How to use ngram in search query param

[bplociennik]

Questions

I'm trying to add ngrams to one of my fields title to use directly in the search query param. I would like to be able to find objects even by part of my name.

I tried examples on the repository and attached code in the advanced section in the documentation but I cannot figure out what I'm doing wrong.

I don't have deep knowledge about Elasticsearch so it may be problematic that I'm using an analyzer or normalizer in the wrong way.

For example, I have a product with the title "product long-description" I would like to get this product when I search (?search=value) for some parts of this title like "long", "description", "long description", "product long" etc.

documents.py:

edge_ngram_completion_filter = token_filter(
    'edge_ngram_completion_filter',
    type="edge_ngram",
    min_gram=3,
    max_gram=20
)

edge_ngram_completion = analyzer(
    "edge_ngram_completion",
    tokenizer="standard",
    filter=["lowercase", edge_ngram_completion_filter],
    char_filter=["html_strip"],
)


@INDEX.doc_type
class ProductDocument(Document):
    title = fields.TextField(
        analyzer=edge_ngram_completion,
        fields={"raw": KeywordField(normalizer=lowercase_normalizer)}
    )

views.py:

class ProductDocumentViewSet(BaseDocumentViewSet):
    document = ProductDocument
    filter_backends = [
        FilteringFilterBackend,
        OrderingFilterBackend,
        CompoundSearchFilterBackend,
        NestedFilteringFilterBackend,
        DefaultOrderingFilterBackend,
    ]
    filter_fields = {
        "id": None,
        "title": {
            "field": "title.raw", "lookups": [
                LOOKUP_FILTER_TERMS,
                LOOKUP_FILTER_PREFIX,
                LOOKUP_FILTER_WILDCARD,
                LOOKUP_QUERY_IN,
                LOOKUP_QUERY_EXCLUDE,
            ],
            },
    }
    search_fields = {"title": {"boost": 25}, "description": None}
    ordering_fields = {
        "id": None,
        "title": "title.raw",
    }

PS. Thanks for the great packages it helped me a lot to save a lot of time and hairs :D

[barseghyanartur]

Hi, For your purpose ngrams are not going to help. Ngrams help when you misspell the words. From how you describe it, what you need is "most words". Search the Elastic docs and the package documentation. It should be supported. Best regards, Artur

…

On Wed, 7 Jul 2021, 19:09 Bartosz Płóciennik, ***@***.***> wrote: *Questions* I'm trying to add ngrams to one of my fields title to use directly in the search query param. I would like to be able to find objects even by part of my name. I tried examples on the repository and attached code in the advanced section in the documentation but I cannot figure out what I'm doing wrong. I don't have deep knowledge about Elasticsearch so it may be problematic that I'm using an analyzer or normalizer in the wrong way. For example, I have a product with the title "product long-description" I would like to get this product when I search (?search=value) for some parts of this title like "long", "description", "long description", "product long" etc. *documents.py:* edge_ngram_completion_filter = token_filter( 'edge_ngram_completion_filter', type="edge_ngram", min_gram=3, max_gram=20 ) edge_ngram_completion = analyzer( "edge_ngram_completion", tokenizer="standard", filter=["lowercase", edge_ngram_completion_filter], char_filter=["html_strip"], ) @INDEX.doc_type class ProductDocument(Document): title = fields.TextField( analyzer=edge_ngram_completion, fields={"raw": KeywordField(normalizer=lowercase_normalizer)} ) *views.py:* class ProductDocumentViewSet(BaseDocumentViewSet): document = ProductDocument filter_backends = [ FilteringFilterBackend, OrderingFilterBackend, CompoundSearchFilterBackend, NestedFilteringFilterBackend, DefaultOrderingFilterBackend, ] filter_fields = { "id": None, "title": { "field": "title.raw", "lookups": [ LOOKUP_FILTER_TERMS, LOOKUP_FILTER_PREFIX, LOOKUP_FILTER_WILDCARD, LOOKUP_QUERY_IN, LOOKUP_QUERY_EXCLUDE, ], }, } search_fields = {"title": {"boost": 25}, "description": None} ordering_fields = { "id": None, "title": "title.raw", } PS. Thanks for the great packages it helped me a lot to save a lot of time and hairs :D — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#259>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABFSRE4WI6LKDYX3C6Q57EDTWSC3JANCNFSM477B6WKA> .

[bplociennik]

you are right, thanks :)