Rubber provides you with tools to easily setup, manage and index your Django models in ElasticSearch. It relies on celery and elasticsearch_dsl (for backward compatibility purposes).

It is designed to allow simultaneous indexing of an object on different indices.

This project is a mutation of django-trampoline.

Settings

Add rubber to your INSTALLED_APPS.

Define the following setting:

RUBBER = {
    'HOSTS': ['localhost:9200'],
    'MODELS': [
        'tests.models.Token'
    ],
    'CONFIG_ROOT': os.path.join(SITE_ROOT, 'es_configs'),
    'OPTIONS': {
        'disabled': False,
        'fail_silently': True,
        'celery_queue': None
    },
}

OPTIONS

celery_queue

None by default.

Specify which Celery queue should handle your indexation tasks.

fail_silently

True by default.

If fail_silently is True exceptions raised while indexing are caught and logged without being re-raised.

disabled

False by default.

ESIndexableMixin

from rubber.mixins import ESIndexableMixin

In order to make your model indexable you must make it inherit from ESIndexableMixin and implement a few things.

get_es_indexers() (required)

Return a dictionnary of indexers to be used for this model.

def get_es_indexers(self):
    return {
        'INDEX_1': {
            'version': 1,
            'index': 'index_1',
            'serializer': Serializer,
            'doc_type': 'token'
        },
        'INDEX_2': {
            'version': 1,
            'dsl_doc_type': DocType,
            'dsl_doc_type_mapping': self.dsl_doc_type_mapping
        },
    }

Whenever an object is saved it will be indexed in both index_1_v1 and index_2_v1.

• INDEX_1 uses a serializer class to map the object's properties to a JSON serializable dictionnary. It takes the object instance as its first argument and implements the data property.

For example:

class Serializer(object):

    def __init__(self, token, *args, **kwargs):
        self.token = token

    @property
    def data(self):
        return {
            'name': self.token.name,
            'number': self.token.number,
            'multi': ['item_1', 'item_2']
        }

• INDEX_2 uses a DocType from elasticsearch_dsl and a method returning a mapped instance of it. This solution is only there for backward compatibility with the older system.

is_indexable() (optional)

def is_indexable(self):
    return True

Tell whether a particular instance of the model should be indexed or skipped (defaults to true).

get_indexable_queryset() (optional)

@classmethod
def get_indexable_queryset(cls):
    return []

Return the list of contents that should be indexed for this model using the command es_create_documents() defined bellow. Remember to use the classmethod decorator.

es_reference_date (optional)

Reference date used by the command es_create_documents --from option. Defaults to modified_at.

Mapping

Management commands

All management commands accept the following arguments:

• --help: Display an help message and the available arguments for the command.

es_create_index

Create the indices passed as arguments using the mappings defined inside CONFIG_ROOT.

Following the previous example for get_es_serializers you would need to create two files index_1_v1.json and index_2_v1 inside CONFIG_ROOT. Remember to include the version in your index name.

You would then run the command: es_create_index index_1_v1 index_2_v1.

es_create_documents

Create documents based on the method get_es_indexers() on the related models.

Arguments:

• --models (optional): Comma separated list of models to be indexed. It must match at least one of the models defined in RUBBER settings. Defaults to all.
• --from (optional): Filter queryset by date. Must be formatted as YYYY-MM-DDTHH:MM:SS.
• --show-tqdm (optional): Show the tqdm progress bar.
• --dry-run (optional): Run the command in dry run mode without actually commiting anything.