Skip to content

Latest commit

 

History

History
219 lines (166 loc) · 7.85 KB

README.rst

File metadata and controls

219 lines (166 loc) · 7.85 KB

PyBankID

https://dev.azure.com/hbldh/github/_apis/build/status/hbldh.pybankid?branchName=master Documentation Status https://coveralls.io/repos/github/hbldh/pybankid/badge.svg?branch=master

PyBankID is a client for providing BankID services as a Relying Party, i.e. providing authentication and signing functionality to end users. This package provides a simplifying interface for initiating authentication and signing orders and then collecting the results from the BankID servers.

If you intend to use PyBankID in your project, you are advised to read the BankID Relying Party Guidelines before doing anything else. There, one can find information about how the BankID methods are defined and how to use them.

An example web application using PyBankID exists and can be found in deployed state on Heroku.

If you use PyBankID in production and want updates on new releases and notifications about important changes to the BankID service, send a mail to the developer of this package to be added to the PyBankID mailing list.

Installation

PyBankID can be installed though pip:

pip install pybankid

To remedy the InsecurePlatformWarning problem detailed below (Python 2, urllib3 and certificate verification), you can install pybankid with the security extras:

pip install pybankid[security]

This installs the pyopenssl, ndg-httpsclient and pyasn1 packages as well.

In Linux, this does however require the installation of some additional system packages:

sudo apt-get install build-essential libssl-dev libffi-dev python-dev

See the cryptography package's documentation for details.

Usage

There are two different clients available in the bankid package: the BankIDClient, which uses the SOAP-based API that is being deprecated in February 2020, and the BankIDJSONClient, which uses the new JSON API released in February 2018.

Any new deployment using PyBankID should use the BankIDJSONClient.

JSON client

>>> from bankid import BankIDJSONClient
>>> client = BankIDJSONClient(certificates=('path/to/certificate.pem',
                                            'path/to/key.pem'))

Connection to production server is the default in the client. If test server is desired, send in the test_server=True keyword in the init of the client.

WHen using the JSON client, all authentication and signing calls requires the end user's ip address to be included in all calls. An authentication order is initiated as such:

>>> client.authenticate(end_user_ip='194.168.2.25',
                        personal_number="YYYYMMDDXXXX")
{
    'autoStartToken': '798c1ea1-e67a-4df6-a2f6-164ac223fd52',
    'orderRef': 'a9b791c3-459f-492b-bf61-23027876140b'
}

and a sign order is initiated in a similar fashion:

>>> client.sign(end_user_ip='194.168.2.25',
                user_visible_data="The information to sign.",
                personal_number="YYYYMMDDXXXX")
{
    'autoStartToken': '798c1ea1-e67a-4df6-a2f6-164ac223fd52',
    'orderRef': 'a9b791c3-459f-492b-bf61-23027876140b'
}

Since the BankIDJSONClient is using the BankID v5 JSON API, the personal_number can now be omitted when calling authenticate and sign. See BankID Relying Party Guidelines for more information about this.

The status of an order can then be studied by polling with the collect method using the received orderRef:

>>> client.collect(order_ref="a9b791c3-459f-492b-bf61-23027876140b")
{
    'hintCode': 'outstandingTransaction',
    'orderRef': 'a9b791c3-459f-492b-bf61-23027876140b',
    'status': 'pending'
}
>>> client.collect(order_ref="a9b791c3-459f-492b-bf61-23027876140b")
{
    'hintCode': 'userSign',
    'orderRef': 'a9b791c3-459f-492b-bf61-23027876140b',
    'status': 'pending'
}
>>> c.collect(order_ref="a9b791c3-459f-492b-bf61-23027876140b")
{
    'completionData': {
        'cert': {
            'notAfter': '1581289199000',
            'notBefore': '1518130800000'
        },
        'device': {
            'ipAddress': '0.0.0.0'
        },
        'ocspResponse': 'MIIHegoBAKCCB[...]',
        'signature': 'PD94bWwgdmVyc2lv[...]',
        'user': {
            'givenName': 'Namn',
            'name': 'Namn Namnsson',
            'personalNumber': 'YYYYMMDDXXXX',
            'surname': 'Namnsson'
        }
    },
    'orderRef': 'a9b791c3-459f-492b-bf61-23027876140b',
    'status': 'complete'
}

Please note that the collect method should be used sparingly: in the BankID Relying Party Guidelines it states that "collect should be called every two seconds and must not be called more frequent than once per second".

SOAP client

For documentation about how to use the SOAP client, see the documentation.

Certificates

Production certificates

If you want to use BankID in a production environment, then you will have to purchase this service from one of the selling banks. They will then provide you with a certificate that can be used to authenticate your company/application with the BankID servers.

This certificate has to be processed somewhat to be able to use with PyBankID, and how to do this depends on what the selling bank provides you with.

Test certificate

The certificate to use when developing against the BankID test servers can be obtained through PyBankID:

>>> import os
>>> import bankid
>>> dir_to_save_cert_and_key_in = os.path.expanduser('~')
>>> cert_and_key = bankid.create_bankid_test_server_cert_and_key(
    dir_to_save_cert_and_key_in)
>>> print(cert_and_key)
['/home/hbldh/certificate.pem', '/home/hbldh/key.pem']
>>> client = bankid.BankIDJSONClient(
    certificates=cert_and_key, test_server=True)

Python 2, urllib3 and certificate verification

An InsecurePlatformWarning is issued when using the client in Python 2 (See urllib3 documentation). This can be remedied by installing pybankid with the security extras as described above, or to manually install pyopenssl according to this issue and docstrings in requests.

Optionally, the environment variable PYBANKID_DISABLE_WARNINGS can be set to disable these warnings.

Testing

The PyBankID solution can be tested with pytest:

py.test tests/