@hamster-note/document-parser 0.1.0 → 0.2.0

package/dist/Parser/index.d.ts

@@ -0,0 +1,19 @@
+import type { IntermediateDocument } from '@hamster-note/types';
+export type ParserInput = ArrayBuffer | ArrayBufferView | Blob;
+export declare abstract class DocumentParser {
+    static readonly exts: readonly string[];
+    /**
+     * 将原始输入解析为中间文档结构。
+     * - 输入必须是内存态数据（ArrayBuffer / ArrayBufferView / Blob）
+     * - 失败时抛出异常；不要返回 undefined。
+     */
+    abstract encode(input: ParserInput): Promise<IntermediateDocument>;
+    /**
+     * 可选：从中间文档逆序列化回原始文件数据。
+     * 默认不支持，需要的子类自行实现。
+     */
+    decode(_intermediateDocument: IntermediateDocument): Promise<ParserInput>;
+    protected static toArrayBuffer(input: ParserInput): Promise<ArrayBuffer>;
+    protected static toUint8Array(input: ParserInput): Promise<Uint8Array>;
+}
+//# sourceMappingURL=index.d.ts.map

package/dist/Parser/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/Parser/index.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,oBAAoB,EAAE,MAAM,qBAAqB,CAAA;AAG/D,MAAM,MAAM,WAAW,GAAG,WAAW,GAAG,eAAe,GAAG,IAAI,CAAA;AAE9D,8BAAsB,cAAc;IAElC,MAAM,CAAC,QAAQ,CAAC,IAAI,EAAE,SAAS,MAAM,EAAE,CAAK;IAE5C;;;;OAIG;IACH,QAAQ,CAAC,MAAM,CAAC,KAAK,EAAE,WAAW,GAAG,OAAO,CAAC,oBAAoB,CAAC;IAElE;;;OAGG;IACH,MAAM,CAAC,qBAAqB,EAAE,oBAAoB,GAAG,OAAO,CAAC,WAAW,CAAC;qBAKlD,aAAa,CAClC,KAAK,EAAE,WAAW,GACjB,OAAO,CAAC,WAAW,CAAC;qBAmBA,YAAY,CAAC,KAAK,EAAE,WAAW,GAAG,OAAO,CAAC,UAAU,CAAC;CAI7E"}

package/dist/Parser/index.js

@@ -0,0 +1,38 @@
+export class DocumentParser {
+    /**
+     * 可选：从中间文档逆序列化回原始文件数据。
+     * 默认不支持，需要的子类自行实现。
+     */
+    decode(_intermediateDocument) {
+        return Promise.reject(new Error('decode is not implemented'));
+    }
+    // 将各种输入转换为 ArrayBuffer，方便子类直接消费二进制
+    static async toArrayBuffer(input) {
+        if (input instanceof ArrayBuffer) {
+            return input;
+        }
+        if (ArrayBuffer.isView(input)) {
+            // 拷贝为独立的 ArrayBuffer，避免 SharedArrayBuffer 兼容性问题
+            const buffer = new ArrayBuffer(input.byteLength);
+            new Uint8Array(buffer).set(new Uint8Array(input.buffer, input.byteOffset, input.byteLength));
+            return buffer;
+        }
+        if (typeof Blob !== 'undefined' && input instanceof Blob) {
+            return input.arrayBuffer();
+        }
+        throw new TypeError('Unsupported parser input');
+    }
+    // 子类若需要 Uint8Array，可直接复用
+    static async toUint8Array(input) {
+        const buffer = await DocumentParser.toArrayBuffer(input);
+        return new Uint8Array(buffer);
+    }
+}
+// 子类必须声明支持的扩展名列表（不带点），用于注册分发
+Object.defineProperty(DocumentParser, "exts", {
+    enumerable: true,
+    configurable: true,
+    writable: true,
+    value: []
+});
+//# sourceMappingURL=index.js.map

package/dist/Parser/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/Parser/index.ts"],"names":[],"mappings":"AAKA,MAAM,OAAgB,cAAc;IAWlC;;;OAGG;IACH,MAAM,CAAC,qBAA2C;QAChD,OAAO,OAAO,CAAC,MAAM,CAAC,IAAI,KAAK,CAAC,2BAA2B,CAAC,CAAC,CAAA;IAC/D,CAAC;IAED,mCAAmC;IACzB,MAAM,CAAC,KAAK,CAAC,aAAa,CAClC,KAAkB;QAElB,IAAI,KAAK,YAAY,WAAW,EAAE,CAAC;YACjC,OAAO,KAAK,CAAA;QACd,CAAC;QACD,IAAI,WAAW,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC;YAC9B,gDAAgD;YAChD,MAAM,MAAM,GAAG,IAAI,WAAW,CAAC,KAAK,CAAC,UAAU,CAAC,CAAA;YAChD,IAAI,UAAU,CAAC,MAAM,CAAC,CAAC,GAAG,CACxB,IAAI,UAAU,CAAC,KAAK,CAAC,MAAM,EAAE,KAAK,CAAC,UAAU,EAAE,KAAK,CAAC,UAAU,CAAC,CACjE,CAAA;YACD,OAAO,MAAM,CAAA;QACf,CAAC;QACD,IAAI,OAAO,IAAI,KAAK,WAAW,IAAI,KAAK,YAAY,IAAI,EAAE,CAAC;YACzD,OAAO,KAAK,CAAC,WAAW,EAAE,CAAA;QAC5B,CAAC;QACD,MAAM,IAAI,SAAS,CAAC,0BAA0B,CAAC,CAAA;IACjD,CAAC;IAED,yBAAyB;IACf,MAAM,CAAC,KAAK,CAAC,YAAY,CAAC,KAAkB;QACpD,MAAM,MAAM,GAAG,MAAM,cAAc,CAAC,aAAa,CAAC,KAAK,CAAC,CAAA;QACxD,OAAO,IAAI,UAAU,CAAC,MAAM,CAAC,CAAA;IAC/B,CAAC;;AA3CD,6BAA6B;AACb;;;;WAA0B,EAAE;GAAA"}

package/dist/index.d.ts

@@ -0,0 +1,4 @@
+export * from './types/DocumentAnchor';
+export * from './Parser';
+export * from './register';
+//# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,cAAc,wBAAwB,CAAA;AACtC,cAAc,UAAU,CAAA;AACxB,cAAc,YAAY,CAAA"}

package/dist/index.js

@@ -0,0 +1,4 @@
+export * from './types/DocumentAnchor';
+export * from './Parser';
+export * from './register';
+//# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,cAAc,wBAAwB,CAAA;AACtC,cAAc,UAAU,CAAA;AACxB,cAAc,YAAY,CAAA"}

package/dist/register/index.d.ts

@@ -0,0 +1,20 @@
+import type { DocumentParser, ParserInput } from '../Parser';
+import type { IntermediateDocument } from '@hamster-note/types';
+export interface DocumentParserConstructor<T extends DocumentParser = DocumentParser> {
+    new (): T;
+    readonly exts: readonly string[];
+}
+export declare const registerParser: (Parser: DocumentParserConstructor) => void;
+export declare function unregisterParser(Parser: DocumentParserConstructor): void;
+export declare function getParserByExt(ext?: string): DocumentParserConstructor | undefined;
+export interface ParseOptions {
+    filename?: string;
+    ext?: string;
+    parser?: DocumentParserConstructor;
+}
+export declare function parse(input: ParserInput, options?: ParseOptions): Promise<IntermediateDocument>;
+export declare function listRegisteredParsers(): {
+    ext: string;
+    parser: DocumentParserConstructor<DocumentParser>;
+}[];
+//# sourceMappingURL=index.d.ts.map

package/dist/register/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/register/index.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,cAAc,EAAE,WAAW,EAAE,MAAM,WAAW,CAAA;AAC5D,OAAO,KAAK,EAAE,oBAAoB,EAAE,MAAM,qBAAqB,CAAA;AAE/D,MAAM,WAAW,yBAAyB,CACxC,CAAC,SAAS,cAAc,GAAG,cAAc;IAEzC,QAAQ,CAAC,CAAA;IACT,QAAQ,CAAC,IAAI,EAAE,SAAS,MAAM,EAAE,CAAA;CACjC;AAOD,eAAO,MAAM,cAAc,GAAI,QAAQ,yBAAyB,SAY/D,CAAA;AAED,wBAAgB,gBAAgB,CAAC,MAAM,EAAE,yBAAyB,QAQjE;AAED,wBAAgB,cAAc,CAC5B,GAAG,CAAC,EAAE,MAAM,GACX,yBAAyB,GAAG,SAAS,CAGvC;AAED,MAAM,WAAW,YAAY;IAC3B,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,GAAG,CAAC,EAAE,MAAM,CAAA;IACZ,MAAM,CAAC,EAAE,yBAAyB,CAAA;CACnC;AAED,wBAAsB,KAAK,CACzB,KAAK,EAAE,WAAW,EAClB,OAAO,GAAE,YAAiB,GACzB,OAAO,CAAC,oBAAoB,CAAC,CAe/B;AAGD,wBAAgB,qBAAqB;;;IAEpC"}

package/dist/register/index.js

@@ -0,0 +1,48 @@
+const parserMap = new Map();
+const normalizeExt = (ext) => ext.replace(/^\./, '').toLowerCase();
+// 注册解析器：防止重复注册同扩展名的不同类
+export const registerParser = (Parser) => {
+    if (!Array.isArray(Parser.exts) || Parser.exts.length === 0) {
+        throw new Error('Parser.exts 必须是非空数组');
+    }
+    Parser.exts.forEach((ext) => {
+        const key = normalizeExt(ext);
+        const exists = parserMap.get(key);
+        if (exists && exists !== Parser) {
+            throw new Error(`扩展名 ${key} 已被其他解析器占用`);
+        }
+        parserMap.set(key, Parser);
+    });
+};
+export function unregisterParser(Parser) {
+    Parser.exts.forEach((ext) => {
+        const key = normalizeExt(ext);
+        const exists = parserMap.get(key);
+        if (exists === Parser) {
+            parserMap.delete(key);
+        }
+    });
+}
+export function getParserByExt(ext) {
+    if (!ext)
+        return undefined;
+    return parserMap.get(normalizeExt(ext));
+}
+export async function parse(input, options = {}) {
+    var _a, _b, _c;
+    const extFromName = (_a = options.filename) === null || _a === void 0 ? void 0 : _a.split('.').pop();
+    const ext = (_b = options.ext) !== null && _b !== void 0 ? _b : extFromName;
+    const Parser = (_c = options.parser) !== null && _c !== void 0 ? _c : getParserByExt(ext);
+    if (!Parser) {
+        throw new Error(`未找到匹配的解析器，扩展名: ${ext !== null && ext !== void 0 ? ext : '未知'}，已注册: ${[
+            ...parserMap.keys()
+        ].join(', ')}`);
+    }
+    const parser = new Parser();
+    return parser.encode(input);
+}
+// 便于测试或运行时可观测性
+export function listRegisteredParsers() {
+    return [...parserMap.entries()].map(([ext, ctor]) => ({ ext, parser: ctor }));
+}
+//# sourceMappingURL=index.js.map

package/dist/register/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/register/index.ts"],"names":[],"mappings":"AAUA,MAAM,SAAS,GAA2C,IAAI,GAAG,EAAE,CAAA;AAEnE,MAAM,YAAY,GAAG,CAAC,GAAW,EAAE,EAAE,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,WAAW,EAAE,CAAA;AAE1E,uBAAuB;AACvB,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,MAAiC,EAAE,EAAE;IAClE,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,MAAM,CAAC,IAAI,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC5D,MAAM,IAAI,KAAK,CAAC,qBAAqB,CAAC,CAAA;IACxC,CAAC;IACD,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE;QAC1B,MAAM,GAAG,GAAG,YAAY,CAAC,GAAG,CAAC,CAAA;QAC7B,MAAM,MAAM,GAAG,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;QACjC,IAAI,MAAM,IAAI,MAAM,KAAK,MAAM,EAAE,CAAC;YAChC,MAAM,IAAI,KAAK,CAAC,OAAO,GAAG,YAAY,CAAC,CAAA;QACzC,CAAC;QACD,SAAS,CAAC,GAAG,CAAC,GAAG,EAAE,MAAM,CAAC,CAAA;IAC5B,CAAC,CAAC,CAAA;AACJ,CAAC,CAAA;AAED,MAAM,UAAU,gBAAgB,CAAC,MAAiC;IAChE,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE;QAC1B,MAAM,GAAG,GAAG,YAAY,CAAC,GAAG,CAAC,CAAA;QAC7B,MAAM,MAAM,GAAG,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;QACjC,IAAI,MAAM,KAAK,MAAM,EAAE,CAAC;YACtB,SAAS,CAAC,MAAM,CAAC,GAAG,CAAC,CAAA;QACvB,CAAC;IACH,CAAC,CAAC,CAAA;AACJ,CAAC;AAED,MAAM,UAAU,cAAc,CAC5B,GAAY;IAEZ,IAAI,CAAC,GAAG;QAAE,OAAO,SAAS,CAAA;IAC1B,OAAO,SAAS,CAAC,GAAG,CAAC,YAAY,CAAC,GAAG,CAAC,CAAC,CAAA;AACzC,CAAC;AAQD,MAAM,CAAC,KAAK,UAAU,KAAK,CACzB,KAAkB,EAClB,UAAwB,EAAE;;IAE1B,MAAM,WAAW,GAAG,MAAA,OAAO,CAAC,QAAQ,0CAAE,KAAK,CAAC,GAAG,EAAE,GAAG,EAAE,CAAA;IACtD,MAAM,GAAG,GAAG,MAAA,OAAO,CAAC,GAAG,mCAAI,WAAW,CAAA;IACtC,MAAM,MAAM,GAAG,MAAA,OAAO,CAAC,MAAM,mCAAI,cAAc,CAAC,GAAG,CAAC,CAAA;IAEpD,IAAI,CAAC,MAAM,EAAE,CAAC;QACZ,MAAM,IAAI,KAAK,CACb,kBAAkB,GAAG,aAAH,GAAG,cAAH,GAAG,GAAI,IAAI,SAAS;YACpC,GAAG,SAAS,CAAC,IAAI,EAAE;SACpB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CACf,CAAA;IACH,CAAC;IAED,MAAM,MAAM,GAAG,IAAI,MAAM,EAAE,CAAA;IAC3B,OAAO,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAA;AAC7B,CAAC;AAED,eAAe;AACf,MAAM,UAAU,qBAAqB;IACnC,OAAO,CAAC,GAAG,SAAS,CAAC,OAAO,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,EAAE,IAAI,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAA;AAC/E,CAAC"}

package/dist/types/DocumentAnchor.d.ts

@@ -0,0 +1,11 @@
+import { Number2 } from '@hamster-note/types';
+export interface DocumentAnchorWithTextId {
+    pageId: string;
+    textId?: string;
+}
+export interface DocumentAnchorWithPosition {
+    pageId: string;
+    position: Number2;
+}
+export type DocumentAnchor = DocumentAnchorWithPosition | DocumentAnchorWithTextId;
+//# sourceMappingURL=DocumentAnchor.d.ts.map

package/dist/types/DocumentAnchor.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"DocumentAnchor.d.ts","sourceRoot":"","sources":["../../src/types/DocumentAnchor.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,OAAO,EAAE,MAAM,qBAAqB,CAAA;AAE7C,MAAM,WAAW,wBAAwB;IACvC,MAAM,EAAE,MAAM,CAAA;IACd,MAAM,CAAC,EAAE,MAAM,CAAA;CAChB;AAGD,MAAM,WAAW,0BAA0B;IACzC,MAAM,EAAE,MAAM,CAAA;IACd,QAAQ,EAAE,OAAO,CAAA;CAClB;AAED,MAAM,MAAM,cAAc,GACtB,0BAA0B,GAC1B,wBAAwB,CAAA"}

package/dist/types/DocumentAnchor.js

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=DocumentAnchor.js.map

package/dist/types/DocumentAnchor.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"DocumentAnchor.js","sourceRoot":"","sources":["../../src/types/DocumentAnchor.ts"],"names":[],"mappings":""}

package/package.json

@@ -1,13 +1,14 @@
 {
   "name": "@hamster-note/document-parser",
-  "version": "0.1.0",
+  "version": "0.2.0",
   "description": "",
   "scripts": {
-    "build:all": "echo 'no-build'",
+    "build:all": "tsc -p tsconfig.build.json",
+    "prepublishOnly": "yarn build:all",
     "test": "echo 'no-test'"
   },
   "dependencies": {
-    "@hamster-note/types": "^0.1.0"
+    "@hamster-note/types": "^0.2.0"
   },
   "devDependencies": {
     "@eslint/js": "^9.39.2",
@@ -27,11 +28,18 @@
     "typescript": "^5.0.2",
     "typescript-eslint": "^8.50.0"
   },
-  "main": "src/index.ts",
-  "module": "src/index.ts",
-  "types": "src/index.ts",
+  "type": "module",
+  "main": "./dist/index.js",
+  "module": "./dist/index.js",
+  "types": "./dist/index.d.ts",
+  "exports": {
+    ".": {
+      "types": "./dist/index.d.ts",
+      "import": "./dist/index.js"
+    }
+  },
   "files": [
-    "src"
+    "dist"
   ],
   "packageManager": "yarn@1.22.22+sha512.a6b2f7906b721bba3d67d4aff083df04dad64c399707841b7acf00f6b133b7ac24255f2652fa22ae3534329dc6180534e98d17432037ff6fd140556e2bb3137e"
-}
+}

package/src/Document/index.ts

@@ -1,13 +0,0 @@
-import { HamsterPage } from '../Page'
-import { IntermediateOutline } from '@hamster-note/types'
-
-export abstract class HamsterDocument {
-  // 获取 Page 对象列表
-  abstract getPages(): Promise<HamsterPage[]>
-  // 获取 Page 对象
-  abstract getPage(pageNumber: number): Promise<HamsterPage | undefined>
-  // 获取大纲（PDF如果有内置的话）
-  abstract getOutline(): Promise<IntermediateOutline | undefined>
-  // 获取封面
-  abstract getCover(): Promise<HTMLCanvasElement | HTMLImageElement>
-}

package/src/Page/index.ts

@@ -1,22 +0,0 @@
-import { Number2 } from '@hamster-note/types'
-
-export enum RenderViews {
-  THUMBNAIL = 'thumbnail',
-  TEXT = 'text'
-}
-
-export interface RenderOptions {
-  views: RenderViews[]
-  scale: number
-}
-
-export abstract class HamsterPage {
-  abstract render(
-    container: HTMLDivElement,
-    options?: RenderOptions
-  ): Promise<void>
-  abstract getNumber(): number
-  abstract getSize(scale: number): Number2
-  abstract getPureText(): string
-  // abstract getTextDom
-}

package/src/Parser/index.ts

@@ -1,28 +0,0 @@
-import { HamsterDocument } from '../Document'
-import { IntermediateDocument } from '@hamster-note/types'
-
-export class DocumentParser {
-  static readonly ext: string
-  // Base static methods to be overridden by concrete parsers
-  static async encode(
-    _file: File | ArrayBuffer
-  ): Promise<IntermediateDocument | HamsterDocument | undefined> {
-    return Promise.resolve(undefined)
-  }
-  static async decode(
-    _intermediateDocument: IntermediateDocument
-  ): Promise<File | ArrayBuffer | undefined> {
-    return Promise.resolve(undefined)
-  }
-  protected static async toArrayBuffer(
-    fileOrBuffer: File | ArrayBuffer
-  ): Promise<ArrayBuffer> {
-    if (fileOrBuffer instanceof ArrayBuffer) return fileOrBuffer
-    return await new Promise<ArrayBuffer>((resolve, reject) => {
-      const reader = new FileReader()
-      reader.readAsArrayBuffer(fileOrBuffer)
-      reader.onload = () => resolve(reader.result as ArrayBuffer)
-      reader.onerror = (e) => reject(e)
-    })
-  }
-}

package/src/index.ts

@@ -1,5 +0,0 @@
-export * from './Document'
-export * from './Page'
-export * from './types/DocumentAnchor'
-export * from './Parser'
-export * from './register'

package/src/register/index.ts

@@ -1,30 +0,0 @@
-import { DocumentParser } from '../Parser'
-
-// Store constructors (classes), not instances
-const parserMap: Map<string, typeof DocumentParser> = new Map()
-
-// Register a parser class (constructor) which exposes a static `ext`
-export const registerParser = (parser: typeof DocumentParser) => {
-  console.log(parser.ext, parser)
-  parserMap.set(parser.ext, parser)
-}
-
-export function getParser(file: File): typeof DocumentParser | undefined {
-  const extension = file.name.split('.').pop()
-  if (!extension) {
-    return undefined
-  }
-  return parserMap.get(extension)
-}
-
-export function parse(file: File) {
-  const Parser = getParser(file)
-  console.log('Parse:', Parser)
-  if (!Parser) {
-    return undefined
-  }
-  // call static encode on the parser class
-  const res = Parser.encode(file)
-  // res.then((_doc) => { /* window.doc1 = _doc */ })
-  return res
-}

package/src/types/DocumentAnchor.ts

@@ -1,17 +0,0 @@
-// 文档锚（标记了文字id）
-import { Number2 } from '@hamster-note/types'
-
-export interface DocumentAnchorWithTextId {
-  pageId: string
-  textId?: string
-}
-
-// 文档锚（标记了位置）
-export interface DocumentAnchorWithPosition {
-  pageId: string
-  position: Number2
-}
-
-export type DocumentAnchor =
-  | DocumentAnchorWithPosition
-  | DocumentAnchorWithTextId