@arela/uploader 1.0.20 → 1.0.21

package/src/config/config.js

@@ -24,6 +24,7 @@ class Config {
     this.watch = this.#loadWatchConfig();
     this.redis = this.#loadRedisConfig();
     this.worker = this.#loadWorkerConfig();
+    this.gdrive = this.#loadGDriveConfig();
   }
 
   /**
@@ -36,10 +37,10 @@ class Config {
       const __dirname = path.dirname(__filename);
       const packageJsonPath = path.resolve(__dirname, '../../package.json');
       const packageJson = JSON.parse(fs.readFileSync(packageJsonPath, 'utf-8'));
-      return packageJson.version || '1.0.20';
+      return packageJson.version || '1.0.21';
     } catch (error) {
       console.warn('⚠️ Could not read package.json version, using fallback');
-      return '1.0.20';
+      return '1.0.21';
     }
   }
 
@@ -579,6 +580,91 @@ class Config {
     return process.env.ARELA_SERVER_ID || null;
   }
 
+  /**
+   * Load Google Drive sync configuration
+   * @private
+   */
+  #loadGDriveConfig() {
+    const rootFolderId = process.env.GDRIVE_ROOT_FOLDER_ID || null;
+
+    // Default mirror destination: <UPLOAD_BASE_PATH>/_gdrive_mirror
+    let localMirrorPath = process.env.GDRIVE_LOCAL_MIRROR_PATH || null;
+    if (!localMirrorPath && process.env.UPLOAD_BASE_PATH) {
+      const base = process.env.UPLOAD_BASE_PATH;
+      if (base !== '*') {
+        localMirrorPath = path.resolve(
+          PathNormalizer.toAbsolutePath(base),
+          '_gdrive_mirror',
+        );
+      }
+    } else if (localMirrorPath) {
+      localMirrorPath = PathNormalizer.toAbsolutePath(localMirrorPath);
+    }
+
+    return {
+      rootFolderId,
+      localMirrorPath,
+      serviceAccountFile: process.env.GDRIVE_SERVICE_ACCOUNT_FILE || null,
+      serviceAccountJson: process.env.GDRIVE_SERVICE_ACCOUNT_JSON || null,
+      skipNativeDocs: process.env.GDRIVE_SKIP_NATIVE_DOCS !== 'false',
+      followShortcuts: process.env.GDRIVE_FOLLOW_SHORTCUTS !== 'false',
+      concurrency: parseInt(process.env.GDRIVE_CONCURRENCY) || 5,
+      pageSize: parseInt(process.env.GDRIVE_PAGE_SIZE) || 1000,
+      maxFileSizeBytes:
+        parseInt(process.env.GDRIVE_MAX_FILE_SIZE_BYTES) ||
+        2 * 1024 * 1024 * 1024, // 2GB default
+    };
+  }
+
+  /**
+   * Get Google Drive configuration
+   * @returns {Object} GDrive sync settings
+   */
+  getGDriveConfig() {
+    return this.gdrive;
+  }
+
+  /**
+   * Validate Google Drive configuration
+   * @throws {Error} If required gdrive configuration is missing
+   */
+  validateGDriveConfig() {
+    const errors = [];
+
+    if (!this.gdrive.rootFolderId) {
+      errors.push(
+        'GDRIVE_ROOT_FOLDER_ID is required (Drive folder ID to sync)',
+      );
+    }
+
+    if (!this.gdrive.localMirrorPath) {
+      errors.push(
+        'Could not determine local mirror path. Set GDRIVE_LOCAL_MIRROR_PATH or UPLOAD_BASE_PATH',
+      );
+    }
+
+    if (!this.gdrive.serviceAccountFile && !this.gdrive.serviceAccountJson) {
+      errors.push(
+        'Either GDRIVE_SERVICE_ACCOUNT_FILE (path to JSON) or GDRIVE_SERVICE_ACCOUNT_JSON (inline JSON) is required',
+      );
+    }
+
+    if (this.gdrive.serviceAccountFile) {
+      const resolved = PathNormalizer.toAbsolutePath(
+        this.gdrive.serviceAccountFile,
+      );
+      if (!fs.existsSync(resolved)) {
+        errors.push(`GDRIVE_SERVICE_ACCOUNT_FILE not found: ${resolved}`);
+      }
+    }
+
+    if (errors.length > 0) {
+      throw new Error(
+        '⚠️ Google Drive configuration errors:\n   - ' + errors.join('\n   - '),
+      );
+    }
+  }
+
   /**
    * Check if worker mode is available (Redis configured)
    * @returns {boolean}

package/src/document-type-shared.js

@@ -2,6 +2,9 @@
 import { dodaPdfDefinition } from './document-types/doda-pdf.js';
 import { dodaXmlDefinition } from './document-types/doda-xml.js';
 import { facturasComerciales } from './document-types/facturas-comerciales.js';
+import { pedimentoCompletoDefinition } from './document-types/pedimento-completo.js';
+// TODO: enable XML pedimento detection — implementation ready in pedimento-completo-xml.js
+// import { pedimentoCompletoXmlDefinition } from './document-types/pedimento-completo-xml.js';
 import { pedimentoSimplificadoDefinition } from './document-types/pedimento-simplificado.js';
 import { proformaDefinition } from './document-types/proforma.js';
 import { supportDocumentDefinition } from './document-types/support-document.js';
@@ -41,6 +44,12 @@ export class DocumentTypeDefinition {
 // proformaDefinition is kept as reference but not used directly in the registry since resolution is handled post-extraction.
 const documentTypes = [
   pedimentoSimplificadoDefinition,
+  pedimentoCompletoDefinition,
+  // TODO: enable XML pedimento detection — uncomment the next line and the
+  // matching import at the top of this file. All downstream code
+  // (composeArelaPath, arela-api SQL filters, IdentifyCommand counters)
+  // already accepts `pedimento_completo_xml`.
+  // pedimentoCompletoXmlDefinition,
   supportDocumentDefinition,
   dodaPdfDefinition,
   dodaXmlDefinition,
@@ -96,12 +105,13 @@ export function extractDocumentFields(source, fileExtension, filePath) {
 
       console.log(`   → Resolved type: ${resolvedType}`);
 
-      // Extract pedimento number and year
+      // Extract pedimento number and year. `filePath` is forwarded so XML
+      // matchers (which compose numPedimento from the filename) can use it.
       const pedimento = docType.extractNumPedimento
-        ? docType.extractNumPedimento(source, fields)
+        ? docType.extractNumPedimento(source, fields, filePath)
         : null;
       const year = docType.extractPedimentoYear
-        ? docType.extractPedimentoYear(source, fields)
+        ? docType.extractPedimentoYear(source, fields, filePath)
         : null;
 
       return [resolvedType, fields, pedimento, year];

package/src/document-types/_pedimento-shared-extractors.js

@@ -0,0 +1,226 @@
+// Shared PDF extractors used by both `pedimento_simplificado` and
+// `pedimento_completo` matchers. Keeping the regexes in a single module
+// prevents drift between the two pedimento variants.
+import { FieldResult } from '../document-type-shared.js';
+
+// 1) Número de Pedimento (15 digits, possibly separated by spaces)
+export const numPedimentoExtractor = {
+  field: 'numPedimento',
+  extract: (source) => {
+    const match = source.match(/\d{2}\s?\d{2}\s?\d{4}\s?\d{7}/);
+    return new FieldResult(
+      'numPedimento',
+      !!match,
+      match ? match[0].replace(/\s/g, '') : null,
+    );
+  },
+};
+
+// 2) Tipo de Operación: 3 chars after the pedimento number
+export const tipoOperacionExtractor = {
+  field: 'tipoOperacion',
+  extract: (source) => {
+    const match = source.match(/\d{2}\s+\d{2}\s+\d{4}\s+\d{7}\s+([A-Z]{3})/);
+    return new FieldResult('tipoOperacion', !!match, match ? match[1] : null);
+  },
+};
+
+// 3) Clave de Pedimento: 2 chars after tipoOperacion
+export const clavePedimentoExtractor = {
+  field: 'clavePedimento',
+  extract: (source) => {
+    const match = source.match(
+      /\d{2}\s+\d{2}\s+\d{4}\s+\d{7}\s+[A-Z]{3}\s+([A-Z0-9]{2})/,
+    );
+    return new FieldResult('clavePedimento', !!match, match ? match[1] : null);
+  },
+};
+
+// 4) Aduana E/S: 3-digit code on the peso-bruto line
+export const aduanaEntradaSalidaExtractor = {
+  field: 'aduanaEntradaSalida',
+  extract: (source) => {
+    const match = source.match(/^\s*\d+\s+[\d,.]+\s+(\d{3})\s*$/m);
+    return new FieldResult(
+      'aduanaEntradaSalida',
+      !!match,
+      match ? match[1] : null,
+    );
+  },
+};
+
+// 5) RFC: 12–13 alphanumeric chars on its own line
+export const rfcExtractor = {
+  field: 'rfc',
+  extract: (source) => {
+    const match = source.match(/\n\s*([A-Z0-9]{12,13})\s*\n/);
+    return new FieldResult('rfc', !!match, match ? match[1] : null);
+  },
+};
+
+// 6) Código de Aceptación: 8 alphanumeric chars on the line right after the RFC
+export const codigoAceptacionExtractor = {
+  field: 'codigoAceptacion',
+  extract: (source) => {
+    const lines = source
+      .split(/\r?\n/)
+      .map((l) => l.trim())
+      .filter((l) => l.length > 0);
+
+    const rfcIndex = lines.findIndex((l) => /^[A-Z0-9]{12,13}$/.test(l));
+    let code = null;
+    if (rfcIndex >= 0 && /^[A-Z0-9]{8}$/.test(lines[rfcIndex + 1] || '')) {
+      code = lines[rfcIndex + 1];
+    }
+    return new FieldResult('codigoAceptacion', code !== null, code);
+  },
+};
+
+// 7) Num. E-Document: collects all 13-char alphanumeric codes following
+//    `NUM. E-DOCUMENT` labels. CoveFact / Pedimento Completo emit one row
+//    per ED clave inside the CLAVE/COMPL. IDENTIFICADOR table.
+export const numEDocumentoExtractor = {
+  field: 'numEDocumento',
+  extract: (source) => {
+    const lines = source.split(/\r?\n/);
+    const edocLines = lines.filter((line) => /NUM\.?\s*E-DOCUMENT/i.test(line));
+
+    if (edocLines.length === 0) {
+      return new FieldResult('numEDocumento', false, null);
+    }
+
+    const extractedCodes = [];
+    edocLines.forEach((line) => {
+      const afterEdoc = line.replace(/.*NUM\.?\s*E-DOCUMENT\s*/i, '');
+      const codes = afterEdoc.match(/[A-Z0-9]{13}/g) || [];
+      extractedCodes.push(...codes);
+    });
+
+    if (extractedCodes.length === 0) {
+      return new FieldResult('numEDocumento', false, null);
+    }
+
+    const uniqueCodes = [...new Set(extractedCodes)];
+    return new FieldResult('numEDocumento', true, `[${uniqueCodes.join(',')}]`);
+  },
+};
+
+// 8) Payment date — multiple known label variants
+export const paymentDateExtractor = {
+  field: 'paymentDate',
+  extract: (source) => {
+    let match = source.match(/2\s+PAGO:\s*(\d{2}\/\d{2}\/\d{4})/);
+    if (!match) {
+      match = source.match(/FECHA DE PAGO:\s*(\d{4}\/\d{2}\/\d{2})/);
+    }
+    if (!match) {
+      match = source.match(/PRESENTACION:\s*(\d{2}\/\d{2}\/\d{4})/);
+    }
+    return new FieldResult('paymentDate', !!match, match ? match[1] : null);
+  },
+};
+
+// 9) COVE — collect all `COVE<alphanum>` tokens from lines containing
+//    `COVE` or `NUMERO DE ACUSE DE VALOR`. CoveFact variant emits
+//    `COVE257W76NF2 / ID250230` → only the leading COVE token is kept
+//    because the COVE regex stops at the space before `/`.
+export const coveExtractor = {
+  field: 'cove',
+  extract: (source) => {
+    const lines = source.split(/\r?\n/);
+    const coveLines = lines.filter(
+      (line) => /COVE/i.test(line) || /NUMERO DE ACUSE DE VALOR/i.test(line),
+    );
+
+    if (coveLines.length === 0) {
+      return new FieldResult('cove', false, null);
+    }
+
+    const coveValues = [];
+    coveLines.forEach((line) => {
+      const coveMatches = line.match(/COVE[A-Z0-9]+/gi) || [];
+      coveValues.push(...coveMatches.map((m) => m.toUpperCase()));
+    });
+
+    if (coveValues.length === 0) {
+      return new FieldResult('cove', false, null);
+    }
+
+    const unique = [...new Set(coveValues)];
+    return new FieldResult('cove', true, `[${unique.join(',')}]`);
+  },
+};
+
+// 10) Patente: from the PATENTE/PEDIMENTO/ADUANA header table
+export const patenteExtractor = {
+  field: 'patente',
+  extract: (source) => {
+    const lines = source.split(/\r?\n/);
+    const patenteHeaderIndex = lines.findIndex((line) =>
+      /PATENTE:.*PEDIMENTO:.*ADUANA:/i.test(line),
+    );
+
+    if (patenteHeaderIndex >= 0) {
+      for (let i = patenteHeaderIndex + 1; i < lines.length; i++) {
+        const line = lines[i].trim();
+        if (/^\d+\s+\d+\s+\d+$/.test(line)) {
+          const parts = line.split(/\s+/);
+          return new FieldResult('patente', true, parts[0]);
+        }
+      }
+    }
+    return new FieldResult('patente', false, null);
+  },
+};
+
+// 11) Fecha de Pago Rectificación — used when clavePedimento is a rectification
+export const fechaPagoRectificacionExtractor = {
+  field: 'fechaPagoRectificacion',
+  extract: (source) => {
+    const rectSectionMatch = source.match(
+      /RECTIFICACION[\s\S]{0,500}?(\d{2}\/\d{2}\/\d{4})/i,
+    );
+    if (rectSectionMatch) {
+      return new FieldResult(
+        'fechaPagoRectificacion',
+        true,
+        rectSectionMatch[1],
+      );
+    }
+    const fechaMatch = source.match(
+      /FECHA PAGO RECT[\s\S]{0,500}?(\d{2}\/\d{2}\/\d{4})/i,
+    );
+    if (fechaMatch) {
+      return new FieldResult('fechaPagoRectificacion', true, fechaMatch[1]);
+    }
+    return new FieldResult('fechaPagoRectificacion', false, null);
+  },
+};
+
+/**
+ * Canonical extractor list for both pedimento_simplificado and pedimento_completo PDFs.
+ * Order matters only for downstream tooling that inspects the result array.
+ */
+export const sharedPedimentoExtractors = [
+  numPedimentoExtractor,
+  tipoOperacionExtractor,
+  clavePedimentoExtractor,
+  aduanaEntradaSalidaExtractor,
+  rfcExtractor,
+  codigoAceptacionExtractor,
+  numEDocumentoExtractor,
+  paymentDateExtractor,
+  coveExtractor,
+  patenteExtractor,
+  fechaPagoRectificacionExtractor,
+];
+
+/**
+ * Shared helper used by both PDF matchers' `extractPedimentoYear`.
+ */
+export function pedimentoYearFromFields(fields) {
+  const numPedimento = fields?.find((f) => f.name === 'numPedimento')?.value;
+  if (!numPedimento) return null;
+  const year = parseInt(numPedimento.substring(0, 2), 10);
+  return year < 50 ? year + 2000 : year + 1900;
+}

package/src/document-types/pedimento-completo-xml.js

@@ -0,0 +1,322 @@
+// VUCEM "consultarPedimentoCompleto" XML matcher.
+//
+// STATUS: implemented but NOT registered in `document-type-shared.js`. To
+// activate, uncomment the import + registration in that file. All downstream
+// code (composeArelaPath, arela-api propagation SQL, IdentifyCommand
+// counters) already includes `pedimento_completo_xml`, so re-enabling is a
+// single-line change.
+//
+// Filename patterns recognized (try in order — patente extraction):
+//   1) VU_PATENTE_ADUANA_PEDIMENTO.xml          → e.g. VU_3429_070_5016101.xml
+//   2) ADUANA-PATENTE-PEDIMENTO.xml             → e.g. 670-3806-2002487.xml
+//   3) {15-digit}[_{15-digit}].xml              → e.g. 260734296016642_260734296016642.xml
+// If none match, arela_path is left null and propagation fills it from a
+// sibling PDF in the same directory.
+//
+// numPedimento is composed from XML body + filename because the body never
+// carries the 15-digit form: YY|AA|PPPP|NNNNNNN.
+import { FieldResult } from '../document-type-shared.js';
+
+// --------------------------- helpers ---------------------------------------
+
+function firstTag(source, tag) {
+  // Match <ns2:tag>value</ns2:tag> or unprefixed <tag>value</tag>
+  const re = new RegExp(
+    `<(?:[a-z0-9]+:)?${tag}>([^<]*)</(?:[a-z0-9]+:)?${tag}>`,
+    'i',
+  );
+  const m = source.match(re);
+  return m ? m[1].trim() : null;
+}
+
+function allTagBlocks(source, tag) {
+  const re = new RegExp(
+    `<(?:[a-z0-9]+:)?${tag}>([\\s\\S]*?)</(?:[a-z0-9]+:)?${tag}>`,
+    'gi',
+  );
+  const out = [];
+  let m;
+  while ((m = re.exec(source)) !== null) {
+    out.push(m[1]);
+  }
+  return out;
+}
+
+function pad(value, length) {
+  if (value == null) return null;
+  return String(value).padStart(length, '0');
+}
+
+/**
+ * Try the three known filename patterns and return {patente, aduana, pedimento}
+ * with any subset of the fields populated. Returns null if no pattern matches.
+ */
+function parseFilenameParts(filePath) {
+  if (!filePath) return null;
+  const fileName = filePath.split(/[\\/]/).pop();
+  if (!fileName) return null;
+
+  let m;
+
+  // 1) VU_PATENTE_ADUANA_PEDIMENTO.xml
+  m = fileName.match(/^VU_(\d{4})_(\d{3})_(\d{7})\.xml$/i);
+  if (m) {
+    return {
+      patente: m[1],
+      aduana: m[2].substring(0, 2),
+      pedimento: m[3],
+      year: null,
+    };
+  }
+
+  // 2) ADUANA-PATENTE-PEDIMENTO.xml
+  m = fileName.match(/^(\d{3})-(\d{4})-(\d{7})\.xml$/i);
+  if (m) {
+    return {
+      patente: m[2],
+      aduana: m[1].substring(0, 2),
+      pedimento: m[3],
+      year: null,
+    };
+  }
+
+  // 3) 15-digit form YY|AA|PPPP|NNNNNNN
+  m = fileName.match(/^(\d{15})(?:_\d{15})?\.xml$/i);
+  if (m) {
+    const fifteen = m[1];
+    return {
+      year: fifteen.substring(0, 2),
+      aduana: fifteen.substring(2, 4),
+      patente: fifteen.substring(4, 8),
+      pedimento: fifteen.substring(8, 15),
+    };
+  }
+
+  return null;
+}
+
+// Capture the YY year from an ISO date string like "2026-05-08-06:00".
+function yyFromIsoDate(iso) {
+  if (!iso) return null;
+  const m = iso.match(/^(\d{4})-/);
+  return m ? m[1].substring(2, 4) : null;
+}
+
+// Find <ns2:fechas> block with nested clave==2 and return its <ns2:fecha>.
+function findPaymentDate(source) {
+  const fechasBlocks = allTagBlocks(source, 'fechas');
+  for (const block of fechasBlocks) {
+    const clave = firstTag(block, 'clave');
+    if (clave === '2') {
+      const fecha = firstTag(block, 'fecha');
+      if (fecha) return fecha;
+    }
+  }
+  return null;
+}
+
+// --------------------------- extractors ------------------------------------
+
+const rfcExtractor = {
+  field: 'rfc',
+  extract: (source) => {
+    const value = firstTag(source, 'rfc');
+    return new FieldResult('rfc', !!value, value);
+  },
+};
+
+const clavePedimentoExtractor = {
+  field: 'clavePedimento',
+  extract: (source) => {
+    // <ns2:claveDocumento><ns2:clave>R1</ns2:clave></ns2:claveDocumento>
+    const blocks = allTagBlocks(source, 'claveDocumento');
+    const clave = blocks.length > 0 ? firstTag(blocks[0], 'clave') : null;
+    return new FieldResult('clavePedimento', !!clave, clave);
+  },
+};
+
+const tipoOperacionExtractor = {
+  field: 'tipoOperacion',
+  extract: (source) => {
+    const blocks = allTagBlocks(source, 'tipoOperacion');
+    const desc = blocks.length > 0 ? firstTag(blocks[0], 'descripcion') : null;
+    return new FieldResult('tipoOperacion', !!desc, desc);
+  },
+};
+
+const aduanaEntradaSalidaExtractor = {
+  field: 'aduanaEntradaSalida',
+  extract: (source) => {
+    const blocks = allTagBlocks(source, 'aduanaEntradaSalida');
+    const clave = blocks.length > 0 ? firstTag(blocks[0], 'clave') : null;
+    return new FieldResult(
+      'aduanaEntradaSalida',
+      !!clave,
+      clave ? pad(clave, 2) : null,
+    );
+  },
+};
+
+const paymentDateExtractor = {
+  field: 'paymentDate',
+  extract: (source) => {
+    const fecha = findPaymentDate(source);
+    return new FieldResult('paymentDate', !!fecha, fecha);
+  },
+};
+
+const fechaPagoRectificacionExtractor = {
+  field: 'fechaPagoRectificacion',
+  extract: (source) => {
+    const rectBlocks = allTagBlocks(source, 'rectificacion');
+    if (rectBlocks.length === 0) {
+      return new FieldResult('fechaPagoRectificacion', false, null);
+    }
+    const fechaPago = firstTag(rectBlocks[0], 'fechaPago');
+    return new FieldResult('fechaPagoRectificacion', !!fechaPago, fechaPago);
+  },
+};
+
+const coveExtractor = {
+  field: 'cove',
+  extract: (source) => {
+    // Collect <ns2:numero> values that live inside <ns2:facturas> blocks.
+    const facturas = allTagBlocks(source, 'facturas');
+    const numeros = [];
+    facturas.forEach((block) => {
+      const numero = firstTag(block, 'numero');
+      if (numero) numeros.push(numero);
+    });
+    if (numeros.length === 0) {
+      return new FieldResult('cove', false, null);
+    }
+    const unique = [...new Set(numeros)];
+    return new FieldResult('cove', true, `[${unique.join(',')}]`);
+  },
+};
+
+const numEDocumentoExtractor = {
+  field: 'numEDocumento',
+  extract: (source) => {
+    // For each <ns2:identificadores> block, check the nested
+    // <ns2:claveIdentificador><ns2:clave> value. If it equals 'ED',
+    // collect the sibling <ns2:complemento1> value.
+    const blocks = allTagBlocks(source, 'identificadores');
+    const codes = [];
+    blocks.forEach((block) => {
+      const claveBlocks = allTagBlocks(block, 'claveIdentificador');
+      const clave =
+        claveBlocks.length > 0 ? firstTag(claveBlocks[0], 'clave') : null;
+      if (clave === 'ED') {
+        const complemento = firstTag(block, 'complemento1');
+        if (complemento) codes.push(complemento);
+      }
+    });
+    if (codes.length === 0) {
+      return new FieldResult('numEDocumento', false, null);
+    }
+    const unique = [...new Set(codes)];
+    return new FieldResult('numEDocumento', true, `[${unique.join(',')}]`);
+  },
+};
+
+// Composed numPedimento — needs the filename + already-extracted fields.
+// We expose it as the LAST extractor so `aduanaEntradaSalida` is available
+// via the `fields` array if the runner exposes it. To stay compatible with
+// the existing extractor signature (which only receives `source`), we
+// re-derive aduanaEntradaSalida inside this extractor and read the filename
+// from a closure populated by `match()` via `extractNumPedimento` below.
+// (See `extractNumPedimento` — that is the canonical place numPedimento is
+// composed for XML.)
+
+// --------------------------- definition ------------------------------------
+
+export const pedimentoCompletoXmlDefinition = {
+  type: 'pedimento_completo_xml',
+  extensions: ['xml'],
+
+  match: (source) => {
+    return /consultarPedimentoCompletoRespuesta/i.test(source);
+  },
+
+  /**
+   * Resolve to `pedimento_completo_xml` only if payment evidence exists:
+   *   - When a <rectificacion> block is present, require fechaPagoRectificacion.
+   *   - Otherwise require paymentDate.
+   */
+  resolveType: (fields) => {
+    const hasRect = !!fields?.find(
+      (f) => f.name === 'fechaPagoRectificacion' && f.found,
+    );
+    if (hasRect) return 'pedimento_completo_xml';
+
+    const paymentDate =
+      fields?.find((f) => f.name === 'paymentDate' && f.found)?.value ?? null;
+    return paymentDate ? 'pedimento_completo_xml' : 'proforma_completo_xml';
+  },
+
+  /**
+   * Compose the 15-digit pedimento number from XML body + filename.
+   *   YY: from rectification fechaPago if present, else from the clave==2
+   *       payment-date fecha; falls back to filename pattern 3.
+   *   AA: from <aduanaEntradaSalida><clave> padded to 2.
+   *   PPPP: from the filename (any of the three patterns).
+   *   NNNNNNN: from <pedimento> padded to 7.
+   * Returns null if any component cannot be resolved.
+   */
+  extractNumPedimento: (source, fields, filePath) => {
+    const parts = parseFilenameParts(filePath);
+
+    const rect = fields?.find(
+      (f) => f.name === 'fechaPagoRectificacion' && f.found,
+    )?.value;
+    const pay = fields?.find((f) => f.name === 'paymentDate' && f.found)?.value;
+
+    let yy =
+      yyFromIsoDate(rect) ||
+      yyFromIsoDate(pay) ||
+      (parts && parts.year) ||
+      null;
+
+    const aduanaField = fields?.find(
+      (f) => f.name === 'aduanaEntradaSalida',
+    )?.value;
+    let aduana = aduanaField || (parts && parts.aduana) || null;
+
+    let patente = parts && parts.patente ? parts.patente : null;
+
+    const pedimentoBody = firstTag(source, 'pedimento');
+    let pedimento = pedimentoBody
+      ? pad(pedimentoBody, 7)
+      : parts && parts.pedimento
+        ? parts.pedimento
+        : null;
+
+    if (!yy || !aduana || !patente || !pedimento) return null;
+
+    return `${pad(yy, 2)}${pad(aduana, 2)}${pad(patente, 4)}${pedimento}`;
+  },
+
+  extractPedimentoYear: (source, fields, filePath) => {
+    // Reuse extractNumPedimento; the year is its leading 2 digits.
+    const num = pedimentoCompletoXmlDefinition.extractNumPedimento(
+      source,
+      fields,
+      filePath,
+    );
+    if (!num) return null;
+    const year = parseInt(num.substring(0, 2), 10);
+    return year < 50 ? year + 2000 : year + 1900;
+  },
+
+  extractors: [
+    rfcExtractor,
+    clavePedimentoExtractor,
+    tipoOperacionExtractor,
+    aduanaEntradaSalidaExtractor,
+    paymentDateExtractor,
+    fechaPagoRectificacionExtractor,
+    coveExtractor,
+    numEDocumentoExtractor,
+  ],
+};