@jongleberry/vurst-html 0.0.1

package/index.d.ts

@@ -0,0 +1,68 @@
+/// <reference types="node" />
+
+export interface NapiSanitizeRssHtmlOptions {
+  proxyImages?: boolean
+  imageProxyUrlPrefix?: string
+  imageProxySigningKeys?: Array<string>
+}
+
+export interface SanitizeRssHtmlResult {
+  html: Buffer
+  firstImageSrc?: string
+}
+
+export interface ExtractDomRemovalsOptions {
+  boilerplatePatterns?: Array<string>
+}
+
+export interface ExtractDomRemovalsResult {
+  cssSelectorsToRemove: Array<string>
+  htmlToRemove: Array<string>
+}
+
+export interface CrawlerHtmlToMarkdownOptions {
+  cssSelectorsToRemove?: Array<string>
+  contentSelectors?: Array<string>
+  linkTextContentToRemove?: Array<string>
+  linkHrefsToRemove?: Array<string>
+  linkRelTokensToRemove?: Array<string>
+  useTextDensityFilter?: boolean
+}
+
+export interface CrawlerHtmlToMarkdownResult {
+  title?: string
+  meta: Record<string, unknown>
+  links: Record<string, unknown>
+  content: string
+  canonicalUrl?: string
+  lang?: string
+}
+
+export declare function sanitizeRssHtml(
+  html: Buffer,
+  options?: NapiSanitizeRssHtmlOptions,
+): Promise<SanitizeRssHtmlResult>
+
+export declare function sanitizeRssHtmlBatch(
+  inputs: Array<Buffer>,
+  options?: NapiSanitizeRssHtmlOptions,
+): Promise<Array<SanitizeRssHtmlResult>>
+
+export declare function htmlToEmbeddingText(html: Buffer): Promise<string>
+
+export declare function extractDomRemovals(
+  htmlPages: Array<Buffer>,
+  options?: ExtractDomRemovalsOptions,
+): Promise<ExtractDomRemovalsResult>
+
+export declare function applyDomRemovalsToHtml(
+  html: Buffer,
+  removals: ExtractDomRemovalsResult,
+): Promise<Buffer>
+
+export declare function getContentFromHtml(
+  htmlBuffer: Buffer,
+  options: CrawlerHtmlToMarkdownOptions,
+): Promise<CrawlerHtmlToMarkdownResult>
+
+export declare function sanitizePromptInjection(content: Buffer, isTitle?: boolean): Promise<Buffer>

package/index.js

@@ -0,0 +1,31 @@
+'use strict'
+
+// Platform dispatcher for @jongleberry/vurst-html.
+
+const { platform, arch } = process
+
+function loadBinding() {
+  switch (`${platform}-${arch}`) {
+    case 'darwin-arm64':
+      return require('./vurst-html.darwin-arm64.node')
+    case 'linux-x64':
+      return require('./vurst-html.linux-x64-gnu.node')
+    case 'linux-arm64':
+      return require('./vurst-html.linux-arm64-gnu.node')
+    default:
+      throw new Error(
+        `Unsupported platform: ${platform}-${arch}. ` +
+          `@jongleberry/vurst-html supports darwin-arm64, linux-x64 (glibc), and linux-arm64 (glibc).`,
+      )
+  }
+}
+
+const binding = loadBinding()
+
+module.exports.sanitizeRssHtml = binding.sanitizeRssHtml
+module.exports.sanitizeRssHtmlBatch = binding.sanitizeRssHtmlBatch
+module.exports.htmlToEmbeddingText = binding.htmlToEmbeddingText
+module.exports.extractDomRemovals = binding.extractDomRemovals
+module.exports.applyDomRemovalsToHtml = binding.applyDomRemovalsToHtml
+module.exports.getContentFromHtml = binding.getContentFromHtml
+module.exports.sanitizePromptInjection = binding.sanitizePromptInjection

package/package.json

@@ -0,0 +1,24 @@
+{
+  "name": "@jongleberry/vurst-html",
+  "version": "0.0.1",
+  "description": "Rust + N-API: HTML sanitization, extraction, and boilerstrip.",
+  "main": "index.js",
+  "types": "index.d.ts",
+  "license": "MIT",
+  "engines": { "node": ">= 18" },
+  "repository": { "type": "git", "url": "git+https://github.com/jonathanong/vurst.git" },
+  "keywords": ["rust", "napi", "html", "sanitize", "embedding", "boilerstrip"],
+  "files": ["index.js", "index.d.ts", "README.md", "*.node"],
+  "napi": {
+    "binaryName": "vurst-html",
+    "targets": ["aarch64-apple-darwin", "aarch64-unknown-linux-gnu", "x86_64-unknown-linux-gnu"]
+  },
+  "scripts": {
+    "build": "napi build --platform --release",
+    "build:debug": "napi build --platform",
+    "test": "node --test __test__/*.test.mjs",
+    "prepublishOnly": "napi prepublish -t npm"
+  },
+  "publishConfig": { "access": "public", "provenance": true },
+  "devDependencies": { "@napi-rs/cli": "^3.6.2", "@types/node": "^24" }
+}