Elasticsearch查询异常：期望ID23结果排首位未达成

当前查询目标是覆盖primaryTitle、primaryFull、primaryText和secondaryText四个字段，期望ID为23（primaryTitle值为"ли джун хёк"）的结果排在首位，但实际返回结果中该条目仅排第三。以下是查询语句、返回结果及问题分析：

执行的查询语句 #

{"from":0,"size":100,"min_score":0.0,"explain":false,
"query":{"bool":{"must":[{"bool":{
"should":[
{"term":{"primaryTitle":{"value":"ли джун хек","boost":1.0}}},
{"function_score":{"query":{"match":{"primaryText":{"query":"ли джун хек","operator":"OR","prefix_length":0,"max_expansions":50,"fuzzy_transpositions":true,"lenient":false,"zero_terms_query":"NONE","auto_generate_synonyms_phrase_query":true,"boost":1.0}}},"functions":[{"filter":{"match_all":{"boost":1.0}},"field_value_factor":{"field":"searchWeight","factor":1.0,"modifier":"none"}}],"score_mode":"multiply","boost_mode":"multiply","max_boost":3.4028235E38,"boost":1.0}},
{"function_score":{"query":{"match":{"secondaryText":{"query":"ли джун хек","operator":"OR","prefix_length":0,"max_expansions":50,"fuzzy_transpositions":true,"lenient":false,"zero_terms_query":"NONE","auto_generate_synonyms_phrase_query":true,"boost":1.0}}},"functions":[{"filter":{"match_all":{"boost":1.0}},"field_value_factor":{"field":"searchWeight","factor":1.0,"modifier":"none"}}],"score_mode":"multiply","boost_mode":"multiply","max_boost":3.4028235E38,"boost":1.0}},
{"function_score":{"query":{"match":{"primaryFull":{"query":"ли джун хек","operator":"OR","prefix_length":0,"max_expansions":50,"fuzzy_transpositions":true,"lenient":false,"zero_terms_query":"NONE","auto_generate_synonyms_phrase_query":true,"boost":1.0}}},"functions":[{"filter":{"match_all":{"boost":1.0}},"field_value_factor":{"field":"searchWeight","factor":1.0,"modifier":"none"}}],"score_mode":"multiply","boost_mode":"multiply","max_boost":3.4028235E38,"boost":1.0}}
],
"adjust_pure_negative":true,"minimum_should_match":"1","boost":1.0}
},{"term":{"moderated":{"value":true,"boost":1.0}}}],"adjust_pure_negative":true,"boost":1.0}}}

返回的查询结果 #

{"took":6,"timed_out":false,"_shards":{"total":15,"successful":15,"skipped":0,"failed":0},"hits":{"total":{"value":5,"relation":"eq"},"max_score":21.115047,"hits":[{"_index":"manga.dev.ru.readmanga.manga_v0","_type":"manga","_id":"99","_score":21.115047,"_source":{"primaryTitle":"ли джун ки v live","primaryFull":"Ли Джун Ки V LIVE","moderated":true,"primaryText":"Ли Джун Ки V LIVE","searchWeight":1.1,"secondaryText":"Ли Джун Ки Lee Joon Ki"}},{"_index":"manga.dev.ru.readmanga.manga_v0","_type":"manga","_id":"100","_score":18.623653,"_source":{"primaryTitle":"стиль ли джун ки!","primaryFull":"Стиль Ли Джун Ки! JG Style! Lee Jun Ki Style!","moderated":true,"primaryText":"Стиль Ли Джун Ки! JG Style! Lee Jun Ki Style!","searchWeight":1.1,"secondaryText":"Ли Джун Ки Lee Joon Ki"}},{"_index":"manga.dev.ru.readmanga.elements.person_v0","_type":"person","_id":"23","_score":10.038389,"_source":{"primaryTitle":"ли джун хёк","personForChapter":true,"primaryFull":"Ли Джун Хёк Lee Joon Hyuk (2000)","personTypeId":7,"moderated":true,"primaryText":"Ли Джун Хёк Lee Joon Hyuk (2000)","searchWeight":0.9,"secondaryText":"actor","searchableHidden":false}},{"_index":"manga.dev.ru.readmanga.manga_v0","_type":"manga","_id":"101","_score":8.480389,"_source":{"primaryTitle":"хитмэн","primaryFull":"Хитмэн Hitman 히트맨","moderated":true,"primaryText":"Хитмэн Hitman 히트맨","searchWeight":1.1,"secondaryText":"Ли Джун Хёк Lee Joon Hyuk (2000)"}},{"_index":"manga.dev.ru.readmanga.elements.person_v0","_type":"person","_id":"22","_score":5.7027664,"_source":{"primaryTitle":"ли джун ки","personForChapter":true,"primaryFull":"Ли Джун Ки Lee Joon Ki","personTypeId":7,"moderated":true,"primaryText":"Ли Джун Ки Lee Joon Ki","searchWeight":0.9,"secondaryText":"actor","searchableHidden":false}}]}}}

查询存在的核心问题 #

• 精确匹配条件不匹配：should子句中的term查询值为"ли джун хек"，但目标条目ID23的primaryTitle是"ли джун хёк"（最后一个字符为ё而非е），term查询是完全精确匹配，因此该条件未命中ID23，导致它丢失了一个关键的高分匹配项。
• 权重优先级设置错误：primaryTitle的精确匹配未设置高权重，当前boost值为1.0，与其他字段的match查询权重一致，无法体现精确匹配的优先级。同时，ID23的searchWeight为0.9，而靠前的条目searchWeight为1.1，function_score中用field_value_factor乘以该权重，进一步拉低了ID23的最终分数。
• 模糊匹配逻辑过于宽松：match查询使用operator: OR，意味着查询词中的任意一个词匹配就会得分。ID99、100的字段中包含"ли джун"，加上较高的searchWeight，最终分数远高于因字符差异（ё和е）仅部分匹配的ID23。

内容的提问来源于stack exchange，提问作者user2572790