bitfire-lark-mcp 1.0.0

package/dist/mcp-tool/document-tool/read-full-content/extract-text.js

@@ -0,0 +1,109 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.BLOCK_TYPE = void 0;
+exports.extractTextFromElements = extractTextFromElements;
+exports.extractBlockText = extractBlockText;
+exports.headingPrefix = headingPrefix;
+exports.formatTableMarkdown = formatTableMarkdown;
+/** 从文本元素列表中提取纯文本 */
+function extractTextFromElements(elements) {
+    if (!(elements === null || elements === void 0 ? void 0 : elements.length))
+        return '';
+    return elements
+        .map((el) => {
+        var _a, _b, _c;
+        if ((_a = el.text_run) === null || _a === void 0 ? void 0 : _a.content)
+            return el.text_run.content;
+        if (el.mention_user)
+            return '@用户';
+        if ((_b = el.mention_doc) === null || _b === void 0 ? void 0 : _b.title)
+            return `@文档: ${el.mention_doc.title}`;
+        if ((_c = el.mention_doc) === null || _c === void 0 ? void 0 : _c.url)
+            return '@文档';
+        return '';
+    })
+        .join('');
+}
+/** 从 block.text 提取纯文本 */
+function extractBlockText(block) {
+    var _a, _b;
+    if (!((_b = (_a = block === null || block === void 0 ? void 0 : block.text) === null || _a === void 0 ? void 0 : _a.elements) === null || _b === void 0 ? void 0 : _b.length))
+        return '';
+    return extractTextFromElements(block.text.elements);
+}
+/** Block 类型枚举（与飞书 API 一致） */
+exports.BLOCK_TYPE = {
+    PAGE: 1,
+    TEXT: 2,
+    HEADING1: 3,
+    HEADING2: 4,
+    HEADING3: 5,
+    HEADING4: 6,
+    HEADING5: 7,
+    HEADING6: 8,
+    HEADING7: 9,
+    HEADING8: 10,
+    HEADING9: 11,
+    BULLET: 12,
+    ORDERED: 13,
+    CODE: 14,
+    QUOTE: 15,
+    TODO: 17,
+    BITABLE: 18,
+    CALLOUT: 19,
+    CHAT_CARD: 20,
+    DIAGRAM: 21,
+    DIVIDER: 22,
+    FILE: 23,
+    GRID: 24,
+    GRID_COLUMN: 25,
+    IFRAME: 26,
+    IMAGE: 27,
+    ISV: 28,
+    MINDNOTE: 29,
+    SHEET: 30,
+    TABLE: 31,
+    TABLE_CELL: 32,
+    VIEW: 33,
+    QUOTE_CONTAINER: 34,
+    TASK: 35,
+    OKR: 36,
+    BOARD: 37,
+    LINK_PREVIEW: 38,
+    /** 画板（文档内嵌画板，部分 API 版本返回 43） */
+    BOARD_DOCX: 43,
+};
+/** 标题层级对应 Markdown 前缀 */
+function headingPrefix(blockType) {
+    const level = blockType - exports.BLOCK_TYPE.HEADING1 + 1;
+    if (level >= 1 && level <= 9)
+        return '#'.repeat(level) + ' ';
+    return '';
+}
+/** 将表格块及其子单元格块格式化为 Markdown 表格 */
+function formatTableMarkdown(tableBlock, cellBlocks, tableProp) {
+    var _a, _b, _c;
+    const rowSize = (_a = tableProp === null || tableProp === void 0 ? void 0 : tableProp.row_size) !== null && _a !== void 0 ? _a : 0;
+    const columnSize = (_b = tableProp === null || tableProp === void 0 ? void 0 : tableProp.column_size) !== null && _b !== void 0 ? _b : 0;
+    if (rowSize <= 0 || columnSize <= 0) {
+        return '[表格]（无法解析行列数）\n';
+    }
+    const cells = [];
+    for (let r = 0; r < rowSize; r++) {
+        cells[r] = [];
+        for (let c = 0; c < columnSize; c++) {
+            const idx = r * columnSize + c;
+            const cell = cellBlocks[idx];
+            const text = cell ? extractBlockText(cell) : '';
+            cells[r][c] = text.replace(/\|/g, '\\|').replace(/\n/g, ' ');
+        }
+    }
+    const lines = [];
+    for (let r = 0; r < rowSize; r++) {
+        lines.push('| ' + cells[r].join(' | ') + ' |');
+        if (r === 0 && ((_c = tableProp === null || tableProp === void 0 ? void 0 : tableProp.header_row) !== null && _c !== void 0 ? _c : true)) {
+            lines.push('| ' + Array(columnSize).fill('---').join(' | ') + ' |');
+        }
+    }
+    return lines.join('\n') + '\n';
+}

package/dist/mcp-tool/document-tool/read-full-content/index.d.ts

@@ -0,0 +1,3 @@
+export { readDocxFullContent } from './reader';
+export type { ReadFullContentOptions } from './types';
+export { extractBlockText, extractTextFromElements, BLOCK_TYPE } from './extract-text';

package/dist/mcp-tool/document-tool/read-full-content/index.js

@@ -0,0 +1,9 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.BLOCK_TYPE = exports.extractTextFromElements = exports.extractBlockText = exports.readDocxFullContent = void 0;
+var reader_1 = require("./reader");
+Object.defineProperty(exports, "readDocxFullContent", { enumerable: true, get: function () { return reader_1.readDocxFullContent; } });
+var extract_text_1 = require("./extract-text");
+Object.defineProperty(exports, "extractBlockText", { enumerable: true, get: function () { return extract_text_1.extractBlockText; } });
+Object.defineProperty(exports, "extractTextFromElements", { enumerable: true, get: function () { return extract_text_1.extractTextFromElements; } });
+Object.defineProperty(exports, "BLOCK_TYPE", { enumerable: true, get: function () { return extract_text_1.BLOCK_TYPE; } });

package/dist/mcp-tool/document-tool/read-full-content/reader.d.ts

@@ -0,0 +1,12 @@
+import type { Client } from '@larksuiteoapi/node-sdk';
+import type { ReadFullContentOptions } from './types';
+/**
+ * 将文档所有块转换为可读文本（含表格、流程图等描述）
+ */
+export declare function readDocxFullContent(client: Client, options: ReadFullContentOptions & {
+    useUAT: boolean;
+    userAccessToken?: string;
+}): Promise<{
+    title?: string;
+    content: string;
+}>;

package/dist/mcp-tool/document-tool/read-full-content/reader.js

@@ -0,0 +1,283 @@
+"use strict";
+var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    var desc = Object.getOwnPropertyDescriptor(m, k);
+    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
+      desc = { enumerable: true, get: function() { return m[k]; } };
+    }
+    Object.defineProperty(o, k2, desc);
+}) : (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    o[k2] = m[k];
+}));
+var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
+    Object.defineProperty(o, "default", { enumerable: true, value: v });
+}) : function(o, v) {
+    o["default"] = v;
+});
+var __importStar = (this && this.__importStar) || (function () {
+    var ownKeys = function(o) {
+        ownKeys = Object.getOwnPropertyNames || function (o) {
+            var ar = [];
+            for (var k in o) if (Object.prototype.hasOwnProperty.call(o, k)) ar[ar.length] = k;
+            return ar;
+        };
+        return ownKeys(o);
+    };
+    return function (mod) {
+        if (mod && mod.__esModule) return mod;
+        var result = {};
+        if (mod != null) for (var k = ownKeys(mod), i = 0; i < k.length; i++) if (k[i] !== "default") __createBinding(result, mod, k[i]);
+        __setModuleDefault(result, mod);
+        return result;
+    };
+})();
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.readDocxFullContent = readDocxFullContent;
+const lark = __importStar(require("@larksuiteoapi/node-sdk"));
+const extract_text_1 = require("./extract-text");
+const whiteboard_1 = require("./whiteboard");
+const DEFAULT_PAGE_SIZE = 200;
+/**
+ * 拉取文档所有块（分页）
+ */
+async function fetchAllBlocks(client, documentId, options) {
+    var _a, _b, _c;
+    const items = [];
+    let pageToken;
+    const call = options.useUAT && options.userAccessToken
+        ? (params) => client.docx.v1.documentBlock.list(params, lark.withUserAccessToken(options.userAccessToken))
+        : (params) => client.docx.v1.documentBlock.list(params);
+    do {
+        const res = await call({
+            path: { document_id: documentId },
+            params: {
+                page_size: (_a = options.pageSize) !== null && _a !== void 0 ? _a : DEFAULT_PAGE_SIZE,
+                page_token: pageToken,
+                document_revision_id: options.documentRevisionId,
+            },
+        });
+        const data = ((_b = res === null || res === void 0 ? void 0 : res.data) !== null && _b !== void 0 ? _b : res);
+        if ((_c = data.items) === null || _c === void 0 ? void 0 : _c.length)
+            items.push(...data.items);
+        pageToken = data.has_more ? data.page_token : undefined;
+    } while (pageToken);
+    return items;
+}
+/**
+ * 将文档所有块转换为可读文本（含表格、流程图等描述）
+ */
+async function readDocxFullContent(client, options) {
+    var _a, _b, _c, _d, _e, _f, _g, _h, _j, _k, _l, _m, _o, _p, _q, _r, _s, _t, _u, _v, _w, _x;
+    const { documentId, documentRevisionId, pageSize } = options;
+    const docCall = options.useUAT && options.userAccessToken
+        ? (params) => client.docx.v1.document.get(params, lark.withUserAccessToken(options.userAccessToken))
+        : (params) => client.docx.v1.document.get(params);
+    let title = '';
+    try {
+        const docRes = await docCall({ path: { document_id: documentId } });
+        const docData = (_a = docRes === null || docRes === void 0 ? void 0 : docRes.data) !== null && _a !== void 0 ? _a : docRes;
+        title = (_c = (_b = docData === null || docData === void 0 ? void 0 : docData.document) === null || _b === void 0 ? void 0 : _b.title) !== null && _c !== void 0 ? _c : '';
+    }
+    catch (_y) {
+        // 仅读块时可能无权限 get document，继续用 blocks
+    }
+    const blocks = await fetchAllBlocks(client, documentId, {
+        useUAT: options.useUAT,
+        userAccessToken: options.userAccessToken,
+        pageSize,
+        documentRevisionId,
+    });
+    const blockMap = new Map();
+    for (const b of blocks) {
+        if (b.block_id)
+            blockMap.set(b.block_id, b);
+    }
+    const out = [];
+    if (title)
+        out.push(`# ${title}\n`);
+    let i = 0;
+    while (i < blocks.length) {
+        const block = blocks[i];
+        const blockType = (_d = block.block_type) !== null && _d !== void 0 ? _d : 0;
+        switch (blockType) {
+            case extract_text_1.BLOCK_TYPE.PAGE:
+                break;
+            case extract_text_1.BLOCK_TYPE.TEXT: {
+                const t = (0, extract_text_1.extractBlockText)(block);
+                if (t)
+                    out.push(t + '\n');
+                break;
+            }
+            case extract_text_1.BLOCK_TYPE.HEADING1:
+            case extract_text_1.BLOCK_TYPE.HEADING2:
+            case extract_text_1.BLOCK_TYPE.HEADING3:
+            case extract_text_1.BLOCK_TYPE.HEADING4:
+            case extract_text_1.BLOCK_TYPE.HEADING5:
+            case extract_text_1.BLOCK_TYPE.HEADING6:
+            case extract_text_1.BLOCK_TYPE.HEADING7:
+            case extract_text_1.BLOCK_TYPE.HEADING8:
+            case extract_text_1.BLOCK_TYPE.HEADING9: {
+                const t = (0, extract_text_1.extractBlockText)(block);
+                out.push((0, extract_text_1.headingPrefix)(blockType) + t + '\n');
+                break;
+            }
+            case extract_text_1.BLOCK_TYPE.BULLET: {
+                const t = (0, extract_text_1.extractBlockText)(block);
+                if (t)
+                    out.push('- ' + t + '\n');
+                break;
+            }
+            case extract_text_1.BLOCK_TYPE.ORDERED: {
+                const t = (0, extract_text_1.extractBlockText)(block);
+                if (t)
+                    out.push('1. ' + t + '\n');
+                break;
+            }
+            case extract_text_1.BLOCK_TYPE.CODE: {
+                const t = (0, extract_text_1.extractBlockText)(block);
+                out.push('```\n' + (t || '') + '\n```\n');
+                break;
+            }
+            case extract_text_1.BLOCK_TYPE.QUOTE: {
+                const t = (0, extract_text_1.extractBlockText)(block);
+                if (t)
+                    out.push('> ' + t.replace(/\n/g, '\n> ') + '\n');
+                break;
+            }
+            case extract_text_1.BLOCK_TYPE.TODO: {
+                const t = (0, extract_text_1.extractBlockText)(block);
+                if (t)
+                    out.push('- [ ] ' + t + '\n');
+                break;
+            }
+            case extract_text_1.BLOCK_TYPE.DIVIDER:
+                out.push('---\n');
+                break;
+            case extract_text_1.BLOCK_TYPE.TABLE: {
+                const prop = (_e = block.table) === null || _e === void 0 ? void 0 : _e.property;
+                const rowSize = (_f = prop === null || prop === void 0 ? void 0 : prop.row_size) !== null && _f !== void 0 ? _f : 0;
+                const columnSize = (_g = prop === null || prop === void 0 ? void 0 : prop.column_size) !== null && _g !== void 0 ? _g : 0;
+                const cellCount = rowSize * columnSize;
+                const cellBlocks = [];
+                for (let j = 1; j <= cellCount && i + j < blocks.length; j++) {
+                    const next = blocks[i + j];
+                    if ((next === null || next === void 0 ? void 0 : next.block_type) === extract_text_1.BLOCK_TYPE.TABLE_CELL) {
+                        cellBlocks.push(next);
+                    }
+                    else {
+                        break;
+                    }
+                }
+                out.push((0, extract_text_1.formatTableMarkdown)(block, cellBlocks, prop));
+                i += cellBlocks.length;
+                break;
+            }
+            case extract_text_1.BLOCK_TYPE.TABLE_CELL:
+                // 已在上方 Table 分支中处理，不应单独出现（若出现则跳过）
+                break;
+            case extract_text_1.BLOCK_TYPE.SHEET:
+                out.push('[电子表格]');
+                if (((_h = block.sheet) === null || _h === void 0 ? void 0 : _h.row_size) != null || ((_j = block.sheet) === null || _j === void 0 ? void 0 : _j.column_size) != null) {
+                    out.push(` (${(_k = block.sheet.row_size) !== null && _k !== void 0 ? _k : '-'} 行 × ${(_l = block.sheet.column_size) !== null && _l !== void 0 ? _l : '-'} 列)`);
+                }
+                out.push('\n');
+                break;
+            case extract_text_1.BLOCK_TYPE.BITABLE:
+                out.push('[多维表格]\n');
+                break;
+            case extract_text_1.BLOCK_TYPE.DIAGRAM:
+                if ((_m = block.diagram) === null || _m === void 0 ? void 0 : _m.content) {
+                    out.push('【流程图/UML】\n' + block.diagram.content + '\n');
+                }
+                else {
+                    out.push('[流程图 / UML 图]\n');
+                }
+                break;
+            case extract_text_1.BLOCK_TYPE.IMAGE:
+                out.push('[图片]\n');
+                break;
+            case extract_text_1.BLOCK_TYPE.FILE:
+                out.push('[文件]\n');
+                break;
+            case extract_text_1.BLOCK_TYPE.IFRAME:
+                if ((_p = (_o = block.iframe) === null || _o === void 0 ? void 0 : _o.component) === null || _p === void 0 ? void 0 : _p.url) {
+                    out.push('[内嵌链接] ' + block.iframe.component.url + '\n');
+                }
+                else {
+                    out.push('[内嵌内容]\n');
+                }
+                break;
+            case extract_text_1.BLOCK_TYPE.CALLOUT: {
+                const t = (0, extract_text_1.extractBlockText)(block);
+                if (t)
+                    out.push('> ' + t.replace(/\n/g, '\n> ') + '\n');
+                break;
+            }
+            case extract_text_1.BLOCK_TYPE.MINDNOTE:
+                out.push('[思维笔记]\n');
+                break;
+            case extract_text_1.BLOCK_TYPE.GRID:
+            case extract_text_1.BLOCK_TYPE.GRID_COLUMN:
+                // 分栏：子块会按顺序出现在后面，文本会在子块中提取，此处仅做占位或跳过
+                break;
+            case extract_text_1.BLOCK_TYPE.BOARD:
+            case extract_text_1.BLOCK_TYPE.BOARD_DOCX:
+            case extract_text_1.BLOCK_TYPE.VIEW: {
+                const viewBlock = block;
+                let whiteboardId = (_u = (_s = (_q = block.token) !== null && _q !== void 0 ? _q : (_r = viewBlock.view) === null || _r === void 0 ? void 0 : _r.token) !== null && _s !== void 0 ? _s : (_t = viewBlock.view) === null || _t === void 0 ? void 0 : _t.whiteboard_token) !== null && _u !== void 0 ? _u : (_v = viewBlock.board) === null || _v === void 0 ? void 0 : _v.token;
+                if (!whiteboardId && viewBlock.view && typeof viewBlock.view === 'object') {
+                    const v = viewBlock.view;
+                    whiteboardId = [v.whiteboard_token, v.whiteboard_id, v.token, v.board_token].find((x) => typeof x === 'string' && x.length > 0);
+                }
+                if (whiteboardId) {
+                    try {
+                        const nodes = await (0, whiteboard_1.fetchWhiteboardNodes)(client, whiteboardId, {
+                            useUAT: options.useUAT,
+                            userAccessToken: options.userAccessToken,
+                        });
+                        const nodeText = (0, whiteboard_1.formatWhiteboardNodes)(nodes);
+                        if (nodeText && !nodeText.includes('（画板无节点）')) {
+                            out.push('[画板]\n' + nodeText + '\n');
+                        }
+                        else if (nodes.length > 0) {
+                            out.push('[画板]\n' + nodeText + '\n');
+                        }
+                        else {
+                            out.push('[画板]\n（已请求画板节点接口，返回 0 个节点；可能需应用具备「画板」相关权限）\n');
+                        }
+                    }
+                    catch (e) {
+                        const msg = (_w = e === null || e === void 0 ? void 0 : e.message) !== null && _w !== void 0 ? _w : '';
+                        out.push('[画板]\n（拉取画板节点失败: ' + (msg.slice(0, 80) || '未知错误') + '）\n');
+                    }
+                }
+                else {
+                    const keys = Object.keys(block).filter((k) => !k.startsWith('_')).join(', ');
+                    out.push('[画板]\n（未获取到画板 ID，无法拉取节点；当前块字段: ' + keys + '）\n');
+                }
+                break;
+            }
+            case extract_text_1.BLOCK_TYPE.LINK_PREVIEW:
+                if ((_x = block.link_preview) === null || _x === void 0 ? void 0 : _x.url) {
+                    out.push('[链接] ' +
+                        block.link_preview.url +
+                        '\n');
+                }
+                break;
+            case extract_text_1.BLOCK_TYPE.CHAT_CARD:
+            case extract_text_1.BLOCK_TYPE.ISV:
+            case extract_text_1.BLOCK_TYPE.QUOTE_CONTAINER:
+            case extract_text_1.BLOCK_TYPE.TASK:
+            case extract_text_1.BLOCK_TYPE.OKR:
+            default: {
+                const t = (0, extract_text_1.extractBlockText)(block);
+                if (t)
+                    out.push(t + '\n');
+                break;
+            }
+        }
+        i++;
+    }
+    return { title: title || undefined, content: out.join('').trim() };
+}

package/dist/mcp-tool/document-tool/read-full-content/types.d.ts

@@ -0,0 +1,78 @@
+/**
+ * 飞书 docx 文档块 API 返回的块结构（与 documentBlock.list 一致）
+ */
+export interface DocxBlockTextElement {
+    text_run?: {
+        content?: string;
+    };
+    mention_user?: {
+        user_id?: string;
+    };
+    mention_doc?: {
+        url?: string;
+        title?: string;
+    };
+}
+export interface DocxBlockText {
+    elements?: DocxBlockTextElement[];
+}
+export interface DocxBlockTableProperty {
+    row_size?: number;
+    column_size?: number;
+    header_row?: boolean;
+    header_column?: boolean;
+}
+export interface DocxBlockTable {
+    property?: DocxBlockTableProperty;
+}
+export interface DocxBlockSheet {
+    row_size?: number;
+    column_size?: number;
+}
+export interface DocxBlockDiagram {
+    diagram_type?: number;
+    /** 流程图 / UML 的文本描述或数据，部分类型可能包含 content */
+    content?: string;
+}
+export interface DocxBlockItem {
+    block_id?: string;
+    block_type?: number;
+    parent_id?: string;
+    children?: string[];
+    index?: number;
+    text?: DocxBlockText;
+    table?: DocxBlockTable;
+    sheet?: DocxBlockSheet;
+    diagram?: DocxBlockDiagram;
+    bitable?: {
+        view_type?: number;
+    };
+    grid?: {
+        column_size?: number;
+    };
+    image?: unknown;
+    file?: unknown;
+    iframe?: {
+        component?: {
+            url?: string;
+        };
+    };
+    callout?: unknown;
+    link_preview?: {
+        url?: string;
+    };
+    board?: unknown;
+    /** 文档 token（如 Bitable/Sheet 可能带 token） */
+    token?: string;
+}
+export interface DocxBlockListResponse {
+    items?: DocxBlockItem[];
+    page_token?: string;
+    has_more?: boolean;
+}
+export interface ReadFullContentOptions {
+    documentId: string;
+    lang?: number;
+    documentRevisionId?: number;
+    pageSize?: number;
+}

package/dist/mcp-tool/document-tool/read-full-content/types.js

@@ -0,0 +1,2 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });

package/dist/mcp-tool/document-tool/read-full-content/whiteboard.d.ts

@@ -0,0 +1,53 @@
+import type { Client } from '@larksuiteoapi/node-sdk';
+/** 画板节点列表 API 返回的单个节点（兼容多种可能字段） */
+export interface WhiteboardNodeItem {
+    node_type?: number | string;
+    resource_type?: string;
+    text?: {
+        elements?: Array<{
+            content?: string;
+        }>;
+    };
+    content?: string;
+    /** 形状/图形节点常见字段 */
+    shape?: {
+        text?: string;
+        label?: string;
+        [k: string]: unknown;
+    };
+    label?: string;
+    title?: string;
+    name?: string;
+    body?: string;
+    description?: string;
+    /** 流程图/状态图可能有的子节点与边 */
+    nodes?: Array<{
+        id?: string;
+        label?: string;
+        text?: string;
+        [k: string]: unknown;
+    }>;
+    edges?: Array<{
+        source?: string;
+        target?: string;
+        label?: string;
+        [k: string]: unknown;
+    }>;
+    [key: string]: unknown;
+}
+export interface WhiteboardNodesResponse {
+    items?: WhiteboardNodeItem[];
+    nodes?: WhiteboardNodeItem[];
+    node_list?: WhiteboardNodeItem[];
+    page_token?: string;
+    has_more?: boolean;
+}
+/**
+ * 拉取画板内所有节点（分页）
+ */
+export declare function fetchWhiteboardNodes(client: Client, whiteboardId: string, options: {
+    useUAT: boolean;
+    userAccessToken?: string;
+}): Promise<WhiteboardNodeItem[]>;
+/** 将画板节点列表格式化为可读文本（含类型标注，便于状态流转图） */
+export declare function formatWhiteboardNodes(nodes: WhiteboardNodeItem[]): string;