@openuji/speculator-search

Build-time search indexing for serverless, instant navigation.

speculator-search extracts searchable content from Speculator documents and generates static JSON indexes. This allows spec readers to find what they need instantly, without requiring a backend search server.

🔍 Features

• Full-Text Indexing: Extracts text from paragraphs, headings, and definitions.
• Context-Aware Results: Search results include section titles and node types for better relevance.
• Zero Runtime Dependencies: The generated index is pure JSON, compatible with any client-side search library (like FlexSearch).
• Hierarchical IDs: Automatically generates stable navigation anchors even if the source document lacks IDs.

🚀 Usage

1. Build the Index (Node.js)

import { buildSearchIndex } from "@openuji/speculator-search";

const searchIndex = buildSearchIndex(workspace);
await fs.writeFile("public/search-index.json", JSON.stringify(searchIndex));

2. Search (Browser)

const response = await fetch("/search-index.json");
const index = await response.json();

// Direct search across all documents
const results = index.documents
  .flatMap((doc) => doc.entries)
  .filter((e) => e.plainText.includes("your query"));

Part of the Speculator ecosystem.