
A small django application for marketing using referral links
## What is this app good for?
Imagine you got a nice website running and you want to make some marketing in order to gain more users.
This can be achieved by spreading links/banners over the internet pointing to your website.
Now in order to find out how your banners performed you need to know how many users registered
after clicking on one of your banners. To tell them apart you add some unique GET parameter to
each link. This app will detect those referral parameters and store for each user what referrer
lead him to your website.
## Key Features
* Manual and automatic referrer generation.
* Referrer organisation in campaigns.
* Automatic associating of referrers with campaigns using patterns.
* Referrers will be cached in sessions, so when a user first browses your site and does not
immediately register the referrer will be cached and assigned when the user finally decides
to register. That way you can link to any site and not necessarily a register page.
* Very easy integration.
* Completely configurable to your needs.
* django 1.7 migrations + legacy south support
* No dependencies
* 100% Test coverage
* Supports django 1.4 - 1.7
* supports python 2 + 3
## Download
### using pip
pip install django-referral
### github
https://github.com/byteweaver/django-referral
## Integration
### 1. Add referral to your INSTALLED_APPS
INSTALLED_APPS = (
...
'referral',
...
)
### 2. Add ReferrerMiddleware to your MIDDLEWARE_CLASSES
Since django-referral uses django's session middleware make sure you add it after SessionMiddleware
MIDDLEWARE_CLASSES = (
...
'django.contrib.sessions.middleware.SessionMiddleware',
'referral.middleware.ReferrerMiddleware',
...
)
### 3. Trigger the referrer association after user creation
from referral.models import UserReferrer
...
def my_user_creation_view(request):
...
UserReferrer.objects.apply_referrer(user, request)
## Settings
###REFERRAL_GET_PARAMETER
The name of the GET parameter used.
Default: "ref"
###REFERRAL_SESSION_KEY
The name of the session key that will hold the detected referrer
Default: "referrer"
###REFERRAL_AUTO_CREATE
Defines whether unknown referrers shall be autocreated
Default: True
###REFERRAL_AUTO_ASSOCIATE
Defines whether referrers should be associated to campaigns automatically using patterns
Default: True
###REFERRAL_CASE_SENSITIVE
Defines whether or referrer names are case-sensitive or not.
Default: False
## Testing
Just run the makefile to set up a virtual environment for testing
make
For simple test
make test
For coverage report
make coverage
For full test in all supported environment (requires tox)
tox
## Changelog
### V1.0.0
* add django 1.7 support
* add pyton3 support
* additional test cases
* code cleanup
* some pep8 fixes
Metadata
Release files for django-referral 1.0.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| django-referral-1.0.1.tar.gz | 10.1 kB | Without provenance | Details |
Release files / django-referral-1.0.1.tar.gz
| Download URL | django-referral-1.0.1.tar.gz |
|---|---|
| Size | 10.1 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
7bc51146eff3610227558c418a3787d4221b9c7511d785b0a3cd0034f13b0d66
|
|
BLAKE2b-256 checksum How to use checksums |
9a0c40a639785862fc4f62170b2f778c9f02920d114a74a065a972607646578f
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
Security for django-referral 1.0.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.