npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@llangtop/pwiki-core

v1.3.9

Published

Wiki knowledge base engine — keyword/semantic/hybrid search with local ONNX embeddings

Downloads

1,605

Readme

@llangtop/pwiki-core

Wiki 知识库搜索引擎。供 pwiki-clipwiki-mcp 调用,也可作为库直接引用。

安装

需要 Node.js 22 或更高版本。

npm i @llangtop/pwiki-core

用法

import { WikiEngine } from "@llangtop/pwiki-core";

const engine = new WikiEngine({ basePath: "~/.pwiki" });

// 加载数据源
engine.addSource("/path/to/notes");
await engine.loadSource("/path/to/notes");

// 搜索(keyword | semantic | hybrid)
const hits = await engine.search("部署流程", "hybrid");

// 限定到物理数据源分片和源内路径。无结果时不会回退全局。
const [source] = engine.sourceRefs;
const scopedHits = await engine.search("部署流程", "hybrid", {
  source: source.id,
  pathPrefix: "技术",
});

// 搜索结果携带 sourceId;读取时继续传入,避免同名相对路径歧义。
const { entry, content } = engine.readEntry(
  scopedHits[0].relPath,
  scopedHits[0].sourceId,
);

Dedicated knowledge areas

WikiEngine also maintains dedicated Markdown files under <WIKI_HOME>/areas/<areaId>.md. These files are domain-context entry points, not loaded sources: writing one does not update BM25, embeddings, manifests, or run an internal query loop.

const area = engine.writeKnowledgeArea(
  "ami/dlms",
  "# AMI / DLMS\n\n## Content map\n\n## Special knowledge\n",
);

console.log(area.id, area.contentHash);
console.log(engine.readKnowledgeArea("ami/dlms")?.content);
console.log(engine.listKnowledgeAreas());

The public mutation is complete replacement semantics. Callers should read the returned contentHash or read the file back after an external write. Area IDs are relative, may contain nested segments, and reject absolute paths, parent segments, and symlinked directories.

Optional Cross-Encoder reranking

engine.search(query, "hybrid") can optionally rerank the final global RRF candidates. It is disabled by default and has no model-loading path until config.json contains reranker.enabled: true. Configure it through setRerankerConfig() or the CLI:

import { setRerankerConfig } from "@llangtop/pwiki-core";

setRerankerConfig({ enabled: true, dtype: "int8", inputTopK: 20, outputTopK: 10 });

SearchHit.score and semanticScore remain untouched; reranked hits add rerankerScore and originalRank. Errors in this optional layer return the original Hybrid/RRF ranking.

CRUD 默认在返回前完成单文件 BM25 增量更新,并等待该文件的向量更新,保持一次性 CLI/脚本的原有完成语义。常驻服务可启用后台向量队列:

CRUD 写操作使用已加载的数据源路径;未注册的数据源不会被隐式创建或写入。 写入路径是 source-relative Markdown 路径,省略 .md 时自动补全,大小写变体会 规范化为小写 .md。省略 source selector 的单文件 rename/move/delete 仍兼容, 但当多个 source 命中同名路径时会拒绝操作,要求显式传入 source。

const engine = new WikiEngine({
  basePath: "~/.pwiki",
  backgroundEmbeddings: true,
});

await engine.modifyEntry(sourcePath, "docs/api.md", newContent);
// 此时 BM25 已可检索;向量任务在后台按文件合并执行。
console.log(engine.backgroundVectorStatus());
await engine.waitForBackgroundTasks(); // 关闭常驻进程前可显式排空

每个数据源的文件索引、BM25、向量和 manifest 分别存放在 <WIKI_HOME>/sources/<sourceId>/。指定 source 后只打开对应分片; pathPrefix 会在 BM25 评分和向量相似度计算之前过滤候选。 BM25 使用标准未加密的 bm25.sqlite3;v3 JSON 快照会在首次访问时自动导入并保留。

主要导出

| 导出 | 类型 | 说明 | |------|------|------| | WikiEngine | class | 唯一对外 API | | SearchScope | type | 数据源和源内路径检索边界 | | SourceRef | type | 稳定数据源 ID、名称和路径 | | keywordSearch | function | 关键词搜索 | | semanticSearch | function | 语义搜索 | | hybridSearch | function | RRF 混合搜索 | | embed | function | 文本 → 向量 | | cosineSimilarity | function | 余弦相似度 | | MODELS | array | 内置嵌入模型列表 |