@jingyi0605/codingns 0.8.5 → 0.9.5

package/dist/server/modules/affairs-indexer/core/src/parser/document-parser.d.ts

@@ -0,0 +1,19 @@
+import type { RuntimeConfig } from "../../../contracts/src/index.js";
+import type { ParsedDocument as ParsedDocumentResult } from "./plain-text-parser.js";
+import type { ParseSkip } from "./parser-adapter.js";
+import { ParserRouter } from "./parser-router.js";
+import type { ParserAdapter } from "./parser-adapter.js";
+/**
+ * 统一解析入口。
+ * 第二阶段改为依赖 ParserRouter，避免解析策略继续散在索引流程里。
+ */
+export declare class DocumentParser {
+    private readonly router;
+    constructor(options: {
+        config: RuntimeConfig;
+        router?: ParserRouter;
+        adapters?: ParserAdapter[];
+    });
+    parse(filePath: string, signal?: AbortSignal): Promise<ParsedDocumentResult>;
+    parseWithOutcome(filePath: string, signal?: AbortSignal): Promise<ParsedDocumentResult | ParseSkip>;
+}

package/dist/server/modules/affairs-indexer/core/src/parser/document-parser.js

@@ -0,0 +1,37 @@
+import { ParserRouter, createDefaultParserAdapters } from "./parser-router.js";
+import { throwIfAborted } from "../utils/abort.js";
+/**
+ * 统一解析入口。
+ * 第二阶段改为依赖 ParserRouter，避免解析策略继续散在索引流程里。
+ */
+export class DocumentParser {
+    router;
+    constructor(options) {
+        this.router = options.router ?? new ParserRouter(options.adapters ?? createDefaultParserAdapters(), {
+            disabledExtensions: options.config.disabledParserExtensions,
+        });
+    }
+    async parse(filePath, signal) {
+        throwIfAborted(signal, "事务文档库解析已取消");
+        const { adapter, extension } = await this.router.resolveForFile(filePath);
+        throwIfAborted(signal, "事务文档库解析已取消");
+        const result = await adapter.parse({
+            filePath,
+            extension,
+        });
+        if ("kind" in result && result.kind === "skip") {
+            throw new Error("parse() 不支持 skip 结果，请改用 parseWithOutcome()");
+        }
+        return result;
+    }
+    async parseWithOutcome(filePath, signal) {
+        throwIfAborted(signal, "事务文档库解析已取消");
+        const { adapter, extension } = await this.router.resolveForFile(filePath);
+        throwIfAborted(signal, "事务文档库解析已取消");
+        return await adapter.parse({
+            filePath,
+            extension,
+        });
+    }
+}
+//# sourceMappingURL=document-parser.js.map

package/dist/server/modules/affairs-indexer/core/src/parser/document-parser.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"document-parser.js","sourceRoot":"","sources":["../../../../../../../src/modules/affairs-indexer/core/src/parser/document-parser.ts"],"names":[],"mappings":"AAGA,OAAO,EAAE,YAAY,EAAE,2BAA2B,EAAE,MAAM,oBAAoB,CAAC;AAE/E,OAAO,EAAE,cAAc,EAAE,MAAM,mBAAmB,CAAC;AAEnD;;;GAGG;AACH,MAAM,OAAO,cAAc;IACR,MAAM,CAAe;IAEtC,YAAY,OAAqF;QAC/F,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,MAAM,IAAI,IAAI,YAAY,CAAC,OAAO,CAAC,QAAQ,IAAI,2BAA2B,EAAE,EAAE;YAClG,kBAAkB,EAAE,OAAO,CAAC,MAAM,CAAC,wBAAwB;SAC5D,CAAC,CAAC;IACL,CAAC;IAED,KAAK,CAAC,KAAK,CAAC,QAAgB,EAAE,MAAoB;QAChD,cAAc,CAAC,MAAM,EAAE,YAAY,CAAC,CAAC;QACrC,MAAM,EAAE,OAAO,EAAE,SAAS,EAAE,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,cAAc,CAAC,QAAQ,CAAC,CAAC;QAC1E,cAAc,CAAC,MAAM,EAAE,YAAY,CAAC,CAAC;QACrC,MAAM,MAAM,GAAG,MAAM,OAAO,CAAC,KAAK,CAAC;YACjC,QAAQ;YACR,SAAS;SACV,CAAC,CAAC;QACH,IAAI,MAAM,IAAI,MAAM,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC/C,MAAM,IAAI,KAAK,CAAC,4CAA4C,CAAC,CAAC;QAChE,CAAC;QACD,OAAO,MAA8B,CAAC;IACxC,CAAC;IAED,KAAK,CAAC,gBAAgB,CAAC,QAAgB,EAAE,MAAoB;QAC3D,cAAc,CAAC,MAAM,EAAE,YAAY,CAAC,CAAC;QACrC,MAAM,EAAE,OAAO,EAAE,SAAS,EAAE,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,cAAc,CAAC,QAAQ,CAAC,CAAC;QAC1E,cAAc,CAAC,MAAM,EAAE,YAAY,CAAC,CAAC;QACrC,OAAO,MAAM,OAAO,CAAC,KAAK,CAAC;YACzB,QAAQ;YACR,SAAS;SACV,CAAqC,CAAC;IACzC,CAAC;CACF"}

package/dist/server/modules/affairs-indexer/core/src/parser/docx-parser-adapter.d.ts

@@ -0,0 +1,7 @@
+import { BaseComplexParserAdapter } from "./base-complex-parser-adapter.js";
+import type { ParseInput, ParsedDocumentPayload } from "./parser-adapter.js";
+export declare class DocxParserAdapter extends BaseComplexParserAdapter {
+    readonly name = "docx_parser";
+    supports(ext: string): boolean;
+    protected parseComplex(input: ParseInput): Promise<ParsedDocumentPayload>;
+}

package/dist/server/modules/affairs-indexer/core/src/parser/docx-parser-adapter.js

@@ -0,0 +1,123 @@
+import fs from "node:fs";
+import path from "node:path";
+import { AppError, APP_ERROR_CODES } from "../../../contracts/src/index.js";
+import { BaseComplexParserAdapter } from "./base-complex-parser-adapter.js";
+import { decodeXmlEntities, extractXmlAttributes, readZipEntries, readZipText, shortSummary, } from "./openxml-utils.js";
+const SUPPORTED_DOCX_EXTENSIONS = new Set([".docx"]);
+function stripXmlTags(raw) {
+    return raw.replace(/<[^>]+>/g, "");
+}
+function parseParagraphText(paragraphXml) {
+    const parts = [];
+    const textPattern = /<w:t\b[^>]*>([\s\S]*?)<\/w:t>/g;
+    for (const match of paragraphXml.matchAll(textPattern)) {
+        parts.push(decodeXmlEntities(match[1]));
+    }
+    const tabCount = [...paragraphXml.matchAll(/<w:tab\b[^>]*\/>/g)].length;
+    for (let index = 0; index < tabCount; index += 1) {
+        parts.push("\t");
+    }
+    const breakCount = [...paragraphXml.matchAll(/<w:br\b[^>]*\/>/g)].length;
+    for (let index = 0; index < breakCount; index += 1) {
+        parts.push("\n");
+    }
+    return parts.join("").replace(/\s+\n/g, "\n").replace(/\n\s+/g, "\n").trim();
+}
+function parseParagraphStyle(paragraphXml) {
+    const match = paragraphXml.match(/<w:pStyle\b([^>]*)\/>/);
+    if (!match) {
+        return null;
+    }
+    const attributes = extractXmlAttributes(match[1]);
+    return attributes["w:val"] ?? null;
+}
+function isHeadingStyle(style) {
+    return Boolean(style && /^Heading[1-6]$/i.test(style));
+}
+function parseDocumentBlocks(documentXml) {
+    const blocks = [];
+    const paragraphPattern = /<w:p\b[^>]*>([\s\S]*?)<\/w:p>/g;
+    for (const match of documentXml.matchAll(paragraphPattern)) {
+        const rawParagraphXml = match[0];
+        const text = parseParagraphText(rawParagraphXml);
+        if (!text) {
+            continue;
+        }
+        const style = parseParagraphStyle(rawParagraphXml);
+        const kind = isHeadingStyle(style) ? "heading" : "paragraph";
+        blocks.push({
+            kind,
+            text,
+            metadata: style
+                ? {
+                    style,
+                }
+                : undefined,
+        });
+    }
+    return blocks;
+}
+function parseCoreTitle(coreXml) {
+    if (!coreXml) {
+        return null;
+    }
+    const match = coreXml.match(/<dc:title\b[^>]*>([\s\S]*?)<\/dc:title>/);
+    if (!match) {
+        return null;
+    }
+    const title = decodeXmlEntities(stripXmlTags(match[1])).trim();
+    return title || null;
+}
+export class DocxParserAdapter extends BaseComplexParserAdapter {
+    name = "docx_parser";
+    supports(ext) {
+        return SUPPORTED_DOCX_EXTENSIONS.has(ext.toLowerCase());
+    }
+    async parseComplex(input) {
+        let fileBuffer;
+        try {
+            fileBuffer = fs.readFileSync(input.filePath);
+        }
+        catch (error) {
+            throw new AppError(`DOCX 文件读取失败：${path.basename(input.filePath)}`, APP_ERROR_CODES.PARSER_COMPLEX_CONTENT_UNREADABLE, { cause: error });
+        }
+        const entries = readZipEntries(fileBuffer, "DOCX");
+        const documentXml = readZipText(entries, "word/document.xml", "DOCX");
+        const coreXml = readZipText(entries, "docProps/core.xml", "DOCX", false);
+        const blocks = parseDocumentBlocks(documentXml);
+        if (blocks.length === 0) {
+            throw new AppError("DOCX 中没有可读取的正文段落", APP_ERROR_CODES.PARSER_COMPLEX_CONTENT_UNREADABLE);
+        }
+        const text = blocks.map(block => block.text).join("\n\n").trim();
+        if (!text) {
+            throw new AppError("DOCX 文本内容为空", APP_ERROR_CODES.PARSER_COMPLEX_CONTENT_UNREADABLE);
+        }
+        const headingCount = blocks.filter(block => block.kind === "heading").length;
+        const structuredBlocks = blocks.map(block => ({
+            kind: block.kind,
+            text: block.text,
+            metadata: block.metadata,
+        }));
+        const title = parseCoreTitle(coreXml) ?? blocks.find(block => block.kind === "heading")?.text
+            ?? path.basename(input.filePath, input.extension);
+        return {
+            title,
+            text,
+            summary: shortSummary(text),
+            parser: "docx",
+            metadata: {
+                adapter: this.name,
+                paragraphCount: blocks.length,
+                headingCount,
+            },
+            structured: {
+                blocks: structuredBlocks,
+                stats: {
+                    paragraphCount: blocks.length,
+                    headingCount,
+                },
+            },
+        };
+    }
+}
+//# sourceMappingURL=docx-parser-adapter.js.map

package/dist/server/modules/affairs-indexer/core/src/parser/docx-parser-adapter.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"docx-parser-adapter.js","sourceRoot":"","sources":["../../../../../../../src/modules/affairs-indexer/core/src/parser/docx-parser-adapter.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,SAAS,CAAC;AACzB,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,QAAQ,EAAE,eAAe,EAAE,MAAM,iCAAiC,CAAC;AAC5E,OAAO,EAAE,wBAAwB,EAAE,MAAM,kCAAkC,CAAC;AAC5E,OAAO,EACL,iBAAiB,EACjB,oBAAoB,EACpB,cAAc,EACd,WAAW,EACX,YAAY,GACb,MAAM,oBAAoB,CAAC;AAG5B,MAAM,yBAAyB,GAAG,IAAI,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC;AAQrD,SAAS,YAAY,CAAC,GAAW;IAC/B,OAAO,GAAG,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC;AACrC,CAAC;AAED,SAAS,kBAAkB,CAAC,YAAoB;IAC9C,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,MAAM,WAAW,GAAG,gCAAgC,CAAC;IACrD,KAAK,MAAM,KAAK,IAAI,YAAY,CAAC,QAAQ,CAAC,WAAW,CAAC,EAAE,CAAC;QACvD,KAAK,CAAC,IAAI,CAAC,iBAAiB,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1C,CAAC;IAED,MAAM,QAAQ,GAAG,CAAC,GAAG,YAAY,CAAC,QAAQ,CAAC,mBAAmB,CAAC,CAAC,CAAC,MAAM,CAAC;IACxE,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,QAAQ,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QACjD,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACnB,CAAC;IAED,MAAM,UAAU,GAAG,CAAC,GAAG,YAAY,CAAC,QAAQ,CAAC,kBAAkB,CAAC,CAAC,CAAC,MAAM,CAAC;IACzE,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,UAAU,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QACnD,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACnB,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,OAAO,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC,OAAO,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;AAC/E,CAAC;AAED,SAAS,mBAAmB,CAAC,YAAoB;IAC/C,MAAM,KAAK,GAAG,YAAY,CAAC,KAAK,CAAC,uBAAuB,CAAC,CAAC;IAC1D,IAAI,CAAC,KAAK,EAAE,CAAC;QACX,OAAO,IAAI,CAAC;IACd,CAAC;IACD,MAAM,UAAU,GAAG,oBAAoB,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAClD,OAAO,UAAU,CAAC,OAAO,CAAC,IAAI,IAAI,CAAC;AACrC,CAAC;AAED,SAAS,cAAc,CAAC,KAAoB;IAC1C,OAAO,OAAO,CAAC,KAAK,IAAI,iBAAiB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;AACzD,CAAC;AAED,SAAS,mBAAmB,CAAC,WAAmB;IAC9C,MAAM,MAAM,GAAqB,EAAE,CAAC;IACpC,MAAM,gBAAgB,GAAG,gCAAgC,CAAC;IAC1D,KAAK,MAAM,KAAK,IAAI,WAAW,CAAC,QAAQ,CAAC,gBAAgB,CAAC,EAAE,CAAC;QAC3D,MAAM,eAAe,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QACjC,MAAM,IAAI,GAAG,kBAAkB,CAAC,eAAe,CAAC,CAAC;QACjD,IAAI,CAAC,IAAI,EAAE,CAAC;YACV,SAAS;QACX,CAAC;QACD,MAAM,KAAK,GAAG,mBAAmB,CAAC,eAAe,CAAC,CAAC;QACnD,MAAM,IAAI,GAAG,cAAc,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,WAAW,CAAC;QAC7D,MAAM,CAAC,IAAI,CAAC;YACV,IAAI;YACJ,IAAI;YACJ,QAAQ,EAAE,KAAK;gBACb,CAAC,CAAC;oBACA,KAAK;iBACN;gBACD,CAAC,CAAC,SAAS;SACd,CAAC,CAAC;IACL,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,cAAc,CAAC,OAAsB;IAC5C,IAAI,CAAC,OAAO,EAAE,CAAC;QACb,OAAO,IAAI,CAAC;IACd,CAAC;IACD,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,yCAAyC,CAAC,CAAC;IACvE,IAAI,CAAC,KAAK,EAAE,CAAC;QACX,OAAO,IAAI,CAAC;IACd,CAAC;IACD,MAAM,KAAK,GAAG,iBAAiB,CAAC,YAAY,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IAC/D,OAAO,KAAK,IAAI,IAAI,CAAC;AACvB,CAAC;AAED,MAAM,OAAO,iBAAkB,SAAQ,wBAAwB;IACpD,IAAI,GAAG,aAAa,CAAC;IAE9B,QAAQ,CAAC,GAAW;QAClB,OAAO,yBAAyB,CAAC,GAAG,CAAC,GAAG,CAAC,WAAW,EAAE,CAAC,CAAC;IAC1D,CAAC;IAES,KAAK,CAAC,YAAY,CAAC,KAAiB;QAC5C,IAAI,UAAkB,CAAC;QACvB,IAAI,CAAC;YACH,UAAU,GAAG,EAAE,CAAC,YAAY,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;QAC/C,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,MAAM,IAAI,QAAQ,CAChB,eAAe,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,EAC9C,eAAe,CAAC,iCAAiC,EACjD,EAAE,KAAK,EAAE,KAAK,EAAE,CACjB,CAAC;QACJ,CAAC;QAED,MAAM,OAAO,GAAG,cAAc,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC;QACnD,MAAM,WAAW,GAAG,WAAW,CAAC,OAAO,EAAE,mBAAmB,EAAE,MAAM,CAAC,CAAC;QACtE,MAAM,OAAO,GAAG,WAAW,CAAC,OAAO,EAAE,mBAAmB,EAAE,MAAM,EAAE,KAAK,CAAC,CAAC;QACzE,MAAM,MAAM,GAAG,mBAAmB,CAAC,WAAW,CAAC,CAAC;QAEhD,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACxB,MAAM,IAAI,QAAQ,CAChB,kBAAkB,EAClB,eAAe,CAAC,iCAAiC,CAClD,CAAC;QACJ,CAAC;QAED,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAAC;QACjE,IAAI,CAAC,IAAI,EAAE,CAAC;YACV,MAAM,IAAI,QAAQ,CAChB,aAAa,EACb,eAAe,CAAC,iCAAiC,CAClD,CAAC;QACJ,CAAC;QAED,MAAM,YAAY,GAAG,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,SAAS,CAAC,CAAC,MAAM,CAAC;QAC7E,MAAM,gBAAgB,GAAsB,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;YAC/D,IAAI,EAAE,KAAK,CAAC,IAAI;YAChB,IAAI,EAAE,KAAK,CAAC,IAAI;YAChB,QAAQ,EAAE,KAAK,CAAC,QAAQ;SACzB,CAAC,CAAC,CAAC;QACJ,MAAM,KAAK,GAAG,cAAc,CAAC,OAAO,CAAC,IAAI,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,SAAS,CAAC,EAAE,IAAI;eACxF,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,QAAQ,EAAE,KAAK,CAAC,SAAS,CAAC,CAAC;QAEpD,OAAO;YACL,KAAK;YACL,IAAI;YACJ,OAAO,EAAE,YAAY,CAAC,IAAI,CAAC;YAC3B,MAAM,EAAE,MAAM;YACd,QAAQ,EAAE;gBACR,OAAO,EAAE,IAAI,CAAC,IAAI;gBAClB,cAAc,EAAE,MAAM,CAAC,MAAM;gBAC7B,YAAY;aACb;YACD,UAAU,EAAE;gBACV,MAAM,EAAE,gBAAgB;gBACxB,KAAK,EAAE;oBACL,cAAc,EAAE,MAAM,CAAC,MAAM;oBAC7B,YAAY;iBACb;aACF;SACF,CAAC;IACJ,CAAC;CACF"}

package/dist/server/modules/affairs-indexer/core/src/parser/openxml-utils.d.ts

@@ -0,0 +1,8 @@
+export declare function normalizeWhitespace(text: string): string;
+export declare function shortSummary(text: string, limit?: number): string;
+export declare function decodeXmlEntities(text: string): string;
+export declare function extractXmlAttributes(raw: string): Record<string, string>;
+export declare function readZipEntries(buffer: Buffer, formatLabel: string): Map<string, Buffer>;
+export declare function readZipText(entries: Map<string, Buffer>, entryPath: string, formatLabel: string): string;
+export declare function readZipText(entries: Map<string, Buffer>, entryPath: string, formatLabel: string, required: true): string;
+export declare function readZipText(entries: Map<string, Buffer>, entryPath: string, formatLabel: string, required: false): string | null;

package/dist/server/modules/affairs-indexer/core/src/parser/openxml-utils.js

@@ -0,0 +1,111 @@
+import { inflateRawSync } from "node:zlib";
+import { AppError, APP_ERROR_CODES } from "../../../contracts/src/index.js";
+const EOCD_SIGNATURE = 0x06054b50;
+const CENTRAL_DIRECTORY_SIGNATURE = 0x02014b50;
+const LOCAL_FILE_SIGNATURE = 0x04034b50;
+const ZIP_METHOD_STORED = 0;
+const ZIP_METHOD_DEFLATE = 8;
+export function normalizeWhitespace(text) {
+    return text.replace(/\s+/g, " ").trim();
+}
+export function shortSummary(text, limit = 180) {
+    const normalized = normalizeWhitespace(text);
+    if (normalized.length <= limit) {
+        return normalized;
+    }
+    return `${normalized.slice(0, limit - 1)}…`;
+}
+export function decodeXmlEntities(text) {
+    return text
+        .replace(/&lt;/g, "<")
+        .replace(/&gt;/g, ">")
+        .replace(/&quot;/g, "\"")
+        .replace(/&apos;/g, "'")
+        .replace(/&amp;/g, "&");
+}
+export function extractXmlAttributes(raw) {
+    const attributes = {};
+    const pattern = /([A-Za-z_][\w:.-]*)="([^"]*)"/g;
+    for (const match of raw.matchAll(pattern)) {
+        attributes[match[1]] = decodeXmlEntities(match[2]);
+    }
+    return attributes;
+}
+function findEndOfCentralDirectory(buffer, formatLabel) {
+    const minimumOffset = Math.max(0, buffer.length - 65557);
+    for (let offset = buffer.length - 22; offset >= minimumOffset; offset -= 1) {
+        if (buffer.readUInt32LE(offset) === EOCD_SIGNATURE) {
+            return offset;
+        }
+    }
+    throw new AppError(`${formatLabel} 文件不是合法的 ZIP 容器`, APP_ERROR_CODES.PARSER_COMPLEX_CONTENT_UNREADABLE);
+}
+function readZipRecords(buffer, formatLabel) {
+    const eocdOffset = findEndOfCentralDirectory(buffer, formatLabel);
+    const entryCount = buffer.readUInt16LE(eocdOffset + 10);
+    const centralDirectoryOffset = buffer.readUInt32LE(eocdOffset + 16);
+    const records = [];
+    let cursor = centralDirectoryOffset;
+    for (let index = 0; index < entryCount; index += 1) {
+        if (cursor + 46 > buffer.length || buffer.readUInt32LE(cursor) !== CENTRAL_DIRECTORY_SIGNATURE) {
+            throw new AppError(`${formatLabel} ZIP 中央目录损坏`, APP_ERROR_CODES.PARSER_COMPLEX_CONTENT_UNREADABLE);
+        }
+        const compressionMethod = buffer.readUInt16LE(cursor + 10);
+        const compressedSize = buffer.readUInt32LE(cursor + 20);
+        const fileNameLength = buffer.readUInt16LE(cursor + 28);
+        const extraFieldLength = buffer.readUInt16LE(cursor + 30);
+        const commentLength = buffer.readUInt16LE(cursor + 32);
+        const localHeaderOffset = buffer.readUInt32LE(cursor + 42);
+        const fileNameStart = cursor + 46;
+        const fileNameEnd = fileNameStart + fileNameLength;
+        const entryPath = buffer.subarray(fileNameStart, fileNameEnd).toString("utf-8");
+        records.push({
+            path: entryPath,
+            compressionMethod,
+            compressedSize,
+            localHeaderOffset,
+        });
+        cursor = fileNameEnd + extraFieldLength + commentLength;
+    }
+    return records;
+}
+function extractZipEntry(buffer, record, formatLabel) {
+    const offset = record.localHeaderOffset;
+    if (offset + 30 > buffer.length || buffer.readUInt32LE(offset) !== LOCAL_FILE_SIGNATURE) {
+        throw new AppError(`${formatLabel} ZIP 本地头损坏：${record.path}`, APP_ERROR_CODES.PARSER_COMPLEX_CONTENT_UNREADABLE);
+    }
+    const fileNameLength = buffer.readUInt16LE(offset + 26);
+    const extraFieldLength = buffer.readUInt16LE(offset + 28);
+    const dataStart = offset + 30 + fileNameLength + extraFieldLength;
+    const dataEnd = dataStart + record.compressedSize;
+    if (dataEnd > buffer.length) {
+        throw new AppError(`${formatLabel} ZIP 条目越界：${record.path}`, APP_ERROR_CODES.PARSER_COMPLEX_CONTENT_UNREADABLE);
+    }
+    const compressed = buffer.subarray(dataStart, dataEnd);
+    if (record.compressionMethod === ZIP_METHOD_STORED) {
+        return Buffer.from(compressed);
+    }
+    if (record.compressionMethod === ZIP_METHOD_DEFLATE) {
+        return inflateRawSync(compressed);
+    }
+    throw new AppError(`${formatLabel} 使用了当前未支持的 ZIP 压缩方法：${record.compressionMethod}`, APP_ERROR_CODES.PARSER_COMPLEX_FORMAT_UNSUPPORTED);
+}
+export function readZipEntries(buffer, formatLabel) {
+    const records = readZipRecords(buffer, formatLabel);
+    const entries = new Map();
+    for (const record of records) {
+        entries.set(record.path, extractZipEntry(buffer, record, formatLabel));
+    }
+    return entries;
+}
+export function readZipText(entries, entryPath, formatLabel, required = true) {
+    const content = entries.get(entryPath);
+    if (!content) {
+        if (!required) {
+            return null;
+        }
+        throw new AppError(`${formatLabel} 缺少必需条目：${entryPath}`, APP_ERROR_CODES.PARSER_COMPLEX_CONTENT_UNREADABLE);
+    }
+    return content.toString("utf-8");
+}
+//# sourceMappingURL=openxml-utils.js.map

package/dist/server/modules/affairs-indexer/core/src/parser/openxml-utils.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"openxml-utils.js","sourceRoot":"","sources":["../../../../../../../src/modules/affairs-indexer/core/src/parser/openxml-utils.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,cAAc,EAAE,MAAM,WAAW,CAAC;AAC3C,OAAO,EAAE,QAAQ,EAAE,eAAe,EAAE,MAAM,iCAAiC,CAAC;AAE5E,MAAM,cAAc,GAAG,UAAU,CAAC;AAClC,MAAM,2BAA2B,GAAG,UAAU,CAAC;AAC/C,MAAM,oBAAoB,GAAG,UAAU,CAAC;AACxC,MAAM,iBAAiB,GAAG,CAAC,CAAC;AAC5B,MAAM,kBAAkB,GAAG,CAAC,CAAC;AAS7B,MAAM,UAAU,mBAAmB,CAAC,IAAY;IAC9C,OAAO,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;AAC1C,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,IAAY,EAAE,KAAK,GAAG,GAAG;IACpD,MAAM,UAAU,GAAG,mBAAmB,CAAC,IAAI,CAAC,CAAC;IAC7C,IAAI,UAAU,CAAC,MAAM,IAAI,KAAK,EAAE,CAAC;QAC/B,OAAO,UAAU,CAAC;IACpB,CAAC;IACD,OAAO,GAAG,UAAU,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,GAAG,CAAC,CAAC,GAAG,CAAC;AAC9C,CAAC;AAED,MAAM,UAAU,iBAAiB,CAAC,IAAY;IAC5C,OAAO,IAAI;SACR,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC;SACrB,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC;SACrB,OAAO,CAAC,SAAS,EAAE,IAAI,CAAC;SACxB,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC;SACvB,OAAO,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAC;AAC5B,CAAC;AAED,MAAM,UAAU,oBAAoB,CAAC,GAAW;IAC9C,MAAM,UAAU,GAA2B,EAAE,CAAC;IAC9C,MAAM,OAAO,GAAG,gCAAgC,CAAC;IACjD,KAAK,MAAM,KAAK,IAAI,GAAG,CAAC,QAAQ,CAAC,OAAO,CAAC,EAAE,CAAC;QAC1C,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,iBAAiB,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IACrD,CAAC;IACD,OAAO,UAAU,CAAC;AACpB,CAAC;AAED,SAAS,yBAAyB,CAAC,MAAc,EAAE,WAAmB;IACpE,MAAM,aAAa,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,MAAM,CAAC,MAAM,GAAG,KAAK,CAAC,CAAC;IACzD,KAAK,IAAI,MAAM,GAAG,MAAM,CAAC,MAAM,GAAG,EAAE,EAAE,MAAM,IAAI,aAAa,EAAE,MAAM,IAAI,CAAC,EAAE,CAAC;QAC3E,IAAI,MAAM,CAAC,YAAY,CAAC,MAAM,CAAC,KAAK,cAAc,EAAE,CAAC;YACnD,OAAO,MAAM,CAAC;QAChB,CAAC;IACH,CAAC;IAED,MAAM,IAAI,QAAQ,CAChB,GAAG,WAAW,iBAAiB,EAC/B,eAAe,CAAC,iCAAiC,CAClD,CAAC;AACJ,CAAC;AAED,SAAS,cAAc,CAAC,MAAc,EAAE,WAAmB;IACzD,MAAM,UAAU,GAAG,yBAAyB,CAAC,MAAM,EAAE,WAAW,CAAC,CAAC;IAClE,MAAM,UAAU,GAAG,MAAM,CAAC,YAAY,CAAC,UAAU,GAAG,EAAE,CAAC,CAAC;IACxD,MAAM,sBAAsB,GAAG,MAAM,CAAC,YAAY,CAAC,UAAU,GAAG,EAAE,CAAC,CAAC;IACpE,MAAM,OAAO,GAAqB,EAAE,CAAC;IACrC,IAAI,MAAM,GAAG,sBAAsB,CAAC;IAEpC,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,UAAU,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QACnD,IAAI,MAAM,GAAG,EAAE,GAAG,MAAM,CAAC,MAAM,IAAI,MAAM,CAAC,YAAY,CAAC,MAAM,CAAC,KAAK,2BAA2B,EAAE,CAAC;YAC/F,MAAM,IAAI,QAAQ,CAChB,GAAG,WAAW,aAAa,EAC3B,eAAe,CAAC,iCAAiC,CAClD,CAAC;QACJ,CAAC;QAED,MAAM,iBAAiB,GAAG,MAAM,CAAC,YAAY,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC;QAC3D,MAAM,cAAc,GAAG,MAAM,CAAC,YAAY,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC;QACxD,MAAM,cAAc,GAAG,MAAM,CAAC,YAAY,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC;QACxD,MAAM,gBAAgB,GAAG,MAAM,CAAC,YAAY,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC;QAC1D,MAAM,aAAa,GAAG,MAAM,CAAC,YAAY,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC;QACvD,MAAM,iBAAiB,GAAG,MAAM,CAAC,YAAY,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC;QAC3D,MAAM,aAAa,GAAG,MAAM,GAAG,EAAE,CAAC;QAClC,MAAM,WAAW,GAAG,aAAa,GAAG,cAAc,CAAC;QACnD,MAAM,SAAS,GAAG,MAAM,CAAC,QAAQ,CAAC,aAAa,EAAE,WAAW,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC;QAEhF,OAAO,CAAC,IAAI,CAAC;YACX,IAAI,EAAE,SAAS;YACf,iBAAiB;YACjB,cAAc;YACd,iBAAiB;SAClB,CAAC,CAAC;QAEH,MAAM,GAAG,WAAW,GAAG,gBAAgB,GAAG,aAAa,CAAC;IAC1D,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,eAAe,CAAC,MAAc,EAAE,MAAsB,EAAE,WAAmB;IAClF,MAAM,MAAM,GAAG,MAAM,CAAC,iBAAiB,CAAC;IACxC,IAAI,MAAM,GAAG,EAAE,GAAG,MAAM,CAAC,MAAM,IAAI,MAAM,CAAC,YAAY,CAAC,MAAM,CAAC,KAAK,oBAAoB,EAAE,CAAC;QACxF,MAAM,IAAI,QAAQ,CAChB,GAAG,WAAW,cAAc,MAAM,CAAC,IAAI,EAAE,EACzC,eAAe,CAAC,iCAAiC,CAClD,CAAC;IACJ,CAAC;IAED,MAAM,cAAc,GAAG,MAAM,CAAC,YAAY,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC;IACxD,MAAM,gBAAgB,GAAG,MAAM,CAAC,YAAY,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC;IAC1D,MAAM,SAAS,GAAG,MAAM,GAAG,EAAE,GAAG,cAAc,GAAG,gBAAgB,CAAC;IAClE,MAAM,OAAO,GAAG,SAAS,GAAG,MAAM,CAAC,cAAc,CAAC;IAClD,IAAI,OAAO,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC;QAC5B,MAAM,IAAI,QAAQ,CAChB,GAAG,WAAW,aAAa,MAAM,CAAC,IAAI,EAAE,EACxC,eAAe,CAAC,iCAAiC,CAClD,CAAC;IACJ,CAAC;IAED,MAAM,UAAU,GAAG,MAAM,CAAC,QAAQ,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC;IACvD,IAAI,MAAM,CAAC,iBAAiB,KAAK,iBAAiB,EAAE,CAAC;QACnD,OAAO,MAAM,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;IACjC,CAAC;IACD,IAAI,MAAM,CAAC,iBAAiB,KAAK,kBAAkB,EAAE,CAAC;QACpD,OAAO,cAAc,CAAC,UAAU,CAAC,CAAC;IACpC,CAAC;IAED,MAAM,IAAI,QAAQ,CAChB,GAAG,WAAW,uBAAuB,MAAM,CAAC,iBAAiB,EAAE,EAC/D,eAAe,CAAC,iCAAiC,CAClD,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,cAAc,CAAC,MAAc,EAAE,WAAmB;IAChE,MAAM,OAAO,GAAG,cAAc,CAAC,MAAM,EAAE,WAAW,CAAC,CAAC;IACpD,MAAM,OAAO,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC1C,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,OAAO,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,EAAE,eAAe,CAAC,MAAM,EAAE,MAAM,EAAE,WAAW,CAAC,CAAC,CAAC;IACzE,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAeD,MAAM,UAAU,WAAW,CACzB,OAA4B,EAC5B,SAAiB,EACjB,WAAmB,EACnB,QAAQ,GAAG,IAAI;IAEf,MAAM,OAAO,GAAG,OAAO,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;IACvC,IAAI,CAAC,OAAO,EAAE,CAAC;QACb,IAAI,CAAC,QAAQ,EAAE,CAAC;YACd,OAAO,IAAI,CAAC;QACd,CAAC;QACD,MAAM,IAAI,QAAQ,CAChB,GAAG,WAAW,WAAW,SAAS,EAAE,EACpC,eAAe,CAAC,iCAAiC,CAClD,CAAC;IACJ,CAAC;IACD,OAAO,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC;AACnC,CAAC"}

package/dist/server/modules/affairs-indexer/core/src/parser/parser-adapter.d.ts

@@ -0,0 +1,42 @@
+export type ParserAvailability = "available" | "degraded" | "unavailable";
+export interface ParseInput {
+    filePath: string;
+    extension: string;
+}
+export type StructuredBlockKind = "paragraph" | "table" | "page" | "sheet" | "slide" | "heading";
+export interface StructuredBlock {
+    kind: StructuredBlockKind;
+    text?: string;
+    page?: number;
+    sheetName?: string;
+    slideIndex?: number;
+    cells?: string[][];
+    metadata?: Record<string, unknown>;
+}
+export interface StructuredDocument {
+    blocks: StructuredBlock[];
+    stats?: Record<string, number>;
+}
+export interface ParseSkip {
+    kind: "skip";
+    adapter: string;
+    reasonCode: string;
+    extension: string;
+    message: string;
+}
+export interface ParsedDocumentPayload {
+    title: string;
+    text: string;
+    summary: string;
+    parser: string;
+    metadata?: Record<string, unknown>;
+    structured?: StructuredDocument;
+}
+export type ParserRouteKind = "primary" | "fallback";
+export interface ParserAdapter {
+    name: string;
+    routeKind?: ParserRouteKind;
+    supports(ext: string): boolean;
+    availability(): Promise<ParserAvailability>;
+    parse(input: ParseInput): Promise<ParsedDocumentPayload | ParseSkip>;
+}

package/dist/server/modules/affairs-indexer/core/src/parser/parser-adapter.js

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

package/dist/server/modules/affairs-indexer/core/src/parser/parser-adapter.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"parser-adapter.js","sourceRoot":"","sources":["../../../../../../../src/modules/affairs-indexer/core/src/parser/parser-adapter.ts"],"names":[],"mappings":""}

package/dist/server/modules/affairs-indexer/core/src/parser/parser-capability-registry.d.ts

@@ -0,0 +1,18 @@
+import type { RuntimeConfig } from "../../../contracts/src/index.js";
+import type { ParserAdapter, ParserAvailability, ParserRouteKind } from "./parser-adapter.js";
+export type ParserCapabilityStatus = "primary" | "fallback" | "disabled";
+export interface ParserCapabilityRecord {
+    extension: string;
+    adapterName: string;
+    routeKind: ParserRouteKind;
+    availability: ParserAvailability;
+    status: ParserCapabilityStatus;
+}
+export declare class ParserCapabilityRegistry {
+    private readonly adapters;
+    private readonly extensions;
+    private readonly disabledExtensions;
+    constructor(adapters: ParserAdapter[], extensions: string[], disabledExtensions?: Set<string>);
+    static createDefault(config?: Pick<RuntimeConfig, "disabledParserExtensions">): ParserCapabilityRegistry;
+    list(): Promise<ParserCapabilityRecord[]>;
+}

package/dist/server/modules/affairs-indexer/core/src/parser/parser-capability-registry.js

@@ -0,0 +1,91 @@
+import { createDefaultParserAdapters } from "./parser-router.js";
+function normalizeRouteKind(adapter) {
+    return adapter.routeKind ?? "primary";
+}
+function normalizeExtension(value) {
+    const normalized = value.trim().toLowerCase();
+    return normalized.startsWith(".") ? normalized : `.${normalized}`;
+}
+function statusFrom(routeKind, availability) {
+    if (availability === "unavailable") {
+        return "disabled";
+    }
+    return routeKind === "fallback" ? "fallback" : "primary";
+}
+export class ParserCapabilityRegistry {
+    adapters;
+    extensions;
+    disabledExtensions;
+    constructor(adapters, extensions, disabledExtensions = new Set()) {
+        this.adapters = adapters;
+        this.extensions = extensions;
+        this.disabledExtensions = disabledExtensions;
+    }
+    static createDefault(config) {
+        return new ParserCapabilityRegistry(createDefaultParserAdapters(), [
+            ".md",
+            ".mdx",
+            ".txt",
+            ".rtf",
+            ".html",
+            ".htm",
+            ".xml",
+            ".json",
+            ".yaml",
+            ".yml",
+            ".tsv",
+            ".csv",
+            ".xlsx",
+            ".ods",
+            ".et",
+            ".numbers",
+            ".doc",
+            ".docx",
+            ".odt",
+            ".wps",
+            ".pdf",
+            ".ppt",
+            ".pptx",
+            ".odp",
+            ".key",
+            ".xls",
+        ], new Set((config?.disabledParserExtensions ?? []).map(normalizeExtension)));
+    }
+    async list() {
+        const records = [];
+        for (const extension of this.extensions) {
+            const supportingAdapters = this.adapters.filter(adapter => adapter.supports(extension));
+            const isDisabled = this.disabledExtensions.has(extension);
+            if (supportingAdapters.length === 0) {
+                records.push({
+                    extension,
+                    adapterName: "none",
+                    routeKind: "fallback",
+                    availability: "unavailable",
+                    status: "disabled",
+                });
+                continue;
+            }
+            const primaryCandidate = supportingAdapters.find(adapter => normalizeRouteKind(adapter) === "primary");
+            const fallbackCandidate = supportingAdapters.find(adapter => normalizeRouteKind(adapter) === "fallback");
+            const preferred = isDisabled
+                ? (fallbackCandidate ?? primaryCandidate ?? supportingAdapters[0])
+                : (primaryCandidate ?? fallbackCandidate ?? supportingAdapters[0]);
+            const preferredAvailability = await preferred.availability();
+            const preferredRouteKind = normalizeRouteKind(preferred);
+            const availability = isDisabled ? "unavailable" : preferredAvailability;
+            const routeKind = isDisabled && fallbackCandidate
+                ? normalizeRouteKind(fallbackCandidate)
+                : preferredRouteKind;
+            records.push({
+                extension,
+                adapterName: preferred.name,
+                routeKind,
+                availability,
+                status: isDisabled ? "disabled" : statusFrom(routeKind, availability),
+            });
+        }
+        return records;
+    }
+}
+//# sourceMappingURL=parser-capability-registry.js.map

package/dist/server/modules/affairs-indexer/core/src/parser/parser-capability-registry.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"parser-capability-registry.js","sourceRoot":"","sources":["../../../../../../../src/modules/affairs-indexer/core/src/parser/parser-capability-registry.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,2BAA2B,EAAE,MAAM,oBAAoB,CAAC;AAYjE,SAAS,kBAAkB,CAAC,OAAsB;IAChD,OAAO,OAAO,CAAC,SAAS,IAAI,SAAS,CAAC;AACxC,CAAC;AAED,SAAS,kBAAkB,CAAC,KAAa;IACvC,MAAM,UAAU,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IAC9C,OAAO,UAAU,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,IAAI,UAAU,EAAE,CAAC;AACpE,CAAC;AAED,SAAS,UAAU,CAAC,SAA0B,EAAE,YAAgC;IAC9E,IAAI,YAAY,KAAK,aAAa,EAAE,CAAC;QACnC,OAAO,UAAU,CAAC;IACpB,CAAC;IACD,OAAO,SAAS,KAAK,UAAU,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,SAAS,CAAC;AAC3D,CAAC;AAED,MAAM,OAAO,wBAAwB;IAEhB;IACA;IACA;IAHnB,YACmB,QAAyB,EACzB,UAAoB,EACpB,qBAAkC,IAAI,GAAG,EAAE;QAF3C,aAAQ,GAAR,QAAQ,CAAiB;QACzB,eAAU,GAAV,UAAU,CAAU;QACpB,uBAAkB,GAAlB,kBAAkB,CAAyB;IAC3D,CAAC;IAEJ,MAAM,CAAC,aAAa,CAAC,MAAwD;QAC3E,OAAO,IAAI,wBAAwB,CAAC,2BAA2B,EAAE,EAAE;YACjE,KAAK;YACL,MAAM;YACN,MAAM;YACN,MAAM;YACN,OAAO;YACP,MAAM;YACN,MAAM;YACN,OAAO;YACP,OAAO;YACP,MAAM;YACN,MAAM;YACN,MAAM;YACN,OAAO;YACP,MAAM;YACN,KAAK;YACL,UAAU;YACV,MAAM;YACN,OAAO;YACP,MAAM;YACN,MAAM;YACN,MAAM;YACN,MAAM;YACN,OAAO;YACP,MAAM;YACN,MAAM;YACN,MAAM;SACP,EAAE,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,wBAAwB,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,kBAAkB,CAAC,CAAC,CAAC,CAAC;IAChF,CAAC;IAED,KAAK,CAAC,IAAI;QACR,MAAM,OAAO,GAA6B,EAAE,CAAC;QAE7C,KAAK,MAAM,SAAS,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC;YACxC,MAAM,kBAAkB,GAAG,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,OAAO,CAAC,EAAE,CAAC,OAAO,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC;YACxF,MAAM,UAAU,GAAG,IAAI,CAAC,kBAAkB,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;YAC1D,IAAI,kBAAkB,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBACpC,OAAO,CAAC,IAAI,CAAC;oBACX,SAAS;oBACT,WAAW,EAAE,MAAM;oBACnB,SAAS,EAAE,UAAU;oBACrB,YAAY,EAAE,aAAa;oBAC3B,MAAM,EAAE,UAAU;iBACnB,CAAC,CAAC;gBACH,SAAS;YACX,CAAC;YAED,MAAM,gBAAgB,GAAG,kBAAkB,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC,kBAAkB,CAAC,OAAO,CAAC,KAAK,SAAS,CAAC,CAAC;YACvG,MAAM,iBAAiB,GAAG,kBAAkB,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC,kBAAkB,CAAC,OAAO,CAAC,KAAK,UAAU,CAAC,CAAC;YACzG,MAAM,SAAS,GAAG,UAAU;gBAC1B,CAAC,CAAC,CAAC,iBAAiB,IAAI,gBAAgB,IAAI,kBAAkB,CAAC,CAAC,CAAC,CAAC;gBAClE,CAAC,CAAC,CAAC,gBAAgB,IAAI,iBAAiB,IAAI,kBAAkB,CAAC,CAAC,CAAC,CAAC,CAAC;YACrE,MAAM,qBAAqB,GAAG,MAAM,SAAS,CAAC,YAAY,EAAE,CAAC;YAC7D,MAAM,kBAAkB,GAAG,kBAAkB,CAAC,SAAS,CAAC,CAAC;YACzD,MAAM,YAAY,GAAG,UAAU,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,qBAAqB,CAAC;YACxE,MAAM,SAAS,GAAG,UAAU,IAAI,iBAAiB;gBAC/C,CAAC,CAAC,kBAAkB,CAAC,iBAAiB,CAAC;gBACvC,CAAC,CAAC,kBAAkB,CAAC;YAEvB,OAAO,CAAC,IAAI,CAAC;gBACX,SAAS;gBACT,WAAW,EAAE,SAAS,CAAC,IAAI;gBAC3B,SAAS;gBACT,YAAY;gBACZ,MAAM,EAAE,UAAU,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,UAAU,CAAC,SAAS,EAAE,YAAY,CAAC;aACtE,CAAC,CAAC;QACL,CAAC;QAED,OAAO,OAAO,CAAC;IACjB,CAAC;CACF"}

package/dist/server/modules/affairs-indexer/core/src/parser/parser-router.d.ts

@@ -0,0 +1,18 @@
+import type { ParserAdapter } from "./parser-adapter.js";
+export declare function createDefaultParserAdapters(): ParserAdapter[];
+/**
+ * 解析器路由器。
+ * 第二阶段把适配器选择逻辑从 DocumentParser 中拆出来，避免索引流程继续硬编码解析策略。
+ */
+export declare class ParserRouter {
+    private readonly adapters;
+    private readonly disabledExtensions;
+    constructor(adapters: ParserAdapter[], options?: {
+        disabledExtensions?: string[];
+    });
+    listAdapters(): ParserAdapter[];
+    resolveForFile(filePath: string): Promise<{
+        adapter: ParserAdapter;
+        extension: string;
+    }>;
+}

package/dist/server/modules/affairs-indexer/core/src/parser/parser-router.js

@@ -0,0 +1,59 @@
+import path from "node:path";
+import { AppError, APP_ERROR_CODES } from "../../../contracts/src/index.js";
+import { ComplexDocumentSkipAdapter } from "./complex-document-skip-adapter.js";
+import { CsvParserAdapter } from "./csv-parser-adapter.js";
+import { DocxParserAdapter } from "./docx-parser-adapter.js";
+import { PdfParserAdapter } from "./pdf-parser-adapter.js";
+import { PlainTextParserAdapter } from "./plain-text-parser-adapter.js";
+import { PptxParserAdapter } from "./pptx-parser-adapter.js";
+import { XlsxParserAdapter } from "./xlsx-parser-adapter.js";
+function isFallbackAdapter(adapter) {
+    return adapter.routeKind === "fallback";
+}
+export function createDefaultParserAdapters() {
+    return [
+        new PlainTextParserAdapter(),
+        new CsvParserAdapter(),
+        new XlsxParserAdapter(),
+        new DocxParserAdapter(),
+        new PdfParserAdapter(),
+        new PptxParserAdapter(),
+        new ComplexDocumentSkipAdapter(),
+    ];
+}
+/**
+ * 解析器路由器。
+ * 第二阶段把适配器选择逻辑从 DocumentParser 中拆出来，避免索引流程继续硬编码解析策略。
+ */
+export class ParserRouter {
+    adapters;
+    disabledExtensions;
+    constructor(adapters, options = {}) {
+        this.adapters = adapters;
+        this.disabledExtensions = new Set((options.disabledExtensions ?? []).map(item => item.toLowerCase()));
+    }
+    listAdapters() {
+        return [...this.adapters];
+    }
+    async resolveForFile(filePath) {
+        const extension = path.extname(filePath).toLowerCase();
+        const primaryAdapter = this.disabledExtensions.has(extension)
+            ? undefined
+            : this.adapters.find(item => !isFallbackAdapter(item) && item.supports(extension));
+        const fallbackAdapter = this.adapters.find(item => isFallbackAdapter(item) && item.supports(extension));
+        const adapter = primaryAdapter ?? fallbackAdapter;
+        if (!adapter) {
+            throw new AppError(`当前没有可用解析器支持扩展名：${extension || "(无扩展名)"}`, APP_ERROR_CODES.PARSER_ROUTE_UNSUPPORTED, {
+                details: {
+                    filePath,
+                    extension,
+                },
+            });
+        }
+        return {
+            adapter,
+            extension,
+        };
+    }
+}
+//# sourceMappingURL=parser-router.js.map

package/dist/server/modules/affairs-indexer/core/src/parser/parser-router.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"parser-router.js","sourceRoot":"","sources":["../../../../../../../src/modules/affairs-indexer/core/src/parser/parser-router.ts"],"names":[],"mappings":"AAAA,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,QAAQ,EAAE,eAAe,EAAE,MAAM,iCAAiC,CAAC;AAE5E,OAAO,EAAE,0BAA0B,EAAE,MAAM,oCAAoC,CAAC;AAChF,OAAO,EAAE,gBAAgB,EAAE,MAAM,yBAAyB,CAAC;AAC3D,OAAO,EAAE,iBAAiB,EAAE,MAAM,0BAA0B,CAAC;AAC7D,OAAO,EAAE,gBAAgB,EAAE,MAAM,yBAAyB,CAAC;AAC3D,OAAO,EAAE,sBAAsB,EAAE,MAAM,gCAAgC,CAAC;AACxE,OAAO,EAAE,iBAAiB,EAAE,MAAM,0BAA0B,CAAC;AAC7D,OAAO,EAAE,iBAAiB,EAAE,MAAM,0BAA0B,CAAC;AAE7D,SAAS,iBAAiB,CAAC,OAAsB;IAC/C,OAAO,OAAO,CAAC,SAAS,KAAK,UAAU,CAAC;AAC1C,CAAC;AAED,MAAM,UAAU,2BAA2B;IACzC,OAAO;QACL,IAAI,sBAAsB,EAAE;QAC5B,IAAI,gBAAgB,EAAE;QACtB,IAAI,iBAAiB,EAAE;QACvB,IAAI,iBAAiB,EAAE;QACvB,IAAI,gBAAgB,EAAE;QACtB,IAAI,iBAAiB,EAAE;QACvB,IAAI,0BAA0B,EAAE;KACjC,CAAC;AACJ,CAAC;AAED;;;GAGG;AACH,MAAM,OAAO,YAAY;IAIJ;IAHF,kBAAkB,CAAc;IAEjD,YACmB,QAAyB,EAC1C,UAEI,EAAE;QAHW,aAAQ,GAAR,QAAQ,CAAiB;QAK1C,IAAI,CAAC,kBAAkB,GAAG,IAAI,GAAG,CAAC,CAAC,OAAO,CAAC,kBAAkB,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC;IACxG,CAAC;IAED,YAAY;QACV,OAAO,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,CAAC;IAC5B,CAAC;IAED,KAAK,CAAC,cAAc,CAAC,QAAgB;QACnC,MAAM,SAAS,GAAG,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;QACvD,MAAM,cAAc,GAAG,IAAI,CAAC,kBAAkB,CAAC,GAAG,CAAC,SAAS,CAAC;YAC3D,CAAC,CAAC,SAAS;YACX,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,iBAAiB,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC;QACrF,MAAM,eAAe,GAAG,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,iBAAiB,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC;QACxG,MAAM,OAAO,GAAG,cAAc,IAAI,eAAe,CAAC;QAElD,IAAI,CAAC,OAAO,EAAE,CAAC;YACb,MAAM,IAAI,QAAQ,CAChB,kBAAkB,SAAS,IAAI,QAAQ,EAAE,EACzC,eAAe,CAAC,wBAAwB,EACxC;gBACE,OAAO,EAAE;oBACP,QAAQ;oBACR,SAAS;iBACV;aACF,CACF,CAAC;QACJ,CAAC;QAED,OAAO;YACL,OAAO;YACP,SAAS;SACV,CAAC;IACJ,CAAC;CACF"}

package/dist/server/modules/affairs-indexer/core/src/parser/parser-skip-repository.d.ts

@@ -0,0 +1,48 @@
+export interface ParserSkipRecordInput {
+    adapter: string;
+    reasonCode: string;
+    extension: string;
+    path: string;
+    message: string;
+    observedAt: string;
+}
+export interface ParserSkipCatalogRecord {
+    skipKey: string;
+    adapter: string;
+    reasonCode: string;
+    extension: string;
+    samplePaths: string[];
+    sampleCount: number;
+    totalCount: number;
+    lastMessage: string | null;
+    firstSeenAt: string;
+    lastSeenAt: string;
+    lastRunAt: string;
+}
+/**
+ * 复杂文档 skip 聚合目录。
+ * 只保存聚合信息，不为每个复杂文档重复生成重失败记录。
+ */
+export declare class ParserSkipRepository {
+    private readonly dbPath;
+    private activeDb;
+    private selectStatement;
+    private upsertStatement;
+    constructor(dbPath: string);
+    beginSession(): void;
+    endSession(): void;
+    private withDatabase;
+    record(input: ParserSkipRecordInput): ParserSkipCatalogRecord;
+    listRecent(limit?: number): ParserSkipCatalogRecord[];
+    summarize(): {
+        totalKinds: number;
+        totalSkipped: number;
+        byReason: Array<{
+            adapter: string;
+            reasonCode: string;
+            extension: string;
+            totalCount: number;
+            lastSeenAt: string;
+        }>;
+    };
+}