This package provides integration with the payment gateway, DataCash. It is designed to work seamlessly with the e-commerce framework django-oscar but can be used without Oscar.
It also supports batch fraud screening using The3rdMan.
When following the below instructions, it may be helpful to browse the sandbox folder above as this is an example Oscar install which has been integrated with Datacash.
From PyPi:
pip install django-oscar-datacash
or from Github:
pip install git+git://github.com/tangentlabs/django-oscar-datacash.git#egg=django-oscar-datacash
Add 'datacash'
to INSTALLED_APPS
and run:
./manage.py migrate datacash
to create the appropriate database tables.
Edit your settings.py
to set the following settings:
DATACASH_HOST = 'testserver.datacash.com' DATACASH_CLIENT = '...' DATACASH_PASSWORD = '...' DATACASH_CURRENCY = 'GBP'
There are other settings available (see below). Obviously, you'll need to specify different settings in your test environment as opposed to your production environment.
You'll need to use a subclass of oscar.apps.checkout.views.PaymentDetailsView
within your own
checkout views. See Oscar's documentation on how to create a local version of the checkout app.
Override the handle_payment
method (which does nothing by default) and add your integration code. See
the sandbox site for an example of integrating Datacash payments
into an Oscar site.
The gateway modules uses the named logger datacash
.
The3rdMan callbacks use the named logger datacash.the3rdman
. It is
recommended that you use django.utils.log.AdminMailHandler
with this logger
to ensure error emails are sent out for 500 responses.
Many Datacash features require your merchant account to be configured correctly. For instance, the default Datacash set-up won't include:
- Payments using historic transactions
- Split settlements
When investigating problems, make sure your Datacash account is set-up correctly.
Using realtime fraud services requires submitting a dict of relevant data as part of the initial transaction. A helper method is provided that will extract all it needs from Oscar's models:
from datacash.the3rdman import build_data_dict
fraud_data = build_data_dict(
request=request,
order_number='1234',
basket=request.basket,
email=email
shipping_address=shipping_address,
billing_addres=billing_address)
then pass this data as a named argument when creating the transaction:
ref = Facade().pre_authorise(..., the3rdman_data=fraud_data)
To receive the callback, include the following in your urls.py
:
urlpatterns = patterns('',
...
(r'^datacash/', include('datacash.urls')),
...
)
When a fraud response is received, a custom signal is raised which your client code should listen for. Example:
from django.dispatch import receiver
from datacash.the3rdman import signals
@receiver(signals.response_received)
def handle_fraud_response(sender, response, **kwargs):
# Do something with response
There are two key components:
The class datacash.gateway.Gateway
provides fine-grained access to the
various DataCash APIs, which involve constructing XML requests and decoding XML
responses. All calls return a datacash.gateway.Response
instance which
provides dictionary-like access to the attributes of the response.
Example calls:
from decimal import Decimal as D
from datacash.gateway import Gateway
gateway = Gateway()
# Single stage processing
response = gateway.auth(amount=D('100.00'), currency='GBP',
merchant_reference='AA_1234',
card_number='4500203021916406',
expiry_date='10/14',
ccv='345')
response = gateway.refund(amount=D('100.00'), currency='GBP',
merchant_reference='AA_1234',
card_number='4500203021916406',
expiry_date='10/14',
ccv='345')
# Two-stage processing (using pre-registered card)
response = gateway.pre(amount=D('50.00'), currency='GBP',
previous_txn_reference='3000000088888888')
response = gateway.fulfill(amount=D('50.00'), currency='GBP',
txn_reference=response['datacash_reference'])
The gateway object know nothing of Oscar's classes and can be used in a stand-alone manner.
The class datacash.facade.Facade
wraps the above gateway object and provides a
less granular API, as well as saving instances of datacash.models.OrderTransaction
to
provide an audit trail for Datacash activity.
DATACASH_HOST
- Host of DataCash serverDATACASH_CLIENT
- UsernameDATACASH_PASSWORD
- PasswordDATACASH_CURRENCY
- Currency to use for transactionsDATACASH_USE_CV2AVS
- Whether to pass CV2AVS dataDATACASH_CAPTURE_METHOD
- The 'capture method' to use. Defaults to 'ecomm'.
To work on django-oscar-datacash
, clone the repo, set up a virtualenv and install
in develop mode:
make install
The test suite can then be run using:
./runtests.py
There is a sandbox Oscar site that can be used for development. Create it with:
make sandbox
and browse it with:
python sandbox/manage.py runserver
Magic card numbers are available on the Datacash site: https://testserver.datacash.com/software/download.cgi?show=magicnumbers
Here's an example:
1000010000000007
Have fun!
- Added support for Python 3
- Added support for Django 1.7
- Drop use of deprecated bankcard attribute
- Ensure compatibility with Oscar 0.7
- Drop support for Oscar 0.4 and 0.5
- Ensure compatibility with Django 1.6 and Oscar 0.6
- Ensure templates are compatible with Django 1.5
- Ensure tests pass with Oscar 0.6
- Format country codes in fraud submission. They must be 3 digits.
- Allow the transaction currency to be set pre transaction. This is to support the new multi-currency features of Oscar 0.6.
- Fix logging formatting bug
- Remove uniqueness constraint for 3rdman response
- Add links to Gatekeeper from dashboard
- Adjust how the response type of callback is determined
- Add support for The3rdMan fraud screening
- Fix mis-handling of datetimes introduced in 0.4.1
- Handle bankcard dates passed as
datetime.datetime
instances instead of strings. This is a compatability fix for Oscar 0.6 development.
- Oscar 0.5 support
- Merchants passwords now removed from saved raw request XML
- A random int is now appended to the merchant ref to avoid having duplicates
- Minor tweak to sort order of transactions in dashboard
- Updated packaging to include HTML templates
- Added handling for split shipment payments
- Added sandbox site
- Added dashboard view of transactions
- Added admin.py
- Added travis.ci support
- Fixed bug with currency in refund transactions
- Fixed issue with submitting currency attribute for historic transactions