Les majuscules impactent les résultats de recherche

[pierre-24]

Du forum:

Par exemple, je cherche 'Nazara', je trouve le bon résultat, pas de problème mais quand je tape 'nazara', je trouve les même résultats mais rangés très différemment, je comprends pas trop la logique, c'est sensible à la casse mais j'ai pourtant les mêmes résultats que quand j'écris avec une majuscule sauf que c'est pas ordonné pareil. ^^

---

Et en effet : nazara vs Nazara.

Ce qui est bizarre, parce que Elasticsearch est configuré chez nous pour être case insensitive. De plus, Elasticsearch comprend à priori la même chose:

[pierre-24]

Du coup, petit devoir pour @vhf : exécuter ces deux commandes et me donner les retours:

curl "localhost:9200/zds_search/_search?explain&pretty" -d '{"from":0, "size": 10, "query": {"function_score": {"query": {"bool": {"should": [{"bool": {"must": [{"match": {"_type": "chapter"}}, {"multi_match": {"fields": ["title", "text"], "query": "nazara"}}]}}, {"function_score": {"query": {"bool": {"must": [{"match": {"_type": "publishedcontent"}}, {"multi_match": {"fields": ["title", "description", "categories", "tags", "text"], "query": "nazara"}}]}}, "functions": [{"filter": {"match": {"content_type": "TUTORIAL"}}, "weight": 1.0}, {"filter": {"match": {"content_type": "ARTICLE"}}, "weight": 1.0}], "boost_mode": "multiply"}}, {"function_score": {"query": {"bool": {"must": [{"match": {"_type": "topic"}}, {"terms": {"forum_pk": [1, 2, 3, 16, 17, 18, 19, 20, 21, 22, 23, 24]}}, {"multi_match": {"fields": ["title", "subtitle", "tags"], "query": "nazara"}}]}}, "functions": [{"filter": {"match": {"is_solved": true}}, "weight": 1.1}, {"filter": {"match": {"is_sticky": true}}, "weight": 1.2}, {"filter": {"match": {"is_locked": true}}, "weight": 0.1}], "boost_mode": "multiply"}}, {"function_score": {"query": {"bool": {"must": [{"match": {"_type": "post"}}, {"terms": {"forum_pk": [1, 2, 3, 16, 17, 18, 19, 20, 21, 22, 23, 24]}}, {"term": {"is_visible": true}}, {"multi_match": {"fields": ["text_html"], "query": "nazara"}}]}}, "functions": [{"filter": {"match": {"position": 1}}, "weight": 1.2}, {"filter": {"match": {"is_useful": true}}, "weight": 1.5}, {"filter": {"range": {"like_dislike_ratio": {"gt": 1}}}, "weight": 1.05}, {"filter": {"range": {"like_dislike_ratio": {"lt": 1}}}, "weight": 0.95}], "boost_mode": "multiply"}}]}}, "functions": [{"filter": {"match": {"_type": "topic"}}, "weight": 2.0}, {"filter": {"match": {"_type": "publishedcontent"}}, "weight": 3.0}, {"filter": {"match": {"_type": "post"}}, "weight": 1.0}, {"filter": {"match": {"_type": "chapter"}}, "weight": 1.5}], "boost_mode": "multiply"}}}'

curl "localhost:9200/zds_search/_search?explain&pretty" -d '{"from":0, "size": 10, "query": {"function_score": {"query": {"bool": {"should": [{"bool": {"must": [{"match": {"_type": "chapter"}}, {"multi_match": {"fields": ["title", "text"], "query": "Nazara"}}]}}, {"function_score": {"query": {"bool": {"must": [{"match": {"_type": "publishedcontent"}}, {"multi_match": {"fields": ["title", "description", "categories", "tags", "text"], "query": "Nazara"}}]}}, "functions": [{"filter": {"match": {"content_type": "TUTORIAL"}}, "weight": 1.0}, {"filter": {"match": {"content_type": "ARTICLE"}}, "weight": 1.0}], "boost_mode": "multiply"}}, {"function_score": {"query": {"bool": {"must": [{"match": {"_type": "topic"}}, {"terms": {"forum_pk": [1, 2, 3, 16, 17, 18, 19, 20, 21, 22, 23, 24]}}, {"multi_match": {"fields": ["title", "subtitle", "tags"], "query": "Nazara"}}]}}, "functions": [{"filter": {"match": {"is_solved": true}}, "weight": 1.1}, {"filter": {"match": {"is_sticky": true}}, "weight": 1.2}, {"filter": {"match": {"is_locked": true}}, "weight": 0.1}], "boost_mode": "multiply"}}, {"function_score": {"query": {"bool": {"must": [{"match": {"_type": "post"}}, {"terms": {"forum_pk": [1, 2, 3, 16, 17, 18, 19, 20, 21, 22, 23, 24]}}, {"term": {"is_visible": true}}, {"multi_match": {"fields": ["text_html"], "query": "Nazara"}}]}}, "functions": [{"filter": {"match": {"position": 1}}, "weight": 1.2}, {"filter": {"match": {"is_useful": true}}, "weight": 1.5}, {"filter": {"range": {"like_dislike_ratio": {"gt": 1}}}, "weight": 1.05}, {"filter": {"range": {"like_dislike_ratio": {"lt": 1}}}, "weight": 0.95}], "boost_mode": "multiply"}}]}}, "functions": [{"filter": {"match": {"_type": "topic"}}, "weight": 2.0}, {"filter": {"match": {"_type": "publishedcontent"}}, "weight": 3.0}, {"filter": {"match": {"_type": "post"}}, "weight": 1.0}, {"filter": {"match": {"_type": "chapter"}}, "weight": 1.5}], "boost_mode": "multiply"}}}'

Bon, ça va générer BEAUCOUP de texte, donc on fait comme la fois passée, utilise des gists. D'avance merci :)

Note pour le futur: la liste des forums en local et en live ne matchent évidement pas. Pas oublier de la changer par [1, 2, 3, 16, 17, 18, 19, 20, 21, 22, 23, 24].

[pierre-24]

Trouvé : https://www.elastic.co/guide/en/elasticsearch/reference/5.2/keyword.html#keyword

They are typically used for filtering (Find me all blog posts where status is published), for sorting, and for aggregations. Keyword fields are only searchable by their exact value.

[gllmc]

Est-ce qu'on n'a pas oublié de fermer ?