my-project-public

repository

loading code, commits, and activity

repositories

loading repo index

my-project-public — gitlawb

#1	export class BM25 {
#2	private documents: string[][];
#3	private k1: number;
#4	private b: number;
#5	private avgDocLength: number;
#6	private docFreq: Map<string, number>;
#7	private docLengths: number[];
#8	private idf: Map<string, number>;
#9
#10	constructor(documents: string[][], k1 = 1.5, b = 0.75) {
#11	this.documents = documents;
#12	this.k1 = k1;
#13	this.b = b;
#14	this.docLengths = documents.map((doc) => doc.length);
#15	this.avgDocLength =
#16	this.docLengths.reduce((a, b) => a + b, 0) / documents.length;
#17	this.docFreq = new Map();
#18	this.idf = new Map();
#19	this.computeIdf();
#20	}
#21
#22	private computeIdf() {
#23	const N = this.documents.length;
#24
#25	// Count document frequency for each term
#26	for (const doc of this.documents) {
#27	const terms = new Set(doc);
#28	for (const term of terms) {
#29	this.docFreq.set(term, (this.docFreq.get(term) \|\| 0) + 1);
#30	}
#31	}
#32
#33	// Compute IDF for each term
#34	for (const [term, freq] of this.docFreq) {
#35	this.idf.set(term, Math.log((N - freq + 0.5) / (freq + 0.5) + 1));
#36	}
#37	}
#38
#39	private score(query: string[], doc: string[], index: number): number {
#40	let score = 0;
#41	const docLength = this.docLengths[index];
#42
#43	for (const term of query) {
#44	const tf = doc.filter((t) => t === term).length;
#45	const idf = this.idf.get(term) \|\| 0;
#46
#47	score +=
#48	(idf * tf * (this.k1 + 1)) /
#49	(tf +
#50	this.k1 * (1 - this.b + (this.b * docLength) / this.avgDocLength));
#51	}
#52
#53	return score;
#54	}
#55
#56	search(query: string[]): string[][] {
#57	const scores = this.documents.map((doc, idx) => ({
#58	doc,
#59	score: this.score(query, doc, idx),
#60	}));
#61
#62	return scores.sort((a, b) => b.score - a.score).map((item) => item.doc);
#63	}
#64	}
#65

z6Mkq5mY3JWtxoxUobWcfNHm7AkRubgSWEZTkBVqZXJviFZ5/my-project-public