@jingyi0605/codingns 0.8.1 → 0.9.0

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

@@ -0,0 +1,193 @@
+import crypto from "node:crypto";
+import { openDatabase } from "../sqlite/open-database.js";
+function makeSkipKey(adapter, reasonCode, extension) {
+    const digest = crypto.createHash("sha1").update(`${adapter}:${reasonCode}:${extension}`).digest("hex").slice(0, 16);
+    return `skip_${digest}`;
+}
+function normalizeSamplePaths(raw, candidatePath, limit = 20) {
+    let values = [];
+    try {
+        const parsed = JSON.parse(raw);
+        if (Array.isArray(parsed)) {
+            values = parsed.filter((item) => typeof item === "string" && item.trim().length > 0);
+        }
+    }
+    catch {
+        values = [];
+    }
+    if (candidatePath && !values.includes(candidatePath)) {
+        values.push(candidatePath);
+    }
+    return values.slice(0, limit);
+}
+/**
+ * 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 class ParserSkipRepository {
+    dbPath;
+    activeDb = null;
+    selectStatement = null;
+    upsertStatement = null;
+    constructor(dbPath) {
+        this.dbPath = dbPath;
+    }
+    beginSession() {
+        if (this.activeDb) {
+            return;
+        }
+        this.activeDb = openDatabase(this.dbPath);
+        this.selectStatement = this.activeDb.prepare(`
+      SELECT sample_paths_json, sample_count, total_count, first_seen_at
+      FROM parser_skip_catalog
+      WHERE skip_key = ?
+    `);
+        this.upsertStatement = this.activeDb.prepare(`
+      INSERT INTO parser_skip_catalog(
+        skip_key, adapter, reason_code, extension, sample_paths_json, sample_count, total_count, last_message, first_seen_at, last_seen_at, last_run_at
+      )
+      VALUES(?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+      ON CONFLICT(skip_key) DO UPDATE SET
+        sample_paths_json = excluded.sample_paths_json,
+        sample_count = excluded.sample_count,
+        total_count = excluded.total_count,
+        last_message = excluded.last_message,
+        last_seen_at = excluded.last_seen_at,
+        last_run_at = excluded.last_run_at
+    `);
+    }
+    endSession() {
+        if (!this.activeDb) {
+            return;
+        }
+        this.activeDb.close();
+        this.activeDb = null;
+        this.selectStatement = null;
+        this.upsertStatement = null;
+    }
+    withDatabase(handler) {
+        if (this.activeDb && this.selectStatement && this.upsertStatement) {
+            return handler(this.activeDb, this.selectStatement, this.upsertStatement);
+        }
+        const db = openDatabase(this.dbPath);
+        const selectStatement = db.prepare(`
+      SELECT sample_paths_json, sample_count, total_count, first_seen_at
+      FROM parser_skip_catalog
+      WHERE skip_key = ?
+    `);
+        const upsertStatement = db.prepare(`
+      INSERT INTO parser_skip_catalog(
+        skip_key, adapter, reason_code, extension, sample_paths_json, sample_count, total_count, last_message, first_seen_at, last_seen_at, last_run_at
+      )
+      VALUES(?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+      ON CONFLICT(skip_key) DO UPDATE SET
+        sample_paths_json = excluded.sample_paths_json,
+        sample_count = excluded.sample_count,
+        total_count = excluded.total_count,
+        last_message = excluded.last_message,
+        last_seen_at = excluded.last_seen_at,
+        last_run_at = excluded.last_run_at
+    `);
+        try {
+            return handler(db, selectStatement, upsertStatement);
+        }
+        finally {
+            db.close();
+        }
+    }
+    record(input) {
+        const skipKey = makeSkipKey(input.adapter, input.reasonCode, input.extension);
+        return this.withDatabase((db, selectStatement, upsertStatement) => {
+            const ownsTransaction = !this.activeDb;
+            try {
+                if (ownsTransaction) {
+                    db.exec("BEGIN IMMEDIATE");
+                }
+                const existing = selectStatement.get(skipKey);
+                const samplePaths = normalizeSamplePaths(existing?.sample_paths_json ?? "[]", input.path);
+                const sampleCount = samplePaths.length;
+                const totalCount = (existing?.total_count ?? 0) + 1;
+                const firstSeenAt = existing?.first_seen_at ?? input.observedAt;
+                upsertStatement.run(skipKey, input.adapter, input.reasonCode, input.extension, JSON.stringify(samplePaths), sampleCount, totalCount, input.message, firstSeenAt, input.observedAt, input.observedAt);
+                if (ownsTransaction) {
+                    db.exec("COMMIT");
+                }
+                return {
+                    skipKey,
+                    adapter: input.adapter,
+                    reasonCode: input.reasonCode,
+                    extension: input.extension,
+                    samplePaths,
+                    sampleCount,
+                    totalCount,
+                    lastMessage: input.message,
+                    firstSeenAt,
+                    lastSeenAt: input.observedAt,
+                    lastRunAt: input.observedAt,
+                };
+            }
+            catch (error) {
+                if (ownsTransaction) {
+                    db.exec("ROLLBACK");
+                }
+                throw error;
+            }
+        });
+    }
+    listRecent(limit = 100) {
+        const db = openDatabase(this.dbPath);
+        try {
+            const rows = db.prepare(`
+        SELECT skip_key, adapter, reason_code, extension, sample_paths_json, sample_count, total_count, last_message, first_seen_at, last_seen_at, last_run_at
+        FROM parser_skip_catalog
+        ORDER BY last_seen_at DESC, extension ASC
+        LIMIT ?
+      `).all(limit);
+            return rows.map(row => ({
+                skipKey: String(row.skip_key),
+                adapter: String(row.adapter),
+                reasonCode: String(row.reason_code),
+                extension: String(row.extension),
+                samplePaths: normalizeSamplePaths(String(row.sample_paths_json ?? "[]")),
+                sampleCount: Number(row.sample_count ?? 0),
+                totalCount: Number(row.total_count ?? 0),
+                lastMessage: row.last_message ? String(row.last_message) : null,
+                firstSeenAt: String(row.first_seen_at),
+                lastSeenAt: String(row.last_seen_at),
+                lastRunAt: String(row.last_run_at),
+            }));
+        }
+        finally {
+            db.close();
+        }
+    }
+    summarize() {
+        const db = openDatabase(this.dbPath);
+        try {
+            const summary = db.prepare(`
+        SELECT COUNT(*) AS total_kinds, COALESCE(SUM(total_count), 0) AS total_skipped
+        FROM parser_skip_catalog
+      `).get();
+            const rows = db.prepare(`
+        SELECT adapter, reason_code, extension, total_count, last_seen_at
+        FROM parser_skip_catalog
+        ORDER BY total_count DESC, extension ASC
+      `).all();
+            return {
+                totalKinds: Number(summary?.total_kinds ?? 0),
+                totalSkipped: Number(summary?.total_skipped ?? 0),
+                byReason: rows.map(row => ({
+                    adapter: String(row.adapter),
+                    reasonCode: String(row.reason_code),
+                    extension: String(row.extension),
+                    totalCount: Number(row.total_count ?? 0),
+                    lastSeenAt: String(row.last_seen_at),
+                })),
+            };
+        }
+        finally {
+            db.close();
+        }
+    }
+}
+//# sourceMappingURL=parser-skip-repository.js.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"parser-skip-repository.js","sourceRoot":"","sources":["../../../../../../../src/modules/affairs-indexer/core/src/parser/parser-skip-repository.ts"],"names":[],"mappings":"AACA,OAAO,MAAM,MAAM,aAAa,CAAC;AACjC,OAAO,EAAE,YAAY,EAAE,MAAM,4BAA4B,CAAC;AAyB1D,SAAS,WAAW,CAAC,OAAe,EAAE,UAAkB,EAAE,SAAiB;IACzE,MAAM,MAAM,GAAG,MAAM,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC,GAAG,OAAO,IAAI,UAAU,IAAI,SAAS,EAAE,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;IACpH,OAAO,QAAQ,MAAM,EAAE,CAAC;AAC1B,CAAC;AAED,SAAS,oBAAoB,CAAC,GAAW,EAAE,aAAsB,EAAE,KAAK,GAAG,EAAE;IAC3E,IAAI,MAAM,GAAa,EAAE,CAAC;IAC1B,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAY,CAAC;QAC1C,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;YAC1B,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,IAAI,EAAkB,EAAE,CAAC,OAAO,IAAI,KAAK,QAAQ,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QACvG,CAAC;IACH,CAAC;IAAC,MAAM,CAAC;QACP,MAAM,GAAG,EAAE,CAAC;IACd,CAAC;IAED,IAAI,aAAa,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,aAAa,CAAC,EAAE,CAAC;QACrD,MAAM,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;IAC7B,CAAC;IAED,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;AAChC,CAAC;AAED;;;GAGG;AACH,MAAM,OAAO,oBAAoB;IAKF;IAJrB,QAAQ,GAAwB,IAAI,CAAC;IACrC,eAAe,GAA+C,IAAI,CAAC;IACnE,eAAe,GAA+C,IAAI,CAAC;IAE3E,YAA6B,MAAc;QAAd,WAAM,GAAN,MAAM,CAAQ;IAAG,CAAC;IAE/C,YAAY;QACV,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAClB,OAAO;QACT,CAAC;QACD,IAAI,CAAC,QAAQ,GAAG,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAC1C,IAAI,CAAC,eAAe,GAAG,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC;;;;KAI5C,CAAC,CAAC;QACH,IAAI,CAAC,eAAe,GAAG,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC;;;;;;;;;;;;KAY5C,CAAC,CAAC;IACL,CAAC;IAED,UAAU;QACR,IAAI,CAAC,IAAI,CAAC,QAAQ,EAAE,CAAC;YACnB,OAAO;QACT,CAAC;QACD,IAAI,CAAC,QAAQ,CAAC,KAAK,EAAE,CAAC;QACtB,IAAI,CAAC,QAAQ,GAAG,IAAI,CAAC;QACrB,IAAI,CAAC,eAAe,GAAG,IAAI,CAAC;QAC5B,IAAI,CAAC,eAAe,GAAG,IAAI,CAAC;IAC9B,CAAC;IAEO,YAAY,CAAI,OAIlB;QACJ,IAAI,IAAI,CAAC,QAAQ,IAAI,IAAI,CAAC,eAAe,IAAI,IAAI,CAAC,eAAe,EAAE,CAAC;YAClE,OAAO,OAAO,CAAC,IAAI,CAAC,QAAQ,EAAE,IAAI,CAAC,eAAe,EAAE,IAAI,CAAC,eAAe,CAAC,CAAC;QAC5E,CAAC;QAED,MAAM,EAAE,GAAG,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACrC,MAAM,eAAe,GAAG,EAAE,CAAC,OAAO,CAAC;;;;KAIlC,CAAC,CAAC;QACH,MAAM,eAAe,GAAG,EAAE,CAAC,OAAO,CAAC;;;;;;;;;;;;KAYlC,CAAC,CAAC;QAEH,IAAI,CAAC;YACH,OAAO,OAAO,CAAC,EAAE,EAAE,eAAe,EAAE,eAAe,CAAC,CAAC;QACvD,CAAC;gBAAS,CAAC;YACT,EAAE,CAAC,KAAK,EAAE,CAAC;QACb,CAAC;IACH,CAAC;IAED,MAAM,CAAC,KAA4B;QACjC,MAAM,OAAO,GAAG,WAAW,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,UAAU,EAAE,KAAK,CAAC,SAAS,CAAC,CAAC;QAC9E,OAAO,IAAI,CAAC,YAAY,CAAC,CAAC,EAAE,EAAE,eAAe,EAAE,eAAe,EAAE,EAAE;YAChE,MAAM,eAAe,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC;YACvC,IAAI,CAAC;gBACH,IAAI,eAAe,EAAE,CAAC;oBACpB,EAAE,CAAC,IAAI,CAAC,iBAAiB,CAAC,CAAC;gBAC7B,CAAC;gBAED,MAAM,QAAQ,GAAG,eAAe,CAAC,GAAG,CAAC,OAAO,CAK/B,CAAC;gBAEd,MAAM,WAAW,GAAG,oBAAoB,CAAC,QAAQ,EAAE,iBAAiB,IAAI,IAAI,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC;gBAC1F,MAAM,WAAW,GAAG,WAAW,CAAC,MAAM,CAAC;gBACvC,MAAM,UAAU,GAAG,CAAC,QAAQ,EAAE,WAAW,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;gBACpD,MAAM,WAAW,GAAG,QAAQ,EAAE,aAAa,IAAI,KAAK,CAAC,UAAU,CAAC;gBAEhE,eAAe,CAAC,GAAG,CACjB,OAAO,EACP,KAAK,CAAC,OAAO,EACb,KAAK,CAAC,UAAU,EAChB,KAAK,CAAC,SAAS,EACf,IAAI,CAAC,SAAS,CAAC,WAAW,CAAC,EAC3B,WAAW,EACX,UAAU,EACV,KAAK,CAAC,OAAO,EACb,WAAW,EACX,KAAK,CAAC,UAAU,EAChB,KAAK,CAAC,UAAU,CACjB,CAAC;gBAEF,IAAI,eAAe,EAAE,CAAC;oBACpB,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;gBACpB,CAAC;gBACD,OAAO;oBACL,OAAO;oBACP,OAAO,EAAE,KAAK,CAAC,OAAO;oBACtB,UAAU,EAAE,KAAK,CAAC,UAAU;oBAC5B,SAAS,EAAE,KAAK,CAAC,SAAS;oBAC1B,WAAW;oBACX,WAAW;oBACX,UAAU;oBACV,WAAW,EAAE,KAAK,CAAC,OAAO;oBAC1B,WAAW;oBACX,UAAU,EAAE,KAAK,CAAC,UAAU;oBAC5B,SAAS,EAAE,KAAK,CAAC,UAAU;iBAC5B,CAAC;YACJ,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,IAAI,eAAe,EAAE,CAAC;oBACpB,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;gBACtB,CAAC;gBACD,MAAM,KAAK,CAAC;YACd,CAAC;QACH,CAAC,CAAC,CAAC;IACL,CAAC;IAED,UAAU,CAAC,KAAK,GAAG,GAAG;QACpB,MAAM,EAAE,GAAG,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACrC,IAAI,CAAC;YACH,MAAM,IAAI,GAAG,EAAE,CAAC,OAAO,CAAC;;;;;OAKvB,CAAC,CAAC,GAAG,CAAC,KAAK,CAAmC,CAAC;YAEhD,OAAO,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;gBACtB,OAAO,EAAE,MAAM,CAAC,GAAG,CAAC,QAAQ,CAAC;gBAC7B,OAAO,EAAE,MAAM,CAAC,GAAG,CAAC,OAAO,CAAC;gBAC5B,UAAU,EAAE,MAAM,CAAC,GAAG,CAAC,WAAW,CAAC;gBACnC,SAAS,EAAE,MAAM,CAAC,GAAG,CAAC,SAAS,CAAC;gBAChC,WAAW,EAAE,oBAAoB,CAAC,MAAM,CAAC,GAAG,CAAC,iBAAiB,IAAI,IAAI,CAAC,CAAC;gBACxE,WAAW,EAAE,MAAM,CAAC,GAAG,CAAC,YAAY,IAAI,CAAC,CAAC;gBAC1C,UAAU,EAAE,MAAM,CAAC,GAAG,CAAC,WAAW,IAAI,CAAC,CAAC;gBACxC,WAAW,EAAE,GAAG,CAAC,YAAY,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC,IAAI;gBAC/D,WAAW,EAAE,MAAM,CAAC,GAAG,CAAC,aAAa,CAAC;gBACtC,UAAU,EAAE,MAAM,CAAC,GAAG,CAAC,YAAY,CAAC;gBACpC,SAAS,EAAE,MAAM,CAAC,GAAG,CAAC,WAAW,CAAC;aACnC,CAAC,CAAC,CAAC;QACN,CAAC;gBAAS,CAAC;YACT,EAAE,CAAC,KAAK,EAAE,CAAC;QACb,CAAC;IACH,CAAC;IAED,SAAS;QAWP,MAAM,EAAE,GAAG,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACrC,IAAI,CAAC;YACH,MAAM,OAAO,GAAG,EAAE,CAAC,OAAO,CAAC;;;OAG1B,CAAC,CAAC,GAAG,EAAkE,CAAC;YAEzE,MAAM,IAAI,GAAG,EAAE,CAAC,OAAO,CAAC;;;;OAIvB,CAAC,CAAC,GAAG,EAAoC,CAAC;YAE3C,OAAO;gBACL,UAAU,EAAE,MAAM,CAAC,OAAO,EAAE,WAAW,IAAI,CAAC,CAAC;gBAC7C,YAAY,EAAE,MAAM,CAAC,OAAO,EAAE,aAAa,IAAI,CAAC,CAAC;gBACjD,QAAQ,EAAE,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;oBACzB,OAAO,EAAE,MAAM,CAAC,GAAG,CAAC,OAAO,CAAC;oBAC5B,UAAU,EAAE,MAAM,CAAC,GAAG,CAAC,WAAW,CAAC;oBACnC,SAAS,EAAE,MAAM,CAAC,GAAG,CAAC,SAAS,CAAC;oBAChC,UAAU,EAAE,MAAM,CAAC,GAAG,CAAC,WAAW,IAAI,CAAC,CAAC;oBACxC,UAAU,EAAE,MAAM,CAAC,GAAG,CAAC,YAAY,CAAC;iBACrC,CAAC,CAAC;aACJ,CAAC;QACJ,CAAC;gBAAS,CAAC;YACT,EAAE,CAAC,KAAK,EAAE,CAAC;QACb,CAAC;IACH,CAAC;CACF"}

package/dist/server/modules/affairs-indexer/core/src/parser/pdf-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 PdfParserAdapter extends BaseComplexParserAdapter {
+    readonly name = "pdf_parser";
+    supports(ext: string): boolean;
+    protected parseComplex(input: ParseInput): Promise<ParsedDocumentPayload>;
+}

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

@@ -0,0 +1,371 @@
+import fs from "node:fs";
+import path from "node:path";
+import { inflateRawSync, inflateSync } from "node:zlib";
+import { AppError, APP_ERROR_CODES } from "../../../contracts/src/index.js";
+import { BaseComplexParserAdapter } from "./base-complex-parser-adapter.js";
+import { shortSummary } from "./openxml-utils.js";
+const SUPPORTED_PDF_EXTENSIONS = new Set([".pdf"]);
+function decodePdfNameToken(token) {
+    return token.replace(/#([0-9A-Fa-f]{2})/g, (_, hex) => String.fromCharCode(Number.parseInt(hex, 16)));
+}
+function normalizePdfText(text) {
+    return text
+        .replace(/\r\n/g, "\n")
+        .replace(/\r/g, "\n")
+        .replace(/[ \t]+\n/g, "\n")
+        .replace(/\n{3,}/g, "\n\n")
+        .replace(/[ \t]{2,}/g, " ")
+        .trim();
+}
+function extractPdfObjects(pdfText) {
+    const objects = new Map();
+    const objectPattern = /(\d+)\s+(\d+)\s+obj\b([\s\S]*?)endobj/g;
+    for (const match of pdfText.matchAll(objectPattern)) {
+        const objectId = `${match[1]} ${match[2]}`;
+        objects.set(objectId, {
+            objectId,
+            body: match[3].trim(),
+        });
+    }
+    return objects;
+}
+function parsePageRecords(objects) {
+    const pages = [];
+    for (const [objectId, record] of objects.entries()) {
+        if (!/\/Type\s*\/Page\b/.test(record.body)) {
+            continue;
+        }
+        if (/\/Type\s*\/Pages\b/.test(record.body)) {
+            continue;
+        }
+        const contentsRefs = [];
+        const arrayMatch = record.body.match(/\/Contents\s*\[([\s\S]*?)\]/);
+        if (arrayMatch) {
+            for (const refMatch of arrayMatch[1].matchAll(/(\d+)\s+(\d+)\s+R/g)) {
+                contentsRefs.push(`${refMatch[1]} ${refMatch[2]}`);
+            }
+        }
+        else {
+            const singleMatch = record.body.match(/\/Contents\s+(\d+)\s+(\d+)\s+R/);
+            if (singleMatch) {
+                contentsRefs.push(`${singleMatch[1]} ${singleMatch[2]}`);
+            }
+        }
+        pages.push({
+            objectNumber: Number.parseInt(objectId.split(" ")[0] ?? "0", 10),
+            record: {
+                contentRefs: contentsRefs,
+            },
+        });
+    }
+    return pages
+        .sort((left, right) => left.objectNumber - right.objectNumber)
+        .map(item => item.record);
+}
+function extractStreamBuffer(objectBody) {
+    const streamStart = objectBody.indexOf("stream");
+    if (streamStart < 0) {
+        return null;
+    }
+    let dataStart = streamStart + "stream".length;
+    if (objectBody[dataStart] === "\r" && objectBody[dataStart + 1] === "\n") {
+        dataStart += 2;
+    }
+    else if (objectBody[dataStart] === "\n") {
+        dataStart += 1;
+    }
+    else if (objectBody[dataStart] === "\r") {
+        dataStart += 1;
+    }
+    const streamEnd = objectBody.lastIndexOf("endstream");
+    if (streamEnd < 0 || streamEnd <= dataStart) {
+        return null;
+    }
+    let rawData = objectBody.slice(dataStart, streamEnd);
+    rawData = rawData.replace(/[\r\n]+$/, "");
+    return Buffer.from(rawData, "latin1");
+}
+function decodePdfStream(record) {
+    const rawStream = extractStreamBuffer(record.body);
+    if (!rawStream) {
+        return null;
+    }
+    const filters = [];
+    const arrayFilterMatch = record.body.match(/\/Filter\s*\[([^\]]+)\]/);
+    if (arrayFilterMatch) {
+        for (const match of arrayFilterMatch[1].matchAll(/\/([A-Za-z0-9#]+)/g)) {
+            filters.push(decodePdfNameToken(match[1]));
+        }
+    }
+    else {
+        const singleFilterMatch = record.body.match(/\/Filter\s*\/([A-Za-z0-9#]+)/);
+        if (singleFilterMatch) {
+            filters.push(decodePdfNameToken(singleFilterMatch[1]));
+        }
+    }
+    let contentBuffer = rawStream;
+    for (const filter of filters) {
+        if (filter === "FlateDecode") {
+            try {
+                contentBuffer = inflateSync(contentBuffer);
+            }
+            catch {
+                contentBuffer = inflateRawSync(contentBuffer);
+            }
+            continue;
+        }
+        throw new AppError(`PDF 当前未支持过滤器：${filter}`, APP_ERROR_CODES.PARSER_COMPLEX_FORMAT_UNSUPPORTED);
+    }
+    return contentBuffer.toString("latin1");
+}
+function decodePdfEscapeSequence(char) {
+    switch (char) {
+        case "n":
+            return "\n";
+        case "r":
+            return "\r";
+        case "t":
+            return "\t";
+        case "b":
+            return "\b";
+        case "f":
+            return "\f";
+        case "(":
+        case ")":
+        case "\\":
+            return char;
+        default:
+            return char;
+    }
+}
+function decodePdfLiteralString(input) {
+    let result = "";
+    for (let index = 0; index < input.length; index += 1) {
+        const char = input[index];
+        if (char !== "\\") {
+            result += char;
+            continue;
+        }
+        const next = input[index + 1];
+        if (!next) {
+            break;
+        }
+        if (/[0-7]/.test(next)) {
+            let octal = next;
+            let consumed = 1;
+            while (consumed < 3 && /[0-7]/.test(input[index + 1 + consumed] ?? "")) {
+                octal += input[index + 1 + consumed];
+                consumed += 1;
+            }
+            result += String.fromCharCode(Number.parseInt(octal, 8));
+            index += consumed;
+            continue;
+        }
+        result += decodePdfEscapeSequence(next);
+        index += 1;
+    }
+    return result;
+}
+function decodePdfHexString(input) {
+    const normalized = input.replace(/\s+/g, "");
+    const padded = normalized.length % 2 === 0 ? normalized : `${normalized}0`;
+    let result = "";
+    for (let index = 0; index < padded.length; index += 2) {
+        const value = Number.parseInt(padded.slice(index, index + 2), 16);
+        if (!Number.isNaN(value)) {
+            result += String.fromCharCode(value);
+        }
+    }
+    return result;
+}
+function extractBalancedParentheses(text, startIndex) {
+    let depth = 1;
+    let current = "";
+    for (let index = startIndex + 1; index < text.length; index += 1) {
+        const char = text[index];
+        if (char === "\\") {
+            current += char;
+            if (index + 1 < text.length) {
+                current += text[index + 1];
+                index += 1;
+            }
+            continue;
+        }
+        if (char === "(") {
+            depth += 1;
+            current += char;
+            continue;
+        }
+        if (char === ")") {
+            depth -= 1;
+            if (depth === 0) {
+                return {
+                    value: current,
+                    endIndex: index,
+                };
+            }
+            current += char;
+            continue;
+        }
+        current += char;
+    }
+    return null;
+}
+function readArrayTokens(raw) {
+    const tokens = [];
+    let index = 0;
+    while (index < raw.length) {
+        const char = raw[index];
+        if (/\s/.test(char)) {
+            index += 1;
+            continue;
+        }
+        if (char === "(") {
+            const literal = extractBalancedParentheses(raw, index);
+            if (!literal) {
+                break;
+            }
+            tokens.push(decodePdfLiteralString(literal.value));
+            index = literal.endIndex + 1;
+            continue;
+        }
+        if (char === "<") {
+            const endIndex = raw.indexOf(">", index + 1);
+            if (endIndex < 0) {
+                break;
+            }
+            tokens.push(decodePdfHexString(raw.slice(index + 1, endIndex)));
+            index = endIndex + 1;
+            continue;
+        }
+        const nextWhitespace = raw.slice(index).search(/\s/);
+        if (nextWhitespace < 0) {
+            break;
+        }
+        index += nextWhitespace + 1;
+    }
+    return tokens;
+}
+function extractTextFromPdfOperators(content) {
+    const parts = [];
+    for (let index = 0; index < content.length; index += 1) {
+        const char = content[index];
+        if (char === "(") {
+            const literal = extractBalancedParentheses(content, index);
+            if (!literal) {
+                continue;
+            }
+            const operatorSegment = content.slice(literal.endIndex + 1, literal.endIndex + 6);
+            if (/\s*Tj\b/.test(operatorSegment) || /\s*'/.test(operatorSegment) || /\s*"/.test(operatorSegment)) {
+                parts.push(decodePdfLiteralString(literal.value));
+            }
+            index = literal.endIndex;
+            continue;
+        }
+        if (char === "<" && content[index + 1] !== "<") {
+            const endIndex = content.indexOf(">", index + 1);
+            if (endIndex < 0) {
+                continue;
+            }
+            const operatorSegment = content.slice(endIndex + 1, endIndex + 6);
+            if (/\s*Tj\b/.test(operatorSegment)) {
+                parts.push(decodePdfHexString(content.slice(index + 1, endIndex)));
+            }
+            index = endIndex;
+            continue;
+        }
+    }
+    const arrayPattern = /\[(.*?)\]\s*TJ\b/gs;
+    for (const match of content.matchAll(arrayPattern)) {
+        const tokens = readArrayTokens(match[1]);
+        if (tokens.length > 0) {
+            parts.push(tokens.join(""));
+        }
+    }
+    return normalizePdfText(parts.join("\n"));
+}
+function extractPageText(page, objects) {
+    const fragments = [];
+    for (const contentRef of page.contentRefs) {
+        const objectRecord = objects.get(contentRef);
+        if (!objectRecord) {
+            continue;
+        }
+        const streamText = decodePdfStream(objectRecord);
+        if (!streamText) {
+            continue;
+        }
+        const text = extractTextFromPdfOperators(streamText);
+        if (text) {
+            fragments.push(text);
+        }
+    }
+    return normalizePdfText(fragments.join("\n\n"));
+}
+export class PdfParserAdapter extends BaseComplexParserAdapter {
+    name = "pdf_parser";
+    supports(ext) {
+        return SUPPORTED_PDF_EXTENSIONS.has(ext.toLowerCase());
+    }
+    async parseComplex(input) {
+        let fileBuffer;
+        try {
+            fileBuffer = fs.readFileSync(input.filePath);
+        }
+        catch (error) {
+            throw new AppError(`PDF 文件读取失败：${path.basename(input.filePath)}`, APP_ERROR_CODES.PARSER_COMPLEX_CONTENT_UNREADABLE, { cause: error });
+        }
+        const pdfText = fileBuffer.toString("latin1");
+        if (!pdfText.startsWith("%PDF-")) {
+            throw new AppError("PDF 文件头非法或内容不可读", APP_ERROR_CODES.PARSER_COMPLEX_CONTENT_UNREADABLE);
+        }
+        const objects = extractPdfObjects(pdfText);
+        if (objects.size === 0) {
+            throw new AppError("PDF 中未发现可读取对象", APP_ERROR_CODES.PARSER_COMPLEX_CONTENT_UNREADABLE);
+        }
+        const pages = parsePageRecords(objects);
+        if (pages.length === 0) {
+            throw new AppError("PDF 中未发现页面对象", APP_ERROR_CODES.PARSER_COMPLEX_CONTENT_UNREADABLE);
+        }
+        const blocks = [];
+        for (let pageIndex = 0; pageIndex < pages.length; pageIndex += 1) {
+            const pageNumber = pageIndex + 1;
+            const pageText = extractPageText(pages[pageIndex], objects);
+            if (!pageText) {
+                continue;
+            }
+            blocks.push({
+                kind: "page",
+                page: pageNumber,
+                text: pageText,
+                metadata: {
+                    pageNumber,
+                },
+            });
+        }
+        if (blocks.length === 0) {
+            throw new AppError("PDF 中没有可提取的文本内容", APP_ERROR_CODES.PARSER_COMPLEX_CONTENT_UNREADABLE);
+        }
+        const text = blocks.map(block => block.text ?? "").filter(Boolean).join("\n\n").trim();
+        const title = path.basename(input.filePath, input.extension);
+        return {
+            title,
+            text,
+            summary: shortSummary(text),
+            parser: "pdf",
+            metadata: {
+                adapter: this.name,
+                pageCount: pages.length,
+                extractedPageCount: blocks.length,
+            },
+            structured: {
+                blocks,
+                stats: {
+                    pageCount: pages.length,
+                    extractedPageCount: blocks.length,
+                },
+            },
+        };
+    }
+}
+//# sourceMappingURL=pdf-parser-adapter.js.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"pdf-parser-adapter.js","sourceRoot":"","sources":["../../../../../../../src/modules/affairs-indexer/core/src/parser/pdf-parser-adapter.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,SAAS,CAAC;AACzB,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,cAAc,EAAE,WAAW,EAAE,MAAM,WAAW,CAAC;AACxD,OAAO,EAAE,QAAQ,EAAE,eAAe,EAAE,MAAM,iCAAiC,CAAC;AAC5E,OAAO,EAAE,wBAAwB,EAAE,MAAM,kCAAkC,CAAC;AAC5E,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAGlD,MAAM,wBAAwB,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC;AAWnD,SAAS,kBAAkB,CAAC,KAAa;IACvC,OAAO,KAAK,CAAC,OAAO,CAAC,oBAAoB,EAAE,CAAC,CAAC,EAAE,GAAW,EAAE,EAAE,CAAC,MAAM,CAAC,YAAY,CAAC,MAAM,CAAC,QAAQ,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC;AAChH,CAAC;AAED,SAAS,gBAAgB,CAAC,IAAY;IACpC,OAAO,IAAI;SACR,OAAO,CAAC,OAAO,EAAE,IAAI,CAAC;SACtB,OAAO,CAAC,KAAK,EAAE,IAAI,CAAC;SACpB,OAAO,CAAC,WAAW,EAAE,IAAI,CAAC;SAC1B,OAAO,CAAC,SAAS,EAAE,MAAM,CAAC;SAC1B,OAAO,CAAC,YAAY,EAAE,GAAG,CAAC;SAC1B,IAAI,EAAE,CAAC;AACZ,CAAC;AAED,SAAS,iBAAiB,CAAC,OAAe;IACxC,MAAM,OAAO,GAAG,IAAI,GAAG,EAA2B,CAAC;IACnD,MAAM,aAAa,GAAG,wCAAwC,CAAC;IAC/D,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,QAAQ,CAAC,aAAa,CAAC,EAAE,CAAC;QACpD,MAAM,QAAQ,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;QAC3C,OAAO,CAAC,GAAG,CAAC,QAAQ,EAAE;YACpB,QAAQ;YACR,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE;SACtB,CAAC,CAAC;IACL,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,gBAAgB,CAAC,OAAqC;IAC7D,MAAM,KAAK,GAA2D,EAAE,CAAC;IACzE,KAAK,MAAM,CAAC,QAAQ,EAAE,MAAM,CAAC,IAAI,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC;QACnD,IAAI,CAAC,mBAAmB,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC;YAC3C,SAAS;QACX,CAAC;QACD,IAAI,oBAAoB,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC;YAC3C,SAAS;QACX,CAAC;QAED,MAAM,YAAY,GAAa,EAAE,CAAC;QAClC,MAAM,UAAU,GAAG,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,6BAA6B,CAAC,CAAC;QACpE,IAAI,UAAU,EAAE,CAAC;YACf,KAAK,MAAM,QAAQ,IAAI,UAAU,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,oBAAoB,CAAC,EAAE,CAAC;gBACpE,YAAY,CAAC,IAAI,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC,IAAI,QAAQ,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;YACrD,CAAC;QACH,CAAC;aAAM,CAAC;YACN,MAAM,WAAW,GAAG,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,gCAAgC,CAAC,CAAC;YACxE,IAAI,WAAW,EAAE,CAAC;gBAChB,YAAY,CAAC,IAAI,CAAC,GAAG,WAAW,CAAC,CAAC,CAAC,IAAI,WAAW,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;YAC3D,CAAC;QACH,CAAC;QAED,KAAK,CAAC,IAAI,CAAC;YACT,YAAY,EAAE,MAAM,CAAC,QAAQ,CAAC,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,GAAG,EAAE,EAAE,CAAC;YAChE,MAAM,EAAE;gBACN,WAAW,EAAE,YAAY;aAC1B;SACF,CAAC,CAAC;IACL,CAAC;IAED,OAAO,KAAK;SACT,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,YAAY,GAAG,KAAK,CAAC,YAAY,CAAC;SAC7D,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;AAC9B,CAAC;AAED,SAAS,mBAAmB,CAAC,UAAkB;IAC7C,MAAM,WAAW,GAAG,UAAU,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC;IACjD,IAAI,WAAW,GAAG,CAAC,EAAE,CAAC;QACpB,OAAO,IAAI,CAAC;IACd,CAAC;IAED,IAAI,SAAS,GAAG,WAAW,GAAG,QAAQ,CAAC,MAAM,CAAC;IAC9C,IAAI,UAAU,CAAC,SAAS,CAAC,KAAK,IAAI,IAAI,UAAU,CAAC,SAAS,GAAG,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QACzE,SAAS,IAAI,CAAC,CAAC;IACjB,CAAC;SAAM,IAAI,UAAU,CAAC,SAAS,CAAC,KAAK,IAAI,EAAE,CAAC;QAC1C,SAAS,IAAI,CAAC,CAAC;IACjB,CAAC;SAAM,IAAI,UAAU,CAAC,SAAS,CAAC,KAAK,IAAI,EAAE,CAAC;QAC1C,SAAS,IAAI,CAAC,CAAC;IACjB,CAAC;IAED,MAAM,SAAS,GAAG,UAAU,CAAC,WAAW,CAAC,WAAW,CAAC,CAAC;IACtD,IAAI,SAAS,GAAG,CAAC,IAAI,SAAS,IAAI,SAAS,EAAE,CAAC;QAC5C,OAAO,IAAI,CAAC;IACd,CAAC;IAED,IAAI,OAAO,GAAG,UAAU,CAAC,KAAK,CAAC,SAAS,EAAE,SAAS,CAAC,CAAC;IACrD,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC;IAC1C,OAAO,MAAM,CAAC,IAAI,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC;AACxC,CAAC;AAED,SAAS,eAAe,CAAC,MAAuB;IAC9C,MAAM,SAAS,GAAG,mBAAmB,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IACnD,IAAI,CAAC,SAAS,EAAE,CAAC;QACf,OAAO,IAAI,CAAC;IACd,CAAC;IAED,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,MAAM,gBAAgB,GAAG,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,yBAAyB,CAAC,CAAC;IACtE,IAAI,gBAAgB,EAAE,CAAC;QACrB,KAAK,MAAM,KAAK,IAAI,gBAAgB,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,oBAAoB,CAAC,EAAE,CAAC;YACvE,OAAO,CAAC,IAAI,CAAC,kBAAkB,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAC7C,CAAC;IACH,CAAC;SAAM,CAAC;QACN,MAAM,iBAAiB,GAAG,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,8BAA8B,CAAC,CAAC;QAC5E,IAAI,iBAAiB,EAAE,CAAC;YACtB,OAAO,CAAC,IAAI,CAAC,kBAAkB,CAAC,iBAAiB,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QACzD,CAAC;IACH,CAAC;IAED,IAAI,aAAa,GAAG,SAAS,CAAC;IAC9B,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,IAAI,MAAM,KAAK,aAAa,EAAE,CAAC;YAC7B,IAAI,CAAC;gBACH,aAAa,GAAG,WAAW,CAAC,aAAa,CAAC,CAAC;YAC7C,CAAC;YAAC,MAAM,CAAC;gBACP,aAAa,GAAG,cAAc,CAAC,aAAa,CAAC,CAAC;YAChD,CAAC;YACD,SAAS;QACX,CAAC;QAED,MAAM,IAAI,QAAQ,CAChB,gBAAgB,MAAM,EAAE,EACxB,eAAe,CAAC,iCAAiC,CAClD,CAAC;IACJ,CAAC;IAED,OAAO,aAAa,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;AAC1C,CAAC;AAED,SAAS,uBAAuB,CAAC,IAAY;IAC3C,QAAQ,IAAI,EAAE,CAAC;QACb,KAAK,GAAG;YACN,OAAO,IAAI,CAAC;QACd,KAAK,GAAG;YACN,OAAO,IAAI,CAAC;QACd,KAAK,GAAG;YACN,OAAO,IAAI,CAAC;QACd,KAAK,GAAG;YACN,OAAO,IAAI,CAAC;QACd,KAAK,GAAG;YACN,OAAO,IAAI,CAAC;QACd,KAAK,GAAG,CAAC;QACT,KAAK,GAAG,CAAC;QACT,KAAK,IAAI;YACP,OAAO,IAAI,CAAC;QACd;YACE,OAAO,IAAI,CAAC;IAChB,CAAC;AACH,CAAC;AAED,SAAS,sBAAsB,CAAC,KAAa;IAC3C,IAAI,MAAM,GAAG,EAAE,CAAC;IAChB,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QACrD,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC;QAC1B,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;YAClB,MAAM,IAAI,IAAI,CAAC;YACf,SAAS;QACX,CAAC;QAED,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;QAC9B,IAAI,CAAC,IAAI,EAAE,CAAC;YACV,MAAM;QACR,CAAC;QAED,IAAI,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YACvB,IAAI,KAAK,GAAG,IAAI,CAAC;YACjB,IAAI,QAAQ,GAAG,CAAC,CAAC;YACjB,OAAO,QAAQ,GAAG,CAAC,IAAI,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,GAAG,QAAQ,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC;gBACvE,KAAK,IAAI,KAAK,CAAC,KAAK,GAAG,CAAC,GAAG,QAAQ,CAAC,CAAC;gBACrC,QAAQ,IAAI,CAAC,CAAC;YAChB,CAAC;YACD,MAAM,IAAI,MAAM,CAAC,YAAY,CAAC,MAAM,CAAC,QAAQ,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,CAAC;YACzD,KAAK,IAAI,QAAQ,CAAC;YAClB,SAAS;QACX,CAAC;QAED,MAAM,IAAI,uBAAuB,CAAC,IAAI,CAAC,CAAC;QACxC,KAAK,IAAI,CAAC,CAAC;IACb,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,kBAAkB,CAAC,KAAa;IACvC,MAAM,UAAU,GAAG,KAAK,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;IAC7C,MAAM,MAAM,GAAG,UAAU,CAAC,MAAM,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,UAAU,GAAG,CAAC;IAC3E,IAAI,MAAM,GAAG,EAAE,CAAC;IAChB,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,MAAM,CAAC,MAAM,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QACtD,MAAM,KAAK,GAAG,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,GAAG,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;QAClE,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC;YACzB,MAAM,IAAI,MAAM,CAAC,YAAY,CAAC,KAAK,CAAC,CAAC;QACvC,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,0BAA0B,CAAC,IAAY,EAAE,UAAkB;IAClE,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,OAAO,GAAG,EAAE,CAAC;IACjB,KAAK,IAAI,KAAK,GAAG,UAAU,GAAG,CAAC,EAAE,KAAK,GAAG,IAAI,CAAC,MAAM,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QACjE,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC;QACzB,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;YAClB,OAAO,IAAI,IAAI,CAAC;YAChB,IAAI,KAAK,GAAG,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;gBAC5B,OAAO,IAAI,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;gBAC3B,KAAK,IAAI,CAAC,CAAC;YACb,CAAC;YACD,SAAS;QACX,CAAC;QACD,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjB,KAAK,IAAI,CAAC,CAAC;YACX,OAAO,IAAI,IAAI,CAAC;YAChB,SAAS;QACX,CAAC;QACD,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjB,KAAK,IAAI,CAAC,CAAC;YACX,IAAI,KAAK,KAAK,CAAC,EAAE,CAAC;gBAChB,OAAO;oBACL,KAAK,EAAE,OAAO;oBACd,QAAQ,EAAE,KAAK;iBAChB,CAAC;YACJ,CAAC;YACD,OAAO,IAAI,IAAI,CAAC;YAChB,SAAS;QACX,CAAC;QACD,OAAO,IAAI,IAAI,CAAC;IAClB,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,eAAe,CAAC,GAAW;IAClC,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,OAAO,KAAK,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC;QAC1B,MAAM,IAAI,GAAG,GAAG,CAAC,KAAK,CAAC,CAAC;QACxB,IAAI,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YACpB,KAAK,IAAI,CAAC,CAAC;YACX,SAAS;QACX,CAAC;QACD,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjB,MAAM,OAAO,GAAG,0BAA0B,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;YACvD,IAAI,CAAC,OAAO,EAAE,CAAC;gBACb,MAAM;YACR,CAAC;YACD,MAAM,CAAC,IAAI,CAAC,sBAAsB,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC;YACnD,KAAK,GAAG,OAAO,CAAC,QAAQ,GAAG,CAAC,CAAC;YAC7B,SAAS;QACX,CAAC;QACD,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjB,MAAM,QAAQ,GAAG,GAAG,CAAC,OAAO,CAAC,GAAG,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC;YAC7C,IAAI,QAAQ,GAAG,CAAC,EAAE,CAAC;gBACjB,MAAM;YACR,CAAC;YACD,MAAM,CAAC,IAAI,CAAC,kBAAkB,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,EAAE,QAAQ,CAAC,CAAC,CAAC,CAAC;YAChE,KAAK,GAAG,QAAQ,GAAG,CAAC,CAAC;YACrB,SAAS;QACX,CAAC;QAED,MAAM,cAAc,GAAG,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;QACrD,IAAI,cAAc,GAAG,CAAC,EAAE,CAAC;YACvB,MAAM;QACR,CAAC;QACD,KAAK,IAAI,cAAc,GAAG,CAAC,CAAC;IAC9B,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,2BAA2B,CAAC,OAAe;IAClD,MAAM,KAAK,GAAa,EAAE,CAAC;IAE3B,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,OAAO,CAAC,MAAM,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QACvD,MAAM,IAAI,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC;QAC5B,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjB,MAAM,OAAO,GAAG,0BAA0B,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;YAC3D,IAAI,CAAC,OAAO,EAAE,CAAC;gBACb,SAAS;YACX,CAAC;YAED,MAAM,eAAe,GAAG,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,QAAQ,GAAG,CAAC,EAAE,OAAO,CAAC,QAAQ,GAAG,CAAC,CAAC,CAAC;YAClF,IAAI,SAAS,CAAC,IAAI,CAAC,eAAe,CAAC,IAAI,MAAM,CAAC,IAAI,CAAC,eAAe,CAAC,IAAI,MAAM,CAAC,IAAI,CAAC,eAAe,CAAC,EAAE,CAAC;gBACpG,KAAK,CAAC,IAAI,CAAC,sBAAsB,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC;YACpD,CAAC;YACD,KAAK,GAAG,OAAO,CAAC,QAAQ,CAAC;YACzB,SAAS;QACX,CAAC;QAED,IAAI,IAAI,KAAK,GAAG,IAAI,OAAO,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,GAAG,EAAE,CAAC;YAC/C,MAAM,QAAQ,GAAG,OAAO,CAAC,OAAO,CAAC,GAAG,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC;YACjD,IAAI,QAAQ,GAAG,CAAC,EAAE,CAAC;gBACjB,SAAS;YACX,CAAC;YACD,MAAM,eAAe,GAAG,OAAO,CAAC,KAAK,CAAC,QAAQ,GAAG,CAAC,EAAE,QAAQ,GAAG,CAAC,CAAC,CAAC;YAClE,IAAI,SAAS,CAAC,IAAI,CAAC,eAAe,CAAC,EAAE,CAAC;gBACpC,KAAK,CAAC,IAAI,CAAC,kBAAkB,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,EAAE,QAAQ,CAAC,CAAC,CAAC,CAAC;YACrE,CAAC;YACD,KAAK,GAAG,QAAQ,CAAC;YACjB,SAAS;QACX,CAAC;IACH,CAAC;IAED,MAAM,YAAY,GAAG,oBAAoB,CAAC;IAC1C,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,QAAQ,CAAC,YAAY,CAAC,EAAE,CAAC;QACnD,MAAM,MAAM,GAAG,eAAe,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QACzC,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACtB,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;QAC9B,CAAC;IACH,CAAC;IAED,OAAO,gBAAgB,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;AAC5C,CAAC;AAED,SAAS,eAAe,CAAC,IAAmB,EAAE,OAAqC;IACjF,MAAM,SAAS,GAAa,EAAE,CAAC;IAC/B,KAAK,MAAM,UAAU,IAAI,IAAI,CAAC,WAAW,EAAE,CAAC;QAC1C,MAAM,YAAY,GAAG,OAAO,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;QAC7C,IAAI,CAAC,YAAY,EAAE,CAAC;YAClB,SAAS;QACX,CAAC;QACD,MAAM,UAAU,GAAG,eAAe,CAAC,YAAY,CAAC,CAAC;QACjD,IAAI,CAAC,UAAU,EAAE,CAAC;YAChB,SAAS;QACX,CAAC;QACD,MAAM,IAAI,GAAG,2BAA2B,CAAC,UAAU,CAAC,CAAC;QACrD,IAAI,IAAI,EAAE,CAAC;YACT,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACvB,CAAC;IACH,CAAC;IACD,OAAO,gBAAgB,CAAC,SAAS,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC;AAClD,CAAC;AAED,MAAM,OAAO,gBAAiB,SAAQ,wBAAwB;IACnD,IAAI,GAAG,YAAY,CAAC;IAE7B,QAAQ,CAAC,GAAW;QAClB,OAAO,wBAAwB,CAAC,GAAG,CAAC,GAAG,CAAC,WAAW,EAAE,CAAC,CAAC;IACzD,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,cAAc,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,EAC7C,eAAe,CAAC,iCAAiC,EACjD,EAAE,KAAK,EAAE,KAAK,EAAE,CACjB,CAAC;QACJ,CAAC;QAED,MAAM,OAAO,GAAG,UAAU,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;QAC9C,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,OAAO,CAAC,EAAE,CAAC;YACjC,MAAM,IAAI,QAAQ,CAChB,iBAAiB,EACjB,eAAe,CAAC,iCAAiC,CAClD,CAAC;QACJ,CAAC;QAED,MAAM,OAAO,GAAG,iBAAiB,CAAC,OAAO,CAAC,CAAC;QAC3C,IAAI,OAAO,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;YACvB,MAAM,IAAI,QAAQ,CAChB,eAAe,EACf,eAAe,CAAC,iCAAiC,CAClD,CAAC;QACJ,CAAC;QAED,MAAM,KAAK,GAAG,gBAAgB,CAAC,OAAO,CAAC,CAAC;QACxC,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACvB,MAAM,IAAI,QAAQ,CAChB,cAAc,EACd,eAAe,CAAC,iCAAiC,CAClD,CAAC;QACJ,CAAC;QAED,MAAM,MAAM,GAAsB,EAAE,CAAC;QACrC,KAAK,IAAI,SAAS,GAAG,CAAC,EAAE,SAAS,GAAG,KAAK,CAAC,MAAM,EAAE,SAAS,IAAI,CAAC,EAAE,CAAC;YACjE,MAAM,UAAU,GAAG,SAAS,GAAG,CAAC,CAAC;YACjC,MAAM,QAAQ,GAAG,eAAe,CAAC,KAAK,CAAC,SAAS,CAAC,EAAE,OAAO,CAAC,CAAC;YAC5D,IAAI,CAAC,QAAQ,EAAE,CAAC;gBACd,SAAS;YACX,CAAC;YACD,MAAM,CAAC,IAAI,CAAC;gBACV,IAAI,EAAE,MAAM;gBACZ,IAAI,EAAE,UAAU;gBAChB,IAAI,EAAE,QAAQ;gBACd,QAAQ,EAAE;oBACR,UAAU;iBACX;aACF,CAAC,CAAC;QACL,CAAC;QAED,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACxB,MAAM,IAAI,QAAQ,CAChB,iBAAiB,EACjB,eAAe,CAAC,iCAAiC,CAClD,CAAC;QACJ,CAAC;QAED,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAAC;QACvF,MAAM,KAAK,GAAG,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,QAAQ,EAAE,KAAK,CAAC,SAAS,CAAC,CAAC;QAE7D,OAAO;YACL,KAAK;YACL,IAAI;YACJ,OAAO,EAAE,YAAY,CAAC,IAAI,CAAC;YAC3B,MAAM,EAAE,KAAK;YACb,QAAQ,EAAE;gBACR,OAAO,EAAE,IAAI,CAAC,IAAI;gBAClB,SAAS,EAAE,KAAK,CAAC,MAAM;gBACvB,kBAAkB,EAAE,MAAM,CAAC,MAAM;aAClC;YACD,UAAU,EAAE;gBACV,MAAM;gBACN,KAAK,EAAE;oBACL,SAAS,EAAE,KAAK,CAAC,MAAM;oBACvB,kBAAkB,EAAE,MAAM,CAAC,MAAM;iBAClC;aACF;SACF,CAAC;IACJ,CAAC;CACF"}

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

@@ -0,0 +1,10 @@
+import type { ParseInput, ParsedDocumentPayload, ParserAdapter, ParserAvailability } from "./parser-adapter.js";
+/**
+ * 轻量文本解析适配器。
+ */
+export declare class PlainTextParserAdapter implements ParserAdapter {
+    readonly name = "plain_text";
+    supports(ext: string): boolean;
+    availability(): Promise<ParserAvailability>;
+    parse(input: ParseInput): Promise<ParsedDocumentPayload>;
+}

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

@@ -0,0 +1,55 @@
+import fs from "node:fs";
+import path from "node:path";
+const SUPPORTED_TEXT_EXTENSIONS = new Set([
+    ".md",
+    ".mdx",
+    ".txt",
+    ".rtf",
+    ".html",
+    ".htm",
+    ".xml",
+    ".json",
+    ".yaml",
+    ".yml",
+    ".tsv",
+]);
+function normalizeWhitespace(text) {
+    return text.replace(/\s+/g, " ").trim();
+}
+function shortSummary(text, limit = 180) {
+    const normalized = normalizeWhitespace(text);
+    if (normalized.length <= limit) {
+        return normalized;
+    }
+    return `${normalized.slice(0, limit - 1)}…`;
+}
+/**
+ * 轻量文本解析适配器。
+ */
+export class PlainTextParserAdapter {
+    name = "plain_text";
+    supports(ext) {
+        return SUPPORTED_TEXT_EXTENSIONS.has(ext.toLowerCase());
+    }
+    async availability() {
+        return "available";
+    }
+    async parse(input) {
+        const raw = fs.readFileSync(input.filePath, "utf-8");
+        const title = path.basename(input.filePath, input.extension);
+        return {
+            title,
+            text: raw,
+            summary: shortSummary(raw),
+            parser: input.extension === ".html" || input.extension === ".htm"
+                ? "html_fallback"
+                : input.extension === ".json" || input.extension === ".yaml" || input.extension === ".yml" || input.extension === ".xml"
+                    ? "structured_text_fallback"
+                    : "plain_text",
+            metadata: {
+                adapter: this.name,
+            },
+        };
+    }
+}
+//# sourceMappingURL=plain-text-parser-adapter.js.map