django-es-drf 1.0.0a1

Simple django - elasticsearch - drf integration

Django ES DRF

Work in progress

A simple integration layer between Django, Elasticsearch and Django rest framework

Model and ES Document example

To declare your own document, create a model and register it with a DjangoDocument

from django.db import models
from django_es_drf import registry, DjangoDocument
from rest_framework.serializers import ModelSerializer


class School(models.Model):
    name = models.CharField(max_length=100)
    address = models.TextField()


@registry.register(School)
class SchoolDocument(DjangoDocument):
    class Index:
        name = "schools"

If you want to split the file into models, serializers and documents, be free to do so but ensure that the file including documents is loaded at the startup - in app's ready function, imported at the bottom of the model etc.

Later on, use SchoolDocument (which is a subclass of elasticsearch-dsl document):

for s in SchoolDocument.search().filter('term', name='Blah'):
    print(s)
    s.name = s.name + '-test'
    s.save()  # will save django and index to ES

DRF example

The simplest case is:

class SchoolAPI(ESViewSet):
    document = SchoolDocument

Search, facets

Luqum search

Django Document and mapping

Custom declaration for fields

Any fields that are already present on the document will be keep as they are. In the example above, to set that name is a text and not keyword (the default), just declare:

import elasticsearch_dsl as e


@registry.register(School, SchoolSerializer)
class SchoolDocument(DjangoDocument):
    name = e.Text()

    class Index:
        name = "schools"

Excluding fields

To exclude a field from the automatic generation, add it to excluded_fields:

@registry.register(School, SchoolSerializer,
                   excluded_fields=['address'])
class SchoolDocument(DjangoDocument):
    class Index:
        name = "schools"

Custom mapping between serializer fields and ES fields

Add your own mapping - the key is the DRF field type, value is a function that takes the field and context and returns ES field: The context is an instance of RegistrationContext.

import elasticsearch_dsl as e


@registry.register(School, SchoolSerializer,
                   mapping={
                       TextField: lambda fld, ctx: e.Keyword()
                   })
class SchoolDocument(DjangoDocument):
    class Index:
        name = "schools"

Disabling the mapping

Add generate=False to decorator's parameters:

import elasticsearch_dsl as e


@registry.register(School, SchoolSerializer,
                   generate=False)
class SchoolDocument(DjangoDocument):
    # you need to provide your own mapping here

    class Index:
        name = "schools"

Relations

The framework does not generate code for relations - if you need this, do it in serializer and add your own mapping, or use a more complete library, such as django-elasticsearch-dsl-drf.

Serializer

The serializer is just a plain DRF serializer that converts django fields to document's fields. When autogenerated mapping is used, just use the plain empty serializer.

Note: see Relations section above if you need to serialize relations

Objects and nested

Viewsets