Identity Lib

This Python package contains shared code related to Identity systems within UIS. It's primary purpose is to encourage code-reuse and to allow for client systems to make use of the same data structures and logic that is contained within our emergent identity APIs.

Use

Install ucam-identitylib using pip:

pip install ucam-identitylib

The module can then be used as identitylib:

from identitylib.identifiers import Identifier

identifier = Identifier.from_string('wgd23@v1.person.identifiers.cam.ac.uk')
print(identifier)

Developer quickstart

This project contains a dockerized testing environment which wraps tox.

Tests can be run using the ./test.sh command:

# Run all PyTest tests and Flake8 checks
$ ./test.sh

# Run PyTest and Flake8 and recreate test environments
$ ./test.sh --recreate

# Run just PyTest
$ ./test.sh -e py3

# Run a single test file within PyTest
$ ./test.sh -e py3 -- tests/test_identifiers.py

# Run a single test file within PyTest with verbose logging
$ ./test.sh -e py3 -- tests/test_identifiers.py -vvv

Pulling latest specs from source repositories

Local copies of the OpenAPI specs used to generate the library should be pulled in to this repo so the specific specs used in each build are under revision control. This can be done using the provided script:

$ ./pull-specs.sh

# If an access token required for https clones from gitlab repositories
# then this can be specified using:
$ ./pull-specs.sh --token "ACCESS_TOKEN_HERE"

# You may need to first set the $USER environment variable to match the GitLab account name.

Generating the identitylib

The identitylib is generated during the docker build process. To create a local copy of the identitylib distribution use the build script:

$ ./build-local.sh

This will create a new folder /dist in the current directory with the wheel and tar package for identitylib.

Interactive testing

An interactive testing environment is provided to smoke-test any changes to the identitylib against staging or local instances of the APIs. Note that to test local instance you will need to set the URL in the config file to host.docker.internal:<PORT NUMBER> on linux.

To set up the interactive testing environment, copy the scripts/api-loader-config.yaml.example file to scripts/api-loader-config.yaml and populate with valid API key/secret pairs for any of the APIs you wish to test.

Run the interactive test environment using the provided script:

$ ./interactive-tests.sh

Additional arguments can be passed to the docker build and docker run commands by setting the variables ITEST_BUILD_ARGS and ITEST_RUN_ARGS, e.g.:

$ ITEST_BUILD_ARGS="--build-arg install_from=dist/ucam_identitylib-1.7.0-py3-none-any.whl" ./interactive-tests.sh

will build the interactive environment from a local wheel stored in the dist directory. By default the latest version avaiable in the test.pypi.org/simple index is installed into the environment.

Once the environment is active you can load up test instances and manipulate them using the python interactive console. The LOADER global variable contains a helper class which reads the config file and generates clients based off it:

# Docker build output...
>>> config, client, inst = LOADER.load_card()
>>> inst.v1_cards_list()
... # Output printed to the screen

Use the builtin help command to view info on the LOADER, and the returned api_inst objects that can be used to make requests.