Development build. This page describes main, not a released package. The latest published Lodestar.Text is 0.6.0 — read its documentation.

Home › Text › Keyword search

The best documents for a query, best first.

public IReadOnlyList<SearchHit> Top(IEnumerable<int> queryTerms, int count)

Parameters — queryTerms are column indices, as Score takes them. count is how many hits to return; fewer come back when the corpus is smaller.

Returns — IReadOnlyList<SearchHit>, descending by score.

Exceptions — ArgumentNullException when queryTerms is null; ArgumentOutOfRangeException when count is negative.

Example — the two best documents for a rare term.

using Lodestar.Abstractions;
using Lodestar.Text.Search;
using Lodestar.Text.Vectorization;

var vectorizer = new CountVectorizer();
CsrMatrix counts = vectorizer.FitTransform(
    ["the cat sat", "the dog sat sat", "a bird flew far away today"]);
int cat = vectorizer.GetFeatureNames().ToList().IndexOf("cat");

IReadOnlyList<SearchHit> hits = new Bm25Index(counts).Top([cat], 2);

int best = hits[0].Document;  // => 0
double score = hits[0].Score;  // => 0.5755…
int howMany = hits.Count;  // => 2

Remarks — the same scores Score returns, ordered; it is not a second scoring path.

Ties break by document index, ascending. A query that separates nothing therefore returns the corpus in its own order rather than in whatever order the sort happened to produce, which is what makes two runs of the same query comparable.

Documents scoring zero come back like any other. Asking for ten hits from a corpus of ten gives ten, however few actually matched — filter on the score if only matches are wanted. That is a deliberate choice: a cutoff belongs to the caller, who knows what the scores mean in their corpus, and this type does not.

Applies to — net10.0, netstandard2.0.

See also — Bm25Index, SearchHit.