2 Commits
0.3.2 ... 0.3.4

Author SHA1 Message Date
57cb9c8bd1 Fixed delete by href
All checks were successful
Publish Library / Build NPM Project (push) Successful in 44s
Publish Library / Tag Version (push) Successful in 26s
2026-08-25 01:00:28 -04:00
7a2f17bfe9 Use kiwix search completely, no fuzzy ranking
All checks were successful
Publish Library / Build NPM Project (push) Successful in 16s
Publish Library / Tag Version (push) Successful in 10s
2026-08-25 00:42:59 -04:00
4 changed files with 12 additions and 28 deletions

View File

@@ -1,6 +1,6 @@
{ {
"name": "@ztimson/zim-utils", "name": "@ztimson/zim-utils",
"version": "0.3.2", "version": "0.3.4",
"description": "Native, dependency-light ZIM archive reader/searcher and Kiwix catalog downloader for Node.js", "description": "Native, dependency-light ZIM archive reader/searcher and Kiwix catalog downloader for Node.js",
"author": "Zak Timson", "author": "Zak Timson",
"license": "MIT", "license": "MIT",

View File

@@ -79,9 +79,10 @@ export class ZimManager {
async delete(nameOrFile) { async delete(nameOrFile) {
const local = await this.list(); const local = await this.list();
const match = local.find(l => l.name === nameOrFile || l.file === nameOrFile); const match = local.find(l => l.name === nameOrFile || l.href === nameOrFile);
if (!match) throw new Error(`ZIM not found locally: ${nameOrFile}`); if (!match) throw new Error(`ZIM not found locally: ${nameOrFile}`);
await fs.promises.rm(path.join(this.#dir, match.file), {force: true}); const file = match.href + (match.href.endsWith('.zim') ? '' : '.zim');
await fs.promises.rm(path.join(this.#dir, file), {force: true});
const server = await this.#ensureServer(); const server = await this.#ensureServer();
await server.reload(); await server.reload();
return {file: match.file, status: 'deleted'}; return {file: match.file, status: 'deleted'};

View File

@@ -6,8 +6,7 @@ import net from 'node:net';
import fs from 'node:fs'; import fs from 'node:fs';
import path from 'node:path'; import path from 'node:path';
import {fileURLToPath} from 'node:url'; import {fileURLToPath} from 'node:url';
import {decodeHtml, fromXml} from '@ztimson/utils'; import {fromXml} from '@ztimson/utils';
import {fuzzyMatch, weightedScore} from './utils.js';
const execFileAsync = promisify(execFile); const execFileAsync = promisify(execFile);
const __dirname = path.dirname(fileURLToPath(import.meta.url)); const __dirname = path.dirname(fileURLToPath(import.meta.url));
@@ -180,28 +179,21 @@ export class KiwixServer {
return {mimetype: res.headers.get('content-type'), data: Buffer.from(await res.arrayBuffer())}; return {mimetype: res.headers.get('content-type'), data: Buffer.from(await res.arrayBuffer())};
} }
/** /** Fulltext search across every local ZIM via kiwix-serve's own xapian index */
* Two-pass fulltext search across every local ZIM: xapian prefilter (kiwix-serve's
* own index), then fuzzy-reranked by title so the strongest matches surface first.
* Returns a flat array matching the catalog/list shape: {id, title, name, category, ..., href, score}
*/
async search(terms, limit = 20) { async search(terms, limit = 20) {
this.#assertRunning(); this.#assertRunning();
const termList = String(terms).split(',').map(t => t.trim()).filter(Boolean); const termList = String(terms).split(/[,\s]+/).map(t => t.trim()).filter(Boolean);
if (!termList.length) return []; if (!termList.length) return [];
const params = new URLSearchParams({pattern: termList.join(' '), format: 'xml', pageLength: String(limit)}); const params = new URLSearchParams({pattern: termList.join(' '), format: 'xml', pageLength: String(limit)});
const res = await fetch(`${this.baseUrl}/search?${params}`); const res = await fetch(`${this.baseUrl}/search?${params}`);
if (!res.ok) return []; if (!res.ok) return [];
const xml = await res.text(); const found = fromXml(await res.text())?.rss?.channel?.item || [];
let found = fromXml(xml);
found = found?.rss?.channel?.item || [];
const books = await this.list(); const books = await this.list();
const bookMap = new Map(books.map(b => [b.title, b])); const bookMap = new Map(books.map(b => [b.title, b]));
const enriched = found.map(hit => { return found.map(hit => {
const book = bookMap.get(hit.book.title); const book = bookMap.get(hit.book.title);
if (!book) return null; if (!book) return null;
const prefix = `/content/${book.href}/`; const prefix = `/content/${book.href}/`;
@@ -216,10 +208,8 @@ export class KiwixServer {
icon: book.icon, icon: book.icon,
viewer: this.baseUrl + hit.link, viewer: this.baseUrl + hit.link,
summary: hit.description, summary: hit.description,
score: weightedScore(hit.title, termList) + weightedScore(hit.description, termList), score: +hit.score || 0,
}; };
}).filter(hit => !!hit && hit.score > 0); }).filter(Boolean);
return enriched.toSorted((a, b) => b.score - a.score).slice(0, limit);
} }
} }

View File

@@ -40,10 +40,3 @@ export function fuzzyMatch(target, ...terms) {
similarities, similarities,
}; };
} }
export function weightedScore(text, termList) {
if (!text) return 0;
const covered = termList.reduce((sum, t) => sum + t.length, 0);
const coverage = Math.min(1, covered / text.length);
return fuzzyMatch(text, ...termList).max * coverage;
}