ApproximateSearchConfig

class lightning_ir.retrieve.base.searcher.ApproximateSearchConfig(k: int = 10, candidate_k: int = 100, imputation_strategy: 'min' | 'gather' | 'zero' = 'gather')

Bases: SearchConfig

Configuration class for approximate searchers in the Lightning IR framework.

__init__(k: int = 10, candidate_k: int = 100, imputation_strategy: 'min' | 'gather' | 'zero' = 'gather') → None

Initialize the ApproximateSearchConfig.

Parameters:

• k (int) – The number of top documents to retrieve. Defaults to 10.

• candidate_k (int) – The number of candidate documents to consider for scoring. Defaults to 100.

• imputation_strategy (Literal["min", "gather", "zero"]) – Strategy for imputing missing scores. Defaults to “gather”.

Methods

__init__([k, candidate_k, imputation_strategy])

Initialize the ApproximateSearchConfig.

Attributes

SUPPORTED_MODELS

search_class

alias of ApproximateSearcher