fuzzy search with trigram

[benzkji]

plus add something like .annotate(len(content)).filter(content_lentg_similaritysearch_length__gt=xy), otherwise way too many searchresults.
[bauart]

            config = conf.LANGUAGE_2_PGCONFIG.get(self.request.LANGUAGE_CODE, "english")
            search_vector = SearchVector(
                "content",
                "title",
                config=config,
            )
            search_query = SearchQuery(q)
            qs = (
                IndexEntry.objects.annotate(
                    search=search_vector,
                    rank=SearchRank(search_vector, search_query),
                    tg_similarity=TrigramSimilarity(Unaccent("title"), q)
                    + TrigramSimilarity(Unaccent("content"), q),
                )
                .filter(index_key=self.request.LANGUAGE_CODE)
                .filter(
                    Q(search=q)
                    | Q(tg_similarity__gt=0)
                    # Q(content__trigram_similar=q)
                    # | Q(title__trigram_similar=q)
                )
                .order_by("-rank", "-tg_similarity")
            )

UPDATE: this is really hard to balance right. Many times it will find too many and awkward results. To be continued.