Search backends

Compound search filter backend

Compound search filter backend aims to replace old style SearchFilterBackend.

Sample view

from django_elasticsearch_dsl_drf.filter_backends import (
    DefaultOrderingFilterBackend,
    CompoundSearchFilterBackend,
    OrderingFilterBackend,
)
from django_elasticsearch_dsl_drf.viewsets import DocumentViewSet

from .documents import BookDocument
from .serializers import BookDocumentSerializer

class BookCompoundSearchBackendDocumentViewSet(DocumentViewSet):

    document = BookDocument
    serializer_class = BookDocumentSerializer
    lookup_field = 'id'

    filter_backends = [
        # ...
        OrderingFilterBackend,
        DefaultOrderingFilterBackend,
        CompoundSearchFilterBackend,
        # ...
    ]

    search_fields = (
        'title',
        'description',
        'summary',
    )

    ordering = ('_score', 'id', 'title', 'price',)

Sample request

http://localhost:8000/search/books-compound-search-backend/?search=enim

Generated query

{
  "from": 0,
  "sort": [
    "id",
    "title",
    "price"
  ],
  "size": 23,
  "query": {
    "bool": {
      "should": [
        {
          "match": {
            "title": {
              "query": "enim"
            }
          }
        },
        {
          "match": {
            "description": {
              "query": "enim"
            }
          }
        },
        {
          "match": {
            "summary": {
              "query": "enim"
            }
          }
        }
      ]
    }
  }
}

Multi match search filter backend

Document and serializer definition are trivial (there are lots of examples in other sections).

Sample view

from django_elasticsearch_dsl_drf.filter_backends import (
    DefaultOrderingFilterBackend,
    MultiMatchSearchFilterBackend,
    OrderingFilterBackend,
)
from django_elasticsearch_dsl_drf.viewsets import DocumentViewSet

from .documents import BookDocument
from .serializers import BookDocumentSerializer


class BookMultiMatchSearchFilterBackendDocumentViewSet(DocumentViewSet):

    document = BookDocument
    serializer_class = BookDocumentSerializer
    lookup_field = 'id'

    filter_backends = [
        # ...
        OrderingFilterBackend,
        DefaultOrderingFilterBackend,
        MultiMatchSearchFilterBackend,
        # ...
    ]

    search_fields = {
        'title': {'boost': 4},
        'summary': {'boost': 2},
        'description': None,
    }

    ordering = ('_score', 'id', 'title', 'price',)

Sample request

Note

Multiple search params (search_multi_match) are not supported. Even if you provide multiple search params, the first one would be picked, having the rest simply ignored.

http://localhost:8000/search/books-multi-match-search-backend/?search_multi_match=debitis%20enim

Generated query

{
  "from": 0,
  "query": {
    "multi_match": {
      "query": "debitis enim",
      "fields": [
        "summary^2",
        "description",
        "title^4"
      ]
    }
  },
  "size": 38,
  "sort": [
    "_score",
    "id",
    "title",
    "price"
  ]
}

Options

All standard multi match query options are available/tunable with help of multi_match_options view property.

Selective list of available options:

• operator
• type
• analyzer
• tie_breaker

Type options

See the Elasticsearch docs for detailed explanation.

• best_fields
• most_fields
• cross_fields
• phrase
• phrase_prefix

Example

class BookMultiMatchSearchFilterBackendDocumentViewSet(DocumentViewSet):

    # ...

    multi_match_options = {
        'type': 'phrase'
    }

Operator options

Can be either and or or.