted environments. Designed for use by services all within the same secure network (AWS VPC etc).

Wraps itsdangerous for the signing and verification (but this could change in the future).

Specifically, this handles:

• Managing multiple different cookies - one for every environment or paired service
• Error correction around sign/verify commands

This package is designed to sign and verify cookies - either ingoing or outgoing. These cookies are not encrypted, so stick to benign data, and only transmit within a trusted environment such as an AWS VPC.

Install and update using pip:

pip install -U Cookie-Manager

Import:

from cookie_manager.cookie_manager import CookieManager

Cookie-Manager is designed to use multiple different signing/verifying keys -- one (or more) per environment. Configure your keys in a dict:

keys = {"key1": "SECRET", "key2": "SECRET2"}

Create an instance (and seed it with your keys):

cookie_manager = CookieManager(keys=keys)

To sign a cookie, start with a dict payload containing your data:

payload = {"key": "value"}

Then sign the payload, making sure to pass a valid key_id as previously configured. The sign method will retrieve your signing key SECRET to sign requests (based on the key_id you pass in). This WILL override any existing key with the name key_id.

signed_cookie = cookie_manager.sign(cookie=payload, key_id="key1")

This will return you a signed cookie (with an additional key_id pair added in):

'{"key": "value", "key_id": "key1"}.XepkCA.CUZtVTCXHbqoalWVCh5xOa4S4WE'

When reading in a signed cookie, verification happens through the cookie payload -> whatever comes in needs to have a key_id in the payload, which is used to lookup the verification key (configured during instantiation). This is added for you by sign:

incoming_signed_cookie = '{"key": "value", "key_id", "key1"}.XepkCA.CUZtVTCXHbqoalWVCh5xOa4S4WE'

Verify this cookie (during which Cookie-Manager will extract key_id from the payload, and lookup the key used to sign the cookie):

payload = cookie_manager.verify(signed_cookie=signed_cookie)

Now, you can access data inside the payload object. The verify function will raise errors if it cannot verify.

This package uses dependency injection to log errors with Python's print. To use your own logger, pass in a logger object which implements critical, error, warning, debug, and info functions. Here's how to patch in the Flask logger, but any object will work providing it meets the Duck Typing rules:

cookie_manager = CookieManager(keys=keys, logger=app.logger)

This will result in logging calls firing to app.logger.<logger-level> with a string passed in.

Like logging, this package uses custom error handling if you need it. By default, all errors will raise as "Exception", but you can pass in a custom object to raise specific errors.

This class will raise Unauthorized, ServiceUnavailable, and BadRequest.

Here's how to pass in a Werkzeug exception object:

from werkzeug import exceptions
cookie_manager = CookieManager(keys=keys, exceptions=exceptions)

The build pipeline require your tests to pass and code to be formatted

Make sure you have Python 3.x installed on your machine (use pyenv).

Install the dependencies with pipenv (making sure to include dev and pre-release packages):

pipenv install --dev --pre

Configure your environment:

pipenv shell && export PYTHONPATH="$PWD"

Run the tests:

pytest

Or with logging:

pytest -s

Or tests with coverage:

pytest --cov=./

Format the code with Black:

black $PWD

Cleanup the (.gitignored) dist folder (if you have one):

rm -rf dist

Notch up the version number in setup.py and build:

python3 setup.py sdist bdist_wheel

Push to PyPi (using the ScholarPack credentials when prompted)

python3 -m twine upload --repository-url https://upload.pypi.org/legacy/ dist/*

• Releases: https://pypi.org/project/cookie-manager/
• Code: https://github.com/ScholarPack/cookie-manager/