Introduction

This package provides a small library for declaratively specifying indexes for Django models using an Elasticsearch backend.

It requires Django, elasticsearch-dsl and an available Elasticsearch instance.

Compatibility

The following compatibility matrix summarizes supported versions of the Elasticsearch engine.

	ES 1.X	ES 2.X	ES 5.X	ES 6.X	ES 7.X	ES 8.X
1.0	Y	N	N	N	N	N
1.1	Y	N	N	N	N	N
1.2	N	Y	N	N	N	N
7.0	N	N	? [1]	? [1]	Y	? [1]
8.0	N	N	N	N	N	Y

[1]	(1, 2, 3) Compatibilty is possible but not officially supported.

Releases prior to 7.0 unofficially supported a variety of Django release series and did not explicitly indicate supported combinations. No official support or guidance is given for these releases.

	Django 3.2	Django 4.2
7.0	Y	N
8.0	Y	planned

Usage

1. Add inelastic_models to INSTALLED_APPS.

2. Mixin the type inelastic_models.indexes.SearchMixin to your models.

3. Implement a type inelastic_models.indexes.Search and bind it to models:

   from .models import Foo
   
   class FooIndex(Search):
       attribute_fields = ('foo', 'baz')
   
   FooIndex.bind_to_model(Foo)
   

You must define ELASTICSEARCH_CONNECTIONS. Pass index and connection parameters to the generated indices and the underlying Elasticsearch instance via the INDEX_OPTIONS and CONNECTION_OPTIONS mappings, respectively:

ELASTICSEARCH_CONNECTIONS = {
    'default': {
        'HOSTS': ['http://localhost:9200'],
        'INDEX_NAME': 'inelastic_models',
        'INDEX_OPTIONS': {
            'number_of_replicas': 3
        },
        'CONNECTION_OPTIONS': {
            'timeout': 42,
            'retry_on_timeout': True
        }
    }
},

Tests

Run tests using the make rule:

make test [venv=<path>] [python=<python executable name, e.g., 'python3.5'>]

It is assumed that you have and Elasticsearch index available at elasticsearch:9200 and that virtualenv available on your path.