/drf-elasticsearch-dsl

django rest framework + elasticsearch-dsl.py

Primary LanguagePythonMIT LicenseMIT

DRF Elasticsearch DSL

https://circleci.com/gh/ajbeach2/drf-elasticsearch-dsl.svg?style=svg

DRF Elasticsearch DSL is losely based on django-haystack and provides a ModelSerializerDocument which supports all of the field types provided by elastic-search-dsl persistence.. ModelSerializerDocument is losely based on the DocType class provided by elasticsearch-dsl.py

The purpose of this libraray is to allow definition of elasticsearch documents with DRF's ModelSerializer class while optionally providing support for async document updates and deletes with celery.

Documentation

The full documentation is at https://drf-elasticsearch-dsl.readthedocs.io.

Quickstart

Install Django Package:

pip install drf-elasticsearch-dsl

Add it to your INSTALLED_APPS:

INSTALLED_APPS = (
    ...
    'drf_elasticsearch_dsl.apps.DrfElasticsearchDsl',
    ...
)

Configure DRF_SERIALIZER_ELASTICSERACH_SETTTINGS in your settings.py file with your elasticsearch url(s)

DRF_SERIALIZER_ELASTICSERACH_SETTTINGS = {
    'elasticsearch_hosts': ['localhost']
}

Create a Model

from django.db import models


class Contact(models.Model):

    first_name = models.CharField(max_length=32, null=False, blank=False)
    last_name = models.CharField(max_length=32, null=False, blank=False)
    url = models.URLField(null=False, blank=False)
    email = models.EmailField(max_length=254, null=False, blank=False)
    bio = models.TextField(null=False, blank=False)
    birthday = models.DateField(null=False, blank=False)

Create a ModelSerializer

from rest_framework import serializers

class ContactSerializer(serializers.ModelSerializer):

    class Meta:
        model = Contact
        fields = '__all__'

Create a search_indexes.py, which should be in the root of the application. Add your ModelSerializerDocument classes here. The specificed index will have its mapping updated for this document.

from drf_elasticsearch_dsl.documents import ModelSerializerDocument
from elasticsearch_dsl import Date, Keyword, Text, String
from .serialziers import ContactSerializer


class ContactSerializerDocument(ModelSerializerDocument):
    first_name = String()
    last_name = String()
    url = Keyword()
    email = Keyword()
    bio = Text()
    birthday = Date()

    class Meta:
        index = 'myapp'
        serializer = ContactSerializer
        doc_type = 'myapp.contact'

Finally, sync your database with elasticsearch by running:

$ python manage.py update_index

Features

Celery Support

By default, dr-elasticsearch-dsl does not setup signals to sync models on save or delete. To enable celery support, add the following to your settings.py configuration:

DRF_SERIALIZER_ELASTICSERACH_SETTTINGS = {
    ...
    'signal_processor_class': 'drf_elasticsearch_dsl.signals.CelerySignalProcessor',
}

See the celery documentation for details setting up celery with django

Running Tests

Does the code actually work?

source <YOURVIRTUALENV>/bin/activate
(myenv) $ pip install -r requirements_test.txt
(myenv) $ tox

TODO:

  • Add search URLS to be automatically added to all ModelSerializerDocument added to search_indexes.py
  • Better documentation
  • Better test coverage

Credits

Tools used in rendering this package: