@aiquants/html-to-markdown 0.1.0 → 0.1.2

package/dist/index.cjs

@@ -1 +1 @@
-"use strict";Object.defineProperty(exports,Symbol.toStringTag,{value:"Module"});const e=require("./core-D1CksXT5.cjs");exports.getMessage=e.getMessage,exports.htmlToMarkdown=e.htmlToMarkdown;
+const e=new(require("happy-dom").Window);Object.assign(global,{document:e.document,window:e,self:e}),Object.defineProperty(exports,Symbol.toStringTag,{value:"Module"});const o=require("./core-C3IjbMzg.cjs");exports.getMessage=o.getMessage,exports.htmlToMarkdown=o.htmlToMarkdown;

package/dist/index.js

@@ -1,4 +1,11 @@
-import { g, h } from "./core-C45e4gDg.js";
+import { Window } from "happy-dom";
+const window = new Window();
+Object.assign(global, {
+  document: window.document,
+  window,
+  self: window
+});
+import { g, h } from "./core-D07ieSmY.js";
 export {
   g as getMessage,
   h as htmlToMarkdown

package/dist/main.cjs

@@ -1,2 +1,2 @@
 #!/usr/bin/env node
-"use strict";Object.create,Object.defineProperty,Object.getOwnPropertyDescriptor,Object.getOwnPropertyNames,Object.getPrototypeOf,Object.prototype.hasOwnProperty;const e=require("./core-D1CksXT5.cjs");var t="undefined"!=typeof document?document.currentScript:null;(("undefined"==typeof document?require("url").pathToFileURL(__filename).href:t&&"SCRIPT"===t.tagName.toUpperCase()&&t.src||new URL("main.cjs",document.baseURI).href).endsWith(process.argv[1])||process.argv[1]&&process.argv[1].includes("vite-node"))&&(async()=>{const t=(await import("yargs-parser")).default,r=await import("node:fs"),c=await import("node:path"),o=t(process.argv.slice(2),{string:["locale"],default:{locale:"en-US"}}),s=o._[0],a=o.locale;s||process.exit(1);try{const t=await e.htmlToMarkdown(s,{locale:a}),o=`${s.replace(/https?:\/\//,"").replace(/[^a-zA-Z0-9]/g,"_")}_${Date.now()}.md`,n=c.join(process.cwd(),".outputs","raw");r.existsSync(n)||r.mkdirSync(n,{recursive:!0});const i=c.join(n,o);r.writeFileSync(i,t)}catch(n){process.exit(1)}})();
+Object.create,Object.defineProperty,Object.getOwnPropertyDescriptor,Object.getOwnPropertyNames,Object.getPrototypeOf,Object.prototype.hasOwnProperty;const e=new(require("happy-dom").Window);Object.assign(global,{document:e.document,window:e,self:e});const t=require("./core-C3IjbMzg.cjs");var r="undefined"!=typeof document?document.currentScript:null;(async()=>{(("undefined"==typeof document?require("url").pathToFileURL(__filename).href:r&&"SCRIPT"===r.tagName.toUpperCase()&&r.src||new URL("main.cjs",document.baseURI).href).endsWith(process.argv[1])||process.argv[1]&&process.argv[1].includes("vite-node"))&&(async()=>{const e=(await import("yargs-parser")).default,r=await import("node:fs"),c=await import("node:path"),o=e(process.argv.slice(2),{string:["locale"],default:{locale:"en-US"}}),s=o._[0],a=o.locale;s||process.exit(1);try{new URL(s)}catch(n){process.exit(1)}try{const e=await t.htmlToMarkdown(s,{locale:a}),o=`${s.replace(/https?:\/\//,"").replace(/[^a-zA-Z0-9]/g,"_")}_${Date.now()}.md`,n=c.join(process.cwd(),".outputs","raw");r.existsSync(n)||r.mkdirSync(n,{recursive:!0});const i=c.join(n,o);r.writeFileSync(i,e)}catch(n){process.exit(1)}})()})();

package/dist/main.js

@@ -1,5 +1,12 @@
 #!/usr/bin/env node
-import { g as getMessage, h as htmlToMarkdown } from "./core-C45e4gDg.js";
+import { Window } from "happy-dom";
+const window = new Window();
+Object.assign(global, {
+  document: window.document,
+  window,
+  self: window
+});
+import { g as getMessage, h as htmlToMarkdown } from "./core-D07ieSmY.js";
 const runCli = async () => {
   const yargsParser = (await import("yargs-parser")).default;
   const fs = await import("node:fs");
@@ -19,6 +26,12 @@ const runCli = async () => {
     console.info(getMessage(locale, "example"));
     process.exit(1);
   }
+  try {
+    new URL(targetUrl);
+  } catch (error) {
+    console.error(getMessage(locale, "error_app"), error);
+    process.exit(1);
+  }
   try {
     console.info(`
 ${getMessage(locale, "cli_start")}`);
@@ -43,11 +56,14 @@ ${getMessage(locale, "error_app")}`, error);
     process.exit(1);
   }
 };
-const isDirectRun = (
-  // `node dist/main.js` や `npx` 経由での実行を判定
-  import.meta.url.endsWith(process.argv[1]) || // `vite-node main.ts` 経由での実行を判定
-  process.argv[1] && process.argv[1].includes("vite-node")
-);
-if (isDirectRun) {
-  runCli();
-}
+const main = async () => {
+  const isDirectRun = (
+    // `node dist/main.js` や `npx` 経由での実行を判定
+    import.meta.url.endsWith(process.argv[1]) || // `vite-node main.ts` 経由での実行を判定
+    process.argv[1] && process.argv[1].includes("vite-node")
+  );
+  if (isDirectRun) {
+    runCli();
+  }
+};
+main();

package/dist/src/converter.d.ts

@@ -3,10 +3,12 @@
  * HTML から Markdown への変換モジュール
  */
 /**
- * HTML文字列をMarkdownに変換する
- * @param html HTMLコンテンツ
- * @param baseUrl リンクを絶対パスに変換するためのベースURL
- * @param locale ロケール
- * @returns Markdownテキスト
+ * Convert HTML string to Markdown format.
+ * HTML 文字列を Markdown 形式に変換する。
+ *
+ * @param html - The HTML content to convert / 変換する HTML コンテンツ
+ * @param baseUrl - Base URL for converting links to absolute paths / リンクを絶対パスに変換するためのベース URL
+ * @param locale - Locale setting for internationalization / 国際化のためのロケール設定
+ * @returns Promise resolving to Markdown text / Markdown テキストを返す Promise
  */
 export declare const convertHtmlToMarkdown: (html: string, baseUrl: string, locale: string) => Promise<string>;

package/dist/src/index.d.ts

@@ -3,5 +3,5 @@
  * htmlToMarkdown ライブラリのメインエクスポートモジュール
  */
 export { htmlToMarkdown } from './core.js';
-export type { HtmlToMarkdownOptions } from './types.js';
 export { getMessage } from './i18n.js';
+export type { HtmlToMarkdownOptions } from './types.js';

package/dist/src/plugins/index.d.ts

@@ -2,6 +2,6 @@
  * Rehype plugins index.
  * Rehype プラグインのインデックス
  */
-export { rehypeSanitizeHtml } from './rehype-sanitize-html.js';
 export { rehypeAbsoluteLinks } from './rehype-absolute-links.js';
+export { rehypeSanitizeHtml } from './rehype-sanitize-html.js';
 export { rehypeWikipediaFootnotes } from './rehype-wikipedia-footnotes.js';

package/dist/src/types.d.ts

@@ -1,3 +1,5 @@
+import { Element, Nodes, Parents } from 'hast';
+import { BlockContent as MdastBlockContent, DefinitionContent as MdastDefinitionContent, Nodes as MdastNodes, Parents as MdastParents, RootContent as MdastRootContent } from 'mdast';
 /**
  * Type definitions for htmlToMarkdown module.
  * htmlToMarkdown モジュールの型定義
@@ -17,9 +19,51 @@ export interface HtmlToMarkdownOptions {
  * Message keys for i18n.
  * i18n 用のメッセージキー
  */
-export type MessageKey = "error_url_required" | "usage" | "example" | "cli_start" | "cli_url" | "cli_locale" | "success_save" | "success_path" | "error_app" | "playwright_launch" | "playwright_goto" | "playwright_get_content" | "playwright_close" | "playwright_error" | "convert_start" | "convert_success";
+export type MessageKey = "error_url_required" | "usage" | "example" | "cli_start" | "cli_url" | "cli_locale" | "success_save" | "success_path" | "error_app" | "playwright_launch" | "playwright_goto" | "playwright_get_content" | "playwright_close" | "playwright_error" | "convert_start" | "convert_success" | "url_parse_error";
 /**
  * Supported locales.
  * サポートされているロケール
  */
 export type SupportedLocale = "ja-JP" | "en-US";
+/**
+ * The following types are based on the type definitions of `hast-util-to-mdast`.
+ * `hast-util-to-mdast` の型定義に基づいた型。
+ *
+ * @see https://github.com/syntax-tree/hast-util-to-mdast/blob/main/lib/state.js
+ */
+export type State = {
+    all: All;
+    baseFound: boolean;
+    elementById: Map<string, Element>;
+    frozenBaseUrl: string | undefined;
+    handlers: Record<string, Handle>;
+    inTable: boolean;
+    nodeHandlers: Record<string, NodeHandle>;
+    one: One;
+    options: Options;
+    patch: Patch;
+    qNesting: number;
+    resolve: Resolve;
+    toFlow: ToFlow;
+    toSpecificContent: ToSpecificContent;
+};
+export type MdastFlowContent = MdastBlockContent | MdastDefinitionContent;
+export type All = (parent: Parents) => MdastRootContent[];
+export type Handle = (state: State, element: Element, parent: Parents | undefined) => Array<MdastNodes> | MdastNodes | undefined | void;
+export type NodeHandle = (state: State, node: Nodes, parent: Parents | undefined) => Array<MdastNodes> | MdastNodes | undefined | void;
+export type One = (node: Nodes, parent: Parents | undefined) => Array<MdastNodes> | MdastNodes | undefined;
+export type Options = {
+    checked?: string | null;
+    document?: boolean | null;
+    handlers?: Record<string, Handle | null | undefined> | null;
+    newlines?: boolean | null;
+    nodeHandlers?: Record<string, NodeHandle | null | undefined> | null;
+    quotes?: Array<string> | null;
+    unchecked?: string | null;
+};
+export type Patch = (from: Nodes, to: MdastNodes) => undefined;
+export type Resolve = (url: string | null | undefined) => string;
+export type ToFlow = (nodes: MdastRootContent[]) => MdastFlowContent[];
+export type ToSpecificContent = <ChildType extends MdastNodes, ParentType extends MdastParents & {
+    children: Array<ChildType>;
+}>(nodes: MdastRootContent[], build: () => ParentType) => ParentType[];

package/dist/tests/browser-100.test.d.ts

@@ -0,0 +1 @@
+export {};

package/dist/tests/browser-coverage.test.d.ts

@@ -0,0 +1 @@
+export {};

package/dist/tests/converter-coverage.test.d.ts

@@ -0,0 +1 @@
+export {};

package/dist/tests/readme-usage.test.d.ts

@@ -0,0 +1 @@
+export {};

package/dist/tests/rehype-absolute-links-coverage.test.d.ts

@@ -0,0 +1 @@
+export {};

package/dist/tests/rehype-sanitize-html-coverage.test.d.ts

@@ -0,0 +1 @@
+export {};

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@aiquants/html-to-markdown",
-  "version": "0.1.0",
+  "version": "0.1.2",
   "description": "HTML to Markdown converter",
   "type": "module",
   "main": "./dist/index.js",
@@ -51,7 +51,8 @@
     "remark-stringify": "^11.0.0",
     "unified": "^11.0.5",
     "unist-util-visit": "^5.0.0",
-    "yargs-parser": "^22.0.0"
+    "yargs-parser": "^22.0.0",
+    "happy-dom": "^18.0.1"
   },
   "devDependencies": {
     "@biomejs/biome": "^2.1.3",
@@ -61,7 +62,6 @@
     "@types/node": "^22.16.5",
     "@types/yargs-parser": "^21.0.3",
     "@vitest/coverage-v8": "^3.2.4",
-    "happy-dom": "^18.0.1",
     "rehype-stringify": "^10.0.1",
     "terser": "^5.43.1",
     "typescript": "^5.8.3",