SeismicIndexConfig

class lightning_ir.retrieve.seismic.seismic_indexer.SeismicIndexConfig(num_postings: int = 3500, centroid_fraction: float = 0.1, min_cluster_size: int = 2, summary_energy: float = 0.4, num_k_nearest_neighbors: int = 0, batch_size: int | None = None, num_threads: int = 0)[source]

Bases: IndexConfig

Configuration for the Seismic indexer.

__init__(num_postings: int = 3500, centroid_fraction: float = 0.1, min_cluster_size: int = 2, summary_energy: float = 0.4, num_k_nearest_neighbors: int = 0, batch_size: int | None = None, num_threads: int = 0) None[source]

Initialize the SeismicIndexConfig.

Parameters:
  • num_postings (int) – Number of postings to keep in the index. Defaults to 3500.

  • centroid_fraction (float) – Fraction of centroids to keep. Defaults to 0.1.

  • min_cluster_size (int) – Minimum size of clusters. Defaults to 2.

  • summary_energy (float) – Energy threshold for summaries. Defaults to 0.4.

  • num_k_nearest_neighbors (int) – Number of nearest neighbors to consider. Defaults to 0.

  • batch_size (int | None) – Batch size for indexing. Defaults to None.

  • num_threads (int) – Number of threads to use for indexing. Defaults to 0.

Methods

__init__([num_postings, centroid_fraction, ...])

Initialize the SeismicIndexConfig.

Attributes

SUPPORTED_MODELS

indexer_class

alias of SeismicIndexer