@undefineds.co/xpod 0.1.0-local.202602081751

package/dist/dns/tencent/TencentDnsProvider.jsonld

@@ -0,0 +1,239 @@
+{
+  "@context": [
+    "https://linkedsoftwaredependencies.org/bundles/npm/@undefineds.co/xpod/^0.0.0/components/context.jsonld"
+  ],
+  "@id": "npmd:@undefineds.co/xpod",
+  "components": [
+    {
+      "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider",
+      "@type": "Class",
+      "requireElement": "TencentDnsProvider",
+      "extends": [
+        "undefineds:dist/dns/DnsProvider.jsonld#ListableDnsProvider"
+      ],
+      "comment": "基于腾讯云 DNSPod v2 API 的 DNS 提供方实现。",
+      "parameters": [
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider_options_tokenId",
+          "range": {
+            "@type": "ParameterRangeUnion",
+            "parameterRangeElements": [
+              "xsd:string",
+              {
+                "@type": "ParameterRangeWildcard"
+              },
+              {
+                "@type": "ParameterRangeUndefined"
+              }
+            ]
+          }
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider_options_token",
+          "range": {
+            "@type": "ParameterRangeUnion",
+            "parameterRangeElements": [
+              "xsd:string",
+              {
+                "@type": "ParameterRangeWildcard"
+              },
+              {
+                "@type": "ParameterRangeUndefined"
+              }
+            ]
+          }
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider_options_baseUrl",
+          "range": {
+            "@type": "ParameterRangeUnion",
+            "parameterRangeElements": [
+              "xsd:string",
+              {
+                "@type": "ParameterRangeWildcard"
+              },
+              {
+                "@type": "ParameterRangeUndefined"
+              }
+            ]
+          },
+          "comment": "API 根地址，默认为 `https://dnsapi.cn`。"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider_options_defaultLineId",
+          "range": {
+            "@type": "ParameterRangeUnion",
+            "parameterRangeElements": [
+              "xsd:string",
+              {
+                "@type": "ParameterRangeWildcard"
+              },
+              {
+                "@type": "ParameterRangeUndefined"
+              }
+            ]
+          },
+          "comment": "默认线路 ID，腾讯云 DNS 默认线路为 `0`。 若未指定且记录已存在，将复用已有线路。"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider_options_timeoutMs",
+          "range": {
+            "@type": "ParameterRangeUnion",
+            "parameterRangeElements": [
+              "xsd:number",
+              "xsd:string",
+              {
+                "@type": "ParameterRangeWildcard"
+              },
+              {
+                "@type": "ParameterRangeUndefined"
+              }
+            ]
+          },
+          "comment": "HTTP 请求超时毫秒数，默认 10s。"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider_options_userAgent",
+          "range": {
+            "@type": "ParameterRangeUnion",
+            "parameterRangeElements": [
+              "xsd:string",
+              {
+                "@type": "ParameterRangeWildcard"
+              },
+              {
+                "@type": "ParameterRangeUndefined"
+              }
+            ]
+          },
+          "comment": "User-Agent 标识，便于审计。"
+        }
+      ],
+      "memberFields": [
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_logger",
+          "memberFieldName": "logger"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_tokenId",
+          "memberFieldName": "tokenId"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_token",
+          "memberFieldName": "token"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_baseUrl",
+          "memberFieldName": "baseUrl"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_defaultLineId",
+          "memberFieldName": "defaultLineId"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_timeoutMs",
+          "memberFieldName": "timeoutMs"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_userAgent",
+          "memberFieldName": "userAgent"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_enabled",
+          "memberFieldName": "enabled"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_constructor",
+          "memberFieldName": "constructor"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_upsertRecord",
+          "memberFieldName": "upsertRecord"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_deleteRecord",
+          "memberFieldName": "deleteRecord"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_listRecords",
+          "memberFieldName": "listRecords"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_toSummary",
+          "memberFieldName": "toSummary"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_findRecord",
+          "memberFieldName": "findRecord"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_compareLine",
+          "memberFieldName": "compareLine"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_compareTtl",
+          "memberFieldName": "compareTtl"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_normalizeTtl",
+          "memberFieldName": "normalizeTtl"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_callApi",
+          "memberFieldName": "callApi"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_normalizeString",
+          "memberFieldName": "normalizeString"
+        },
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider__member_normalizeTimeout",
+          "memberFieldName": "normalizeTimeout"
+        }
+      ],
+      "constructorArguments": [
+        {
+          "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider_options__constructorArgument",
+          "fields": [
+            {
+              "keyRaw": "tokenId",
+              "value": {
+                "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider_options_tokenId"
+              }
+            },
+            {
+              "keyRaw": "token",
+              "value": {
+                "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider_options_token"
+              }
+            },
+            {
+              "keyRaw": "baseUrl",
+              "value": {
+                "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider_options_baseUrl"
+              }
+            },
+            {
+              "keyRaw": "defaultLineId",
+              "value": {
+                "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider_options_defaultLineId"
+              }
+            },
+            {
+              "keyRaw": "timeoutMs",
+              "value": {
+                "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider_options_timeoutMs"
+              }
+            },
+            {
+              "keyRaw": "userAgent",
+              "value": {
+                "@id": "undefineds:dist/dns/tencent/TencentDnsProvider.jsonld#TencentDnsProvider_options_userAgent"
+              }
+            }
+          ]
+        }
+      ]
+    }
+  ]
+}

package/dist/document/Chunker.d.ts

@@ -0,0 +1,64 @@
+/**
+ * Chunker - 文档分块接口
+ *
+ * 将 Markdown 文档按标题层级分块，保留结构信息
+ */
+/**
+ * 文本块
+ */
+export interface TextChunk {
+    /** 块 ID */
+    id: string;
+    /** 标题层级 (1-6)，0 表示文档开头无标题部分 */
+    level: number;
+    /** 标题文本 */
+    heading: string;
+    /** 块内容（包含标题和正文） */
+    content: string;
+    /** 标题路径 (如 ["Introduction", "Getting Started"]) */
+    path: string[];
+    /** 在原文中的起始偏移量 */
+    startOffset: number;
+    /** 在原文中的结束偏移量 */
+    endOffset: number;
+    /** 父块 ID */
+    parentId?: string;
+    /** 子块 */
+    children: TextChunk[];
+}
+/**
+ * 分块器接口
+ */
+export interface Chunker {
+    /**
+     * 将 Markdown 分块为树形结构
+     * @param markdown Markdown 内容
+     * @returns 根级块数组（树形结构）
+     */
+    chunk(markdown: string): TextChunk[];
+    /**
+     * 将树形结构扁平化
+     * @param chunks 树形结构的块
+     * @returns 扁平化的块数组
+     */
+    flatten(chunks: TextChunk[]): TextChunk[];
+}
+/**
+ * 块元数据（用于存储到 .meta 文件）
+ */
+export interface ChunkMetadata {
+    /** 块 ID */
+    id: string;
+    /** 标题层级 */
+    level: number;
+    /** 标题文本 */
+    heading: string;
+    /** 起始偏移量 */
+    startOffset: number;
+    /** 结束偏移量 */
+    endOffset: number;
+    /** 父块 ID */
+    parentId?: string;
+    /** 向量 ID（存储后填充） */
+    vectorId?: number;
+}

package/dist/document/Chunker.js

@@ -0,0 +1,8 @@
+"use strict";
+/**
+ * Chunker - 文档分块接口
+ *
+ * 将 Markdown 文档按标题层级分块，保留结构信息
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+//# sourceMappingURL=Chunker.js.map

package/dist/document/Chunker.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"Chunker.js","sourceRoot":"","sources":["../../src/document/Chunker.ts"],"names":[],"mappings":";AAAA;;;;GAIG","sourcesContent":["/**\n * Chunker - 文档分块接口\n *\n * 将 Markdown 文档按标题层级分块，保留结构信息\n */\n\n/**\n * 文本块\n */\nexport interface TextChunk {\n  /** 块 ID */\n  id: string;\n\n  /** 标题层级 (1-6)，0 表示文档开头无标题部分 */\n  level: number;\n\n  /** 标题文本 */\n  heading: string;\n\n  /** 块内容（包含标题和正文） */\n  content: string;\n\n  /** 标题路径 (如 [\"Introduction\", \"Getting Started\"]) */\n  path: string[];\n\n  /** 在原文中的起始偏移量 */\n  startOffset: number;\n\n  /** 在原文中的结束偏移量 */\n  endOffset: number;\n\n  /** 父块 ID */\n  parentId?: string;\n\n  /** 子块 */\n  children: TextChunk[];\n}\n\n/**\n * 分块器接口\n */\nexport interface Chunker {\n  /**\n   * 将 Markdown 分块为树形结构\n   * @param markdown Markdown 内容\n   * @returns 根级块数组（树形结构）\n   */\n  chunk(markdown: string): TextChunk[];\n\n  /**\n   * 将树形结构扁平化\n   * @param chunks 树形结构的块\n   * @returns 扁平化的块数组\n   */\n  flatten(chunks: TextChunk[]): TextChunk[];\n}\n\n/**\n * 块元数据（用于存储到 .meta 文件）\n */\nexport interface ChunkMetadata {\n  /** 块 ID */\n  id: string;\n  /** 标题层级 */\n  level: number;\n  /** 标题文本 */\n  heading: string;\n  /** 起始偏移量 */\n  startOffset: number;\n  /** 结束偏移量 */\n  endOffset: number;\n  /** 父块 ID */\n  parentId?: string;\n  /** 向量 ID（存储后填充） */\n  vectorId?: number;\n}\n"]}

package/dist/document/DocumentParser.d.ts

@@ -0,0 +1,58 @@
+/**
+ * DocumentParser - 文档解析接口
+ *
+ * 将各种格式的文档（PDF, Office, HTML 等）转换为 Markdown
+ */
+/**
+ * 解析后的文档
+ */
+export interface ParsedDocument {
+    /** Markdown 格式的内容 */
+    markdown: string;
+    /** 文档元数据 */
+    metadata: {
+        /** 文档标题 */
+        title?: string;
+        /** 文档描述 */
+        description?: string;
+        /** 原始 URL */
+        url: string;
+        /** 内容类型 */
+        contentType?: string;
+        /** 字数统计 */
+        wordCount?: number;
+    };
+    /** 原始 URL */
+    rawUrl: string;
+}
+/**
+ * 文档解析器接口
+ */
+export interface DocumentParser {
+    /**
+     * 解析文档为 Markdown
+     * @param url 文档 URL
+     * @param options 解析选项
+     * @returns 解析后的文档
+     */
+    parse(url: string, options?: ParseOptions): Promise<ParsedDocument>;
+    /**
+     * 检查是否支持该 URL
+     * @param url 文档 URL
+     * @returns 是否支持
+     */
+    supports(url: string): boolean;
+}
+/**
+ * 解析选项
+ */
+export interface ParseOptions {
+    /** 访问令牌（用于需要认证的资源） */
+    accessToken?: string;
+    /** 超时时间（毫秒） */
+    timeout?: number;
+    /** 是否包含图片描述 */
+    includeImages?: boolean;
+    /** 是否包含链接 */
+    includeLinks?: boolean;
+}

package/dist/document/DocumentParser.js

@@ -0,0 +1,8 @@
+"use strict";
+/**
+ * DocumentParser - 文档解析接口
+ *
+ * 将各种格式的文档（PDF, Office, HTML 等）转换为 Markdown
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+//# sourceMappingURL=DocumentParser.js.map

package/dist/document/DocumentParser.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"DocumentParser.js","sourceRoot":"","sources":["../../src/document/DocumentParser.ts"],"names":[],"mappings":";AAAA;;;;GAIG","sourcesContent":["/**\n * DocumentParser - 文档解析接口\n *\n * 将各种格式的文档（PDF, Office, HTML 等）转换为 Markdown\n */\n\n/**\n * 解析后的文档\n */\nexport interface ParsedDocument {\n  /** Markdown 格式的内容 */\n  markdown: string;\n\n  /** 文档元数据 */\n  metadata: {\n    /** 文档标题 */\n    title?: string;\n    /** 文档描述 */\n    description?: string;\n    /** 原始 URL */\n    url: string;\n    /** 内容类型 */\n    contentType?: string;\n    /** 字数统计 */\n    wordCount?: number;\n  };\n\n  /** 原始 URL */\n  rawUrl: string;\n}\n\n/**\n * 文档解析器接口\n */\nexport interface DocumentParser {\n  /**\n   * 解析文档为 Markdown\n   * @param url 文档 URL\n   * @param options 解析选项\n   * @returns 解析后的文档\n   */\n  parse(url: string, options?: ParseOptions): Promise<ParsedDocument>;\n\n  /**\n   * 检查是否支持该 URL\n   * @param url 文档 URL\n   * @returns 是否支持\n   */\n  supports(url: string): boolean;\n}\n\n/**\n * 解析选项\n */\nexport interface ParseOptions {\n  /** 访问令牌（用于需要认证的资源） */\n  accessToken?: string;\n  /** 超时时间（毫秒） */\n  timeout?: number;\n  /** 是否包含图片描述 */\n  includeImages?: boolean;\n  /** 是否包含链接 */\n  includeLinks?: boolean;\n}\n"]}

package/dist/document/HeadingChunker.d.ts

@@ -0,0 +1,34 @@
+/**
+ * HeadingChunker - 按标题层级分块
+ *
+ * 将 Markdown 文档按标题层级分块，构建树形结构
+ * 每个块记录 startOffset 和 endOffset
+ */
+import type { Chunker, TextChunk } from './Chunker';
+/**
+ * 按标题层级分块的实现
+ */
+export declare class HeadingChunker implements Chunker {
+    /**
+     * 将 Markdown 分块为树形结构
+     */
+    chunk(markdown: string): TextChunk[];
+    /**
+     * 将树形结构扁平化
+     */
+    flatten(chunks: TextChunk[]): TextChunk[];
+    /**
+     * 更新块的 endOffset
+     * 每个块的 endOffset 应该是下一个同级或更高级块的 startOffset - 1
+     * 或者文档末尾
+     */
+    private updateEndOffsets;
+    /**
+     * 找到下一个同级或更高级的块
+     */
+    private findNextSiblingOrHigher;
+    /**
+     * 生成块 ID
+     */
+    private generateChunkId;
+}

package/dist/document/HeadingChunker.js

@@ -0,0 +1,182 @@
+"use strict";
+/**
+ * HeadingChunker - 按标题层级分块
+ *
+ * 将 Markdown 文档按标题层级分块，构建树形结构
+ * 每个块记录 startOffset 和 endOffset
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.HeadingChunker = void 0;
+const crypto_1 = require("crypto");
+/**
+ * 按标题层级分块的实现
+ */
+class HeadingChunker {
+    /**
+     * 将 Markdown 分块为树形结构
+     */
+    chunk(markdown) {
+        const lines = markdown.split('\n');
+        const rootChunks = [];
+        const stack = []; // 用于构建树形结构
+        let currentOffset = 0;
+        let currentChunk = null;
+        let contentBuffer = '';
+        let inCodeBlock = false;
+        for (let i = 0; i < lines.length; i++) {
+            const line = lines[i];
+            const lineLength = line.length + (i < lines.length - 1 ? 1 : 0); // +1 for \n except last line
+            // 检查代码块边界
+            if (line.trim().startsWith('```')) {
+                inCodeBlock = !inCodeBlock;
+                // 代码块边界添加到当前块内容
+                if (currentChunk) {
+                    contentBuffer += line + '\n';
+                }
+                currentOffset += lineLength;
+                continue;
+            }
+            // 在代码块内，不解析标题
+            if (inCodeBlock) {
+                if (currentChunk) {
+                    contentBuffer += line + '\n';
+                }
+                currentOffset += lineLength;
+                continue;
+            }
+            // 检查是否是标题行
+            const headingMatch = line.match(/^(#{1,6})\s+(.+)$/);
+            if (headingMatch) {
+                // 发现新标题，先保存之前的块内容
+                if (currentChunk) {
+                    currentChunk.content = contentBuffer.trim();
+                    currentChunk.endOffset = currentOffset;
+                }
+                const level = headingMatch[1].length;
+                const heading = headingMatch[2].trim();
+                // 创建新块
+                const newChunk = {
+                    id: this.generateChunkId(),
+                    level,
+                    heading,
+                    content: '',
+                    path: [heading], // 初始路径包含自己
+                    startOffset: currentOffset,
+                    endOffset: currentOffset + lineLength, // 临时值，后面会更新
+                    children: [],
+                };
+                // 构建树形结构
+                // 找到合适的父节点
+                while (stack.length > 0 && stack[stack.length - 1].level >= level) {
+                    stack.pop();
+                }
+                if (stack.length > 0) {
+                    const parent = stack[stack.length - 1];
+                    newChunk.parentId = parent.id;
+                    // path 包含父级的 path 加上当前标题
+                    newChunk.path = [...parent.path, heading];
+                    parent.children.push(newChunk);
+                }
+                else {
+                    newChunk.path = [heading];
+                    rootChunks.push(newChunk);
+                }
+                stack.push(newChunk);
+                currentChunk = newChunk;
+                contentBuffer = line + '\n';
+            }
+            else {
+                // 非标题行，添加到当前块
+                if (currentChunk) {
+                    contentBuffer += line + '\n';
+                }
+                else {
+                    // 文档开头没有标题的部分
+                    // 创建一个 level 0 的块
+                    if (line.trim().length > 0 && rootChunks.length === 0) {
+                        const introChunk = {
+                            id: this.generateChunkId(),
+                            level: 0,
+                            heading: '',
+                            content: '',
+                            path: [],
+                            startOffset: 0,
+                            endOffset: 0,
+                            children: [],
+                        };
+                        rootChunks.push(introChunk);
+                        currentChunk = introChunk;
+                        contentBuffer = line + '\n';
+                    }
+                    else if (currentChunk === null && rootChunks.length > 0) {
+                        // 继续添加到第一个块
+                        currentChunk = rootChunks[0];
+                        contentBuffer = currentChunk.content + line + '\n';
+                    }
+                }
+            }
+            currentOffset += lineLength;
+        }
+        // 保存最后一个块的内容
+        if (currentChunk) {
+            currentChunk.content = contentBuffer.trim();
+            currentChunk.endOffset = currentOffset;
+        }
+        // 更新所有块的 endOffset（递归）
+        this.updateEndOffsets(rootChunks, markdown.length);
+        return rootChunks;
+    }
+    /**
+     * 将树形结构扁平化
+     */
+    flatten(chunks) {
+        const result = [];
+        const walk = (chunk) => {
+            result.push(chunk);
+            for (const child of chunk.children) {
+                walk(child);
+            }
+        };
+        for (const chunk of chunks) {
+            walk(chunk);
+        }
+        return result;
+    }
+    /**
+     * 更新块的 endOffset
+     * 每个块的 endOffset 应该是下一个同级或更高级块的 startOffset - 1
+     * 或者文档末尾
+     */
+    updateEndOffsets(chunks, docLength) {
+        const flat = this.flatten(chunks);
+        for (let i = 0; i < flat.length; i++) {
+            const chunk = flat[i];
+            const nextSiblingOrHigher = this.findNextSiblingOrHigher(flat, i, chunk.level);
+            if (nextSiblingOrHigher) {
+                chunk.endOffset = nextSiblingOrHigher.startOffset;
+            }
+            else {
+                chunk.endOffset = docLength;
+            }
+        }
+    }
+    /**
+     * 找到下一个同级或更高级的块
+     */
+    findNextSiblingOrHigher(flat, startIndex, level) {
+        for (let i = startIndex + 1; i < flat.length; i++) {
+            if (flat[i].level <= level && flat[i].level > 0) {
+                return flat[i];
+            }
+        }
+        return null;
+    }
+    /**
+     * 生成块 ID
+     */
+    generateChunkId() {
+        return `chunk-${(0, crypto_1.randomBytes)(6).toString('hex')}`;
+    }
+}
+exports.HeadingChunker = HeadingChunker;
+//# sourceMappingURL=HeadingChunker.js.map

package/dist/document/HeadingChunker.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"HeadingChunker.js","sourceRoot":"","sources":["../../src/document/HeadingChunker.ts"],"names":[],"mappings":";AAAA;;;;;GAKG;;;AAEH,mCAAqC;AAGrC;;GAEG;AACH,MAAa,cAAc;IACzB;;OAEG;IACI,KAAK,CAAC,QAAgB;QAC3B,MAAM,KAAK,GAAG,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACnC,MAAM,UAAU,GAAgB,EAAE,CAAC;QACnC,MAAM,KAAK,GAAgB,EAAE,CAAC,CAAC,WAAW;QAE1C,IAAI,aAAa,GAAG,CAAC,CAAC;QACtB,IAAI,YAAY,GAAqB,IAAI,CAAC;QAC1C,IAAI,aAAa,GAAG,EAAE,CAAC;QACvB,IAAI,WAAW,GAAG,KAAK,CAAC;QAExB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACtC,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;YACtB,MAAM,UAAU,GAAG,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,6BAA6B;YAE9F,UAAU;YACV,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;gBAClC,WAAW,GAAG,CAAC,WAAW,CAAC;gBAC3B,gBAAgB;gBAChB,IAAI,YAAY,EAAE,CAAC;oBACjB,aAAa,IAAI,IAAI,GAAG,IAAI,CAAC;gBAC/B,CAAC;gBACD,aAAa,IAAI,UAAU,CAAC;gBAC5B,SAAS;YACX,CAAC;YAED,cAAc;YACd,IAAI,WAAW,EAAE,CAAC;gBAChB,IAAI,YAAY,EAAE,CAAC;oBACjB,aAAa,IAAI,IAAI,GAAG,IAAI,CAAC;gBAC/B,CAAC;gBACD,aAAa,IAAI,UAAU,CAAC;gBAC5B,SAAS;YACX,CAAC;YAED,WAAW;YACX,MAAM,YAAY,GAAG,IAAI,CAAC,KAAK,CAAC,mBAAmB,CAAC,CAAC;YAErD,IAAI,YAAY,EAAE,CAAC;gBACjB,kBAAkB;gBAClB,IAAI,YAAY,EAAE,CAAC;oBACjB,YAAY,CAAC,OAAO,GAAG,aAAa,CAAC,IAAI,EAAE,CAAC;oBAC5C,YAAY,CAAC,SAAS,GAAG,aAAa,CAAC;gBACzC,CAAC;gBAED,MAAM,KAAK,GAAG,YAAY,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;gBACrC,MAAM,OAAO,GAAG,YAAY,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;gBAEvC,OAAO;gBACP,MAAM,QAAQ,GAAc;oBAC1B,EAAE,EAAE,IAAI,CAAC,eAAe,EAAE;oBAC1B,KAAK;oBACL,OAAO;oBACP,OAAO,EAAE,EAAE;oBACX,IAAI,EAAE,CAAC,OAAO,CAAC,EAAE,WAAW;oBAC5B,WAAW,EAAE,aAAa;oBAC1B,SAAS,EAAE,aAAa,GAAG,UAAU,EAAE,YAAY;oBACnD,QAAQ,EAAE,EAAE;iBACb,CAAC;gBAEF,SAAS;gBACT,WAAW;gBACX,OAAO,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,KAAK,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI,KAAK,EAAE,CAAC;oBAClE,KAAK,CAAC,GAAG,EAAE,CAAC;gBACd,CAAC;gBAED,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBACrB,MAAM,MAAM,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;oBACvC,QAAQ,CAAC,QAAQ,GAAG,MAAM,CAAC,EAAE,CAAC;oBAC9B,yBAAyB;oBACzB,QAAQ,CAAC,IAAI,GAAG,CAAC,GAAG,MAAM,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;oBAC1C,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;gBACjC,CAAC;qBAAM,CAAC;oBACN,QAAQ,CAAC,IAAI,GAAG,CAAC,OAAO,CAAC,CAAC;oBAC1B,UAAU,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;gBAC5B,CAAC;gBAED,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;gBACrB,YAAY,GAAG,QAAQ,CAAC;gBACxB,aAAa,GAAG,IAAI,GAAG,IAAI,CAAC;YAC9B,CAAC;iBAAM,CAAC;gBACN,cAAc;gBACd,IAAI,YAAY,EAAE,CAAC;oBACjB,aAAa,IAAI,IAAI,GAAG,IAAI,CAAC;gBAC/B,CAAC;qBAAM,CAAC;oBACN,cAAc;oBACd,kBAAkB;oBAClB,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;wBACtD,MAAM,UAAU,GAAc;4BAC5B,EAAE,EAAE,IAAI,CAAC,eAAe,EAAE;4BAC1B,KAAK,EAAE,CAAC;4BACR,OAAO,EAAE,EAAE;4BACX,OAAO,EAAE,EAAE;4BACX,IAAI,EAAE,EAAE;4BACR,WAAW,EAAE,CAAC;4BACd,SAAS,EAAE,CAAC;4BACZ,QAAQ,EAAE,EAAE;yBACb,CAAC;wBACF,UAAU,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;wBAC5B,YAAY,GAAG,UAAU,CAAC;wBAC1B,aAAa,GAAG,IAAI,GAAG,IAAI,CAAC;oBAC9B,CAAC;yBAAM,IAAI,YAAY,KAAK,IAAI,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;wBAC1D,YAAY;wBACZ,YAAY,GAAG,UAAU,CAAC,CAAC,CAAC,CAAC;wBAC7B,aAAa,GAAG,YAAY,CAAC,OAAO,GAAG,IAAI,GAAG,IAAI,CAAC;oBACrD,CAAC;gBACH,CAAC;YACH,CAAC;YAED,aAAa,IAAI,UAAU,CAAC;QAC9B,CAAC;QAED,aAAa;QACb,IAAI,YAAY,EAAE,CAAC;YACjB,YAAY,CAAC,OAAO,GAAG,aAAa,CAAC,IAAI,EAAE,CAAC;YAC5C,YAAY,CAAC,SAAS,GAAG,aAAa,CAAC;QACzC,CAAC;QAED,uBAAuB;QACvB,IAAI,CAAC,gBAAgB,CAAC,UAAU,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC;QAEnD,OAAO,UAAU,CAAC;IACpB,CAAC;IAED;;OAEG;IACI,OAAO,CAAC,MAAmB;QAChC,MAAM,MAAM,GAAgB,EAAE,CAAC;QAE/B,MAAM,IAAI,GAAG,CAAC,KAAgB,EAAQ,EAAE;YACtC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACnB,KAAK,MAAM,KAAK,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;gBACnC,IAAI,CAAC,KAAK,CAAC,CAAC;YACd,CAAC;QACH,CAAC,CAAC;QAEF,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;YAC3B,IAAI,CAAC,KAAK,CAAC,CAAC;QACd,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAED;;;;OAIG;IACK,gBAAgB,CAAC,MAAmB,EAAE,SAAiB;QAC7D,MAAM,IAAI,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;QAElC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACrC,MAAM,KAAK,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;YACtB,MAAM,mBAAmB,GAAG,IAAI,CAAC,uBAAuB,CAAC,IAAI,EAAE,CAAC,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;YAE/E,IAAI,mBAAmB,EAAE,CAAC;gBACxB,KAAK,CAAC,SAAS,GAAG,mBAAmB,CAAC,WAAW,CAAC;YACpD,CAAC;iBAAM,CAAC;gBACN,KAAK,CAAC,SAAS,GAAG,SAAS,CAAC;YAC9B,CAAC;QACH,CAAC;IACH,CAAC;IAED;;OAEG;IACK,uBAAuB,CAC7B,IAAiB,EACjB,UAAkB,EAClB,KAAa;QAEb,KAAK,IAAI,CAAC,GAAG,UAAU,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAClD,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,IAAI,KAAK,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,EAAE,CAAC;gBAChD,OAAO,IAAI,CAAC,CAAC,CAAC,CAAC;YACjB,CAAC;QACH,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC;IAED;;OAEG;IACK,eAAe;QACrB,OAAO,SAAS,IAAA,oBAAW,EAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;IACnD,CAAC;CACF;AA7LD,wCA6LC","sourcesContent":["/**\n * HeadingChunker - 按标题层级分块\n *\n * 将 Markdown 文档按标题层级分块，构建树形结构\n * 每个块记录 startOffset 和 endOffset\n */\n\nimport { randomBytes } from 'crypto';\nimport type { Chunker, TextChunk } from './Chunker';\n\n/**\n * 按标题层级分块的实现\n */\nexport class HeadingChunker implements Chunker {\n  /**\n   * 将 Markdown 分块为树形结构\n   */\n  public chunk(markdown: string): TextChunk[] {\n    const lines = markdown.split('\\n');\n    const rootChunks: TextChunk[] = [];\n    const stack: TextChunk[] = []; // 用于构建树形结构\n\n    let currentOffset = 0;\n    let currentChunk: TextChunk | null = null;\n    let contentBuffer = '';\n    let inCodeBlock = false;\n\n    for (let i = 0; i < lines.length; i++) {\n      const line = lines[i];\n      const lineLength = line.length + (i < lines.length - 1 ? 1 : 0); // +1 for \\n except last line\n\n      // 检查代码块边界\n      if (line.trim().startsWith('```')) {\n        inCodeBlock = !inCodeBlock;\n        // 代码块边界添加到当前块内容\n        if (currentChunk) {\n          contentBuffer += line + '\\n';\n        }\n        currentOffset += lineLength;\n        continue;\n      }\n\n      // 在代码块内，不解析标题\n      if (inCodeBlock) {\n        if (currentChunk) {\n          contentBuffer += line + '\\n';\n        }\n        currentOffset += lineLength;\n        continue;\n      }\n\n      // 检查是否是标题行\n      const headingMatch = line.match(/^(#{1,6})\\s+(.+)$/);\n\n      if (headingMatch) {\n        // 发现新标题，先保存之前的块内容\n        if (currentChunk) {\n          currentChunk.content = contentBuffer.trim();\n          currentChunk.endOffset = currentOffset;\n        }\n\n        const level = headingMatch[1].length;\n        const heading = headingMatch[2].trim();\n\n        // 创建新块\n        const newChunk: TextChunk = {\n          id: this.generateChunkId(),\n          level,\n          heading,\n          content: '',\n          path: [heading], // 初始路径包含自己\n          startOffset: currentOffset,\n          endOffset: currentOffset + lineLength, // 临时值，后面会更新\n          children: [],\n        };\n\n        // 构建树形结构\n        // 找到合适的父节点\n        while (stack.length > 0 && stack[stack.length - 1].level >= level) {\n          stack.pop();\n        }\n\n        if (stack.length > 0) {\n          const parent = stack[stack.length - 1];\n          newChunk.parentId = parent.id;\n          // path 包含父级的 path 加上当前标题\n          newChunk.path = [...parent.path, heading];\n          parent.children.push(newChunk);\n        } else {\n          newChunk.path = [heading];\n          rootChunks.push(newChunk);\n        }\n\n        stack.push(newChunk);\n        currentChunk = newChunk;\n        contentBuffer = line + '\\n';\n      } else {\n        // 非标题行，添加到当前块\n        if (currentChunk) {\n          contentBuffer += line + '\\n';\n        } else {\n          // 文档开头没有标题的部分\n          // 创建一个 level 0 的块\n          if (line.trim().length > 0 && rootChunks.length === 0) {\n            const introChunk: TextChunk = {\n              id: this.generateChunkId(),\n              level: 0,\n              heading: '',\n              content: '',\n              path: [],\n              startOffset: 0,\n              endOffset: 0,\n              children: [],\n            };\n            rootChunks.push(introChunk);\n            currentChunk = introChunk;\n            contentBuffer = line + '\\n';\n          } else if (currentChunk === null && rootChunks.length > 0) {\n            // 继续添加到第一个块\n            currentChunk = rootChunks[0];\n            contentBuffer = currentChunk.content + line + '\\n';\n          }\n        }\n      }\n\n      currentOffset += lineLength;\n    }\n\n    // 保存最后一个块的内容\n    if (currentChunk) {\n      currentChunk.content = contentBuffer.trim();\n      currentChunk.endOffset = currentOffset;\n    }\n\n    // 更新所有块的 endOffset（递归）\n    this.updateEndOffsets(rootChunks, markdown.length);\n\n    return rootChunks;\n  }\n\n  /**\n   * 将树形结构扁平化\n   */\n  public flatten(chunks: TextChunk[]): TextChunk[] {\n    const result: TextChunk[] = [];\n\n    const walk = (chunk: TextChunk): void => {\n      result.push(chunk);\n      for (const child of chunk.children) {\n        walk(child);\n      }\n    };\n\n    for (const chunk of chunks) {\n      walk(chunk);\n    }\n\n    return result;\n  }\n\n  /**\n   * 更新块的 endOffset\n   * 每个块的 endOffset 应该是下一个同级或更高级块的 startOffset - 1\n   * 或者文档末尾\n   */\n  private updateEndOffsets(chunks: TextChunk[], docLength: number): void {\n    const flat = this.flatten(chunks);\n\n    for (let i = 0; i < flat.length; i++) {\n      const chunk = flat[i];\n      const nextSiblingOrHigher = this.findNextSiblingOrHigher(flat, i, chunk.level);\n\n      if (nextSiblingOrHigher) {\n        chunk.endOffset = nextSiblingOrHigher.startOffset;\n      } else {\n        chunk.endOffset = docLength;\n      }\n    }\n  }\n\n  /**\n   * 找到下一个同级或更高级的块\n   */\n  private findNextSiblingOrHigher(\n    flat: TextChunk[],\n    startIndex: number,\n    level: number,\n  ): TextChunk | null {\n    for (let i = startIndex + 1; i < flat.length; i++) {\n      if (flat[i].level <= level && flat[i].level > 0) {\n        return flat[i];\n      }\n    }\n    return null;\n  }\n\n  /**\n   * 生成块 ID\n   */\n  private generateChunkId(): string {\n    return `chunk-${randomBytes(6).toString('hex')}`;\n  }\n}\n"]}

package/dist/document/JinaDocumentParser.d.ts

@@ -0,0 +1,40 @@
+/**
+ * JinaDocumentParser - 基于 JINA Reader API 的文档解析器
+ *
+ * 使用 JINA Reader API (https://r.jina.ai/) 将各种格式的文档转换为 Markdown
+ * 支持：PDF, Office 文档, HTML, 图片等
+ *
+ * @see https://jina.ai/reader/
+ */
+import type { DocumentParser, ParsedDocument, ParseOptions } from './DocumentParser';
+export interface JinaDocumentParserOptions {
+    /** JINA API Key */
+    apiKey: string;
+    /** 自定义 API 端点（默认 https://r.jina.ai） */
+    baseUrl?: string;
+    /** 默认超时时间（毫秒） */
+    defaultTimeout?: number;
+}
+/**
+ * JINA Reader API 文档解析器
+ */
+export declare class JinaDocumentParser implements DocumentParser {
+    protected readonly logger: import("global-logger-factory").Logger<unknown>;
+    private readonly apiKey;
+    private readonly baseUrl;
+    private readonly defaultTimeout;
+    constructor(options: JinaDocumentParserOptions);
+    /**
+     * 解析文档为 Markdown
+     */
+    parse(url: string, options?: ParseOptions): Promise<ParsedDocument>;
+    /**
+     * 检查是否支持该 URL
+     * JINA Reader 支持大多数 URL
+     */
+    supports(url: string): boolean;
+    /**
+     * 从 Markdown 内容中提取元数据
+     */
+    private extractMetadata;
+}