python3-core-api-client 1.0

Python client for Core API.

python3-core-api-client

Python client for Core API.

This client was built for and tested on the 1.241 version of the API.

Support

This client is officially supported by Cyberfusion.

Have questions? Ask your support questions on the platform. No access to the platform? Send an email to [e-mail]. GitHub issues are not actively monitored.

Install

This client can be used in any Python project and with any framework.

This client requires Python 3.11 or higher.

PyPI

Run the following command to install the package from PyPI:

pip3 install python3-core-api-client

Debian

Run the following commands to build a Debian package:

mk-build-deps -i -t 'apt -o Debug::pkgProblemResolver=yes --no-install-recommends -y'
dpkg-buildpackage -us -uc

Usage

API documentation

Refer to the API documentation for information about API requests.

Enums and Models are auto-generated based on the OpenAPI spec - so the client is completely in line with the Core API. Requests and Resources are not auto-generated.

Getting started

Initialise the CoreApiConnector with your username and password or API key.

The connector takes care of authentication, and offers several resources (e.g. virtual_hosts) and endpoints (e.g. list_virtual_hosts).

from cyberfusion.CoreApiClient.connector import CoreApiConnector

# Using username and password

connector = CoreApiConnector(
    username='username', password='password'
)

# Or using API key

connector = CoreApiConnector(
    api_key='api_key'
)

virtual_hosts = connector.virtual_hosts.list_virtual_hosts()

Authentication

This client takes care of authentication.

If authentication using username and password fails, cyberfusion.CoreApiClient.exceptions.AuthenticationException is thrown.

If authentication using API key fails, the regular CallException exception is raised.

Don't have an API user? Contact Cyberfusion.

Requests

The client uses a fluent interface to build requests.

• Start with the connector
• Go to the desired resource
• Call the desired endpoint

Code example:

from cyberfusion.CoreApiClient.connector import CoreApiConnector

from cyberfusion.CoreApiClient.models import MailDomainCreateRequest

connector = CoreApiConnector(
    username='username', password='password'
)

connector.mail_domains.create_mail_domain(
    MailDomainCreateRequest(
        domain='cyberfusion.io',
        unix_user_id=1,
        is_local=True,
        catch_all_forward_email_addresses=[],
    )
)

Models are validated before sending the request (using Pydantic). If invalid data is provided, pydantic.ValidationError is thrown.

Code example:

from cyberfusion.CoreApiClient.connector import CoreApiConnector

from cyberfusion.CoreApiClient.models import MailAliasCreateRequest

connector = CoreApiConnector(
    username='username', password='password'
)

connector.mail_aliases.create_mail_alias(
    MailAliasCreateRequest(
        local_part='&^@$#^&@$#^&',
        mail_domain_id=1,
    )
)
# throw pydantic.ValidationError

The exception has an errors() method to get all validation errors.

Code example:

from cyberfusion.CoreApiClient.connector import CoreApiConnector

from cyberfusion.CoreApiClient.models import MailAliasCreateRequest

import pydantic

connector = CoreApiConnector(
    username='username', password='password'
)

try:
    connector.mail_aliases.create_mail_alias(
        MailAliasCreateRequest(
            local_part='&^@$#^&@$#^&',
            mail_domain_id=1,
        )
    )
except pydantic.ValidationError as e:
    errors = e.errors()

for error in errors:
    print(error['loc'])
    print(error['msg'])
    print(error['type'])

Responses

Get model from response

Calling an endpoint returns the resource model.

Code example:

from cyberfusion.CoreApiClient.connector import CoreApiConnector

from cyberfusion.CoreApiClient.models import MailDomainCreateRequest

connector = CoreApiConnector(
    username='username', password='password'
)

mail_domain_resource = connector.mail_domains.create_mail_domain(
    MailDomainCreateRequest(
        domain='cyberfusion.io',
        unix_user_id=1,
        is_local=True,
        catch_all_forward_email_addresses=[],
    )
)
# mail_domain_resource is a model representing the API resource

Throw exception on failure

If a request returns an unexpected HTTP status code, cyberfusion.CoreApiClient.exceptions.CallException is thrown.

The exception includes the response, and the HTTP status code.

Enums

Some properties only accept certain values (enums).

Find these values in cyberfusion.CoreApiClient.models.

Deep dive

Custom requests session

Want to provide your own requests session? Pass it to the connector:

import requests

from cyberfusion.CoreApiClient.connector import CoreApiConnector

connector = CoreApiConnector(
    ..., requests_session=requests.Session()
)

Don't pass a custom session? A default one is created, with retries enabled.

Manual requests

Don't want to use the full SDK, but easily send requests and retrieve responses from the Core API?

Initialise the connector as usual, and call send:

import requests

from cyberfusion.CoreApiClient.connector import CoreApiConnector

connector = CoreApiConnector(...)

response = connector.send(method='GET', path='/foobar', data={}, query_parameters={})

response.status_code
response.json
response.body
response.headers
response.failed

To raise cyberfusion.CoreApiClient.exceptions.CallException in case of an unexpected HTTP status code, use send_or_fail instead of send. It takes the same parameters.

Generating models

Auto-generate models as follows:

datamodel-codegen --input-file-type openapi --input $file --output src/cyberfusion/CoreApiClient/models.py --target-python-version 3.11

Replace $file by the path to the OpenAPI spec (JSON).

Note: don't replace models.py in full - it contains customisations.

Test strategy

Tests use a mock server, Stoplight Prism.

Prism checks requests' syntactic validity - based on the OpenAPI spec.

Therefore, the resources' test suites solely call methods without asserting specifics: nearly all possible issues - invalid requests, mismatch between resource models and endpoint, etc. - are already caught by having a validating mock server.