langchain_community.vectorstores.vectara.VectaraQueryConfig

class langchain_community.vectorstores.vectara.VectaraQueryConfig(k: int = 10, lambda_val: float = 0.0, filter: str = '', score_threshold: ~typing.Optional[float] = None, n_sentence_context: int = 2, mmr_config: ~langchain_community.vectorstores.vectara.MMRConfig = <factory>, summary_config: ~langchain_community.vectorstores.vectara.SummaryConfig = <factory>)[source]

用于Vectara查询的配置。

k: 要返回的文档数量。默认为10。 lambda_val: 用于混合搜索的词汇匹配参数。 filter: 元数据过滤参数的字典。例如，过滤器可以是 “doc.rating > 3.0 and part.lang = ‘deu’”，更多详情请参见https://docs.vectara.com/docs/search-apis/sql/filter-overview。 score_threshold: 结果的最小得分阈值。如果定义了，得分低于此值的结果将被过滤掉。 n_sentence_context: 匹配段前/后要添加的句子数量，默认为2。 mmr_config: MMRConfig配置数据类 summary_config: SummaryConfig配置数据类

Attributes

filter

k

lambda_val

n_sentence_context

score_threshold

mmr_config

summary_config

Methods

__init__([k, lambda_val, filter, ...])

Parameters

  • k (int) –

  • lambda_val (float) –

  • filter (str) –

  • score_threshold (Optional[float]) –

  • n_sentence_context (int) –

  • mmr_config (MMRConfig) –

  • summary_config (SummaryConfig) –

Return type

None

__init__(k: int = 10, lambda_val: float = 0.0, filter: str = '', score_threshold: ~typing.Optional[float] = None, n_sentence_context: int = 2, mmr_config: ~langchain_community.vectorstores.vectara.MMRConfig = <factory>, summary_config: ~langchain_community.vectorstores.vectara.SummaryConfig = <factory>) → None
Parameters

  • k (int) –

  • lambda_val (float) –

  • filter (str) –

  • score_threshold (Optional[float]) –

  • n_sentence_context (int) –

  • mmr_config (MMRConfig) –

  • summary_config (SummaryConfig) –

Return type

None