Django ManyToMany History

Django ManyToMany relation field with history of changes. Like usual Django's ManyToManyField, it's generate intermediary join table to represent the many-to-many relationship, but with two additional columns: 'time_from' and 'time_to'. Using updated interface of field it's possible to retreive history of all versions of this field's value.

Compatibility

• python v2.7, v3.4
• django >=1.5, <=1.8. Does not compatible with Django 1.9
• postgres

Installation

pip install django-m2m-history

Add into settings.py lines:

INSTALLED_APPS = (
    ...
    'm2m_history',
)

Usage example

Make 2 models with many-to-many relationship:

class Publication(models.Model):
    title = models.CharField(max_length=30)

    def __str__(self):
        return self.title

class Article(models.Model):
    headline = models.CharField(max_length=100)
    publications = ManyToManyHistoryField(Publication)

    def __str__(self):
        return self.headline

Create objects and relations betweeen them:

>>> p1 = Publication.objects.create(title='Pub1')
>>> p2 = Publication.objects.create(title='Pub2')
>>> p3 = Publication.objects.create(title='Pub3')

>>> article = Article.objects.create(headline='Article1')
>>> state_time1 = datetime.now()

>>> article.publications = [p1, p2]
>>> state_time2 = article.publications.last_update_time()
>>> article.publications.all()
[<Publication: Pub1>, <Publication: Pub2>]
>>> article.publications.count()
2
>>> article.publications.through.objects.count()
2

>>> article.publications = [p3]
>>> state_time3 = article.publications.last_update_time()
>>> article.publications.all()
[<Publication: Pub3>]
>>> article.publications.count()
1
>>> article.publications.through.objects.count()
3

>>> article.publications.add(p2, p1)
>>> state_time4 = article.publications.last_update_time()
>>> article.publications.all()
[<Publication: Pub1>, <Publication: Pub2>, <Publication: Pub3>]
>>> article.publications.count()
3
>>> article.publications.through.objects.count()
5

>>> article.publications.remove(p2, p1)
>>> state_time5 = article.publications.last_update_time()
>>> article.publications.all()
[<Publication: Pub3>]
>>> article.publications.count()
1
>>> article.publications.through.objects.count()
5

>>> article.publications = [p1, p2]
>>> state_time6 = article.publications.last_update_time()
>>> article.publications.all()
[<Publication: Pub1>, <Publication: Pub2>]
>>> article.publications.count()
2
>>> article.publications.through.objects.count()
7

>>> article.publications.clear()
>>> state_time7 = article.publications.last_update_time()
>>> article.publications.all()
[]
 >>> article.publications.count()
0
>>> article.publications.through.objects.count()
7

Get objects of history states by timestamps:

>>> article.publications.were_at(state_time1)
[]

>>> article.publications.were_at(state_time2)
[<Publication: Pub1>, <Publication: Pub2>]

>>> article.publications.were_at(state_time3)
[<Publication: Pub3>]

>>> article.publications.were_at(state_time4)
[<Publication: Pub1>, <Publication: Pub2>, <Publication: Pub3>]

>>> article.publications.were_at(state_time5)
[<Publication: Pub3>]

>>> article.publications.were_at(state_time6)
[<Publication: Pub1>, <Publication: Pub2>]

>>> article.publications.were_at(state_time7)
[]

Get added and removed objects of history states by timestamps:

>>> article.publications.added_at(state_time3)
[<Publication: Pub3>]

>>> article.publications.removed_at(state_time3)
[<Publication: Pub1>, <Publication: Pub2>]

>>> article.publications.added_at(state_time4)
[<Publication: Pub1>, <Publication: Pub2>]

>>> article.publications.removed_at(state_time5)
[<Publication: Pub1>, <Publication: Pub2>]

>>> article.publications.added_at(state_time6)
[<Publication: Pub1>, <Publication: Pub2>]

>>> article.publications.removed_at(state_time6)
[<Publication: Pub3>]