einvoice/test/suite/einvoice_parsing/test.parse-08.xpath-evaluation.ts

import { expect, tap } from '@git.zone/tstest/tapbundle';
import * as einvoice from '../../../ts/index.js';
import * as plugins from '../../plugins.js';

tap.test('PARSE-08: XPath evaluation for e-invoice data extraction', async () => {
  console.log('Testing XPath-like data extraction from e-invoices...\n');
  
  // Test extracting specific fields from different invoice formats
  const invoiceExtractionTests = [
    {
      name: 'UBL Invoice field extraction',
      xml: `<?xml version="1.0" encoding="UTF-8"?>
<ubl:Invoice xmlns:ubl="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2"
             xmlns:cac="urn:oasis:names:specification:ubl:schema:xsd:CommonAggregateComponents-2"
             xmlns:cbc="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2">
  <cbc:ID>UBL-XPATH-001</cbc:ID>
  <cbc:IssueDate>2024-01-15</cbc:IssueDate>
  <cbc:DueDate>2024-02-15</cbc:DueDate>
  <cac:AccountingSupplierParty>
    <cac:Party>
      <cac:PartyName>
        <cbc:Name>XPath Test Supplier</cbc:Name>
      </cac:PartyName>
      <cac:PostalAddress>
        <cbc:StreetName>123 Test Street</cbc:StreetName>
        <cbc:CityName>Berlin</cbc:CityName>
        <cbc:PostalZone>10115</cbc:PostalZone>
        <cac:Country>
          <cbc:IdentificationCode>DE</cbc:IdentificationCode>
        </cac:Country>
      </cac:PostalAddress>
    </cac:Party>
  </cac:AccountingSupplierParty>
  <cac:AccountingCustomerParty>
    <cac:Party>
      <cac:PartyName>
        <cbc:Name>XPath Test Customer</cbc:Name>
      </cac:PartyName>
    </cac:Party>
  </cac:AccountingCustomerParty>
  <cac:InvoiceLine>
    <cbc:ID>1</cbc:ID>
    <cbc:InvoicedQuantity unitCode="EA">10</cbc:InvoicedQuantity>
    <cbc:LineExtensionAmount currencyID="EUR">100.00</cbc:LineExtensionAmount>
    <cac:Item>
      <cbc:Name>Test Product A</cbc:Name>
      <cbc:Description>Detailed description of product A</cbc:Description>
    </cac:Item>
  </cac:InvoiceLine>
  <cac:InvoiceLine>
    <cbc:ID>2</cbc:ID>
    <cbc:InvoicedQuantity unitCode="KG">5.5</cbc:InvoicedQuantity>
    <cbc:LineExtensionAmount currencyID="EUR">55.00</cbc:LineExtensionAmount>
    <cac:Item>
      <cbc:Name>Test Product B</cbc:Name>
    </cac:Item>
  </cac:InvoiceLine>
  <cac:LegalMonetaryTotal>
    <cbc:TaxInclusiveAmount currencyID="EUR">184.45</cbc:TaxInclusiveAmount>
  </cac:LegalMonetaryTotal>
</ubl:Invoice>`,
      expectedData: {
        id: 'UBL-XPATH-001',
        issueDate: '2024-01-15',
        dueDate: '2024-02-15',
        supplierName: 'XPath Test Supplier',
        customerName: 'XPath Test Customer',
        lineItemCount: 2,
        totalAmount: 184.45
      }
    },
    {
      name: 'CII Invoice field extraction',
      xml: `<?xml version="1.0" encoding="UTF-8"?>
<rsm:CrossIndustryInvoice 
  xmlns:rsm="urn:un:unece:uncefact:data:standard:CrossIndustryInvoice:100"
  xmlns:qdt="urn:un:unece:uncefact:data:standard:QualifiedDataType:100"
  xmlns:ram="urn:un:unece:uncefact:data:standard:ReusableAggregateBusinessInformationEntity:100"
  xmlns:udt="urn:un:unece:uncefact:data:standard:UnqualifiedDataType:100">
  <rsm:ExchangedDocument>
    <ram:ID>CII-XPATH-001</ram:ID>
    <ram:TypeCode>380</ram:TypeCode>
    <ram:IssueDateTime>
      <udt:DateTimeString format="102">20240115</udt:DateTimeString>
    </ram:IssueDateTime>
  </rsm:ExchangedDocument>
  <rsm:SupplyChainTradeTransaction>
    <ram:ApplicableHeaderTradeAgreement>
      <ram:SellerTradeParty>
        <ram:Name>CII XPath Supplier</ram:Name>
      </ram:SellerTradeParty>
      <ram:BuyerTradeParty>
        <ram:Name>CII XPath Customer</ram:Name>
      </ram:BuyerTradeParty>
    </ram:ApplicableHeaderTradeAgreement>
  </rsm:SupplyChainTradeTransaction>
</rsm:CrossIndustryInvoice>`,
      expectedData: {
        id: 'CII-XPATH-001',
        supplierName: 'CII XPath Supplier',
        customerName: 'CII XPath Customer'
      }
    }
  ];
  
  for (const test of invoiceExtractionTests) {
    console.log(`\n${test.name}:`);
    
    try {
      const invoice = new einvoice.EInvoice();
      await invoice.fromXmlString(test.xml);
      
      console.log('  ✓ Invoice parsed successfully');
      
      // Extract and verify data
      const extractedData: any = {
        id: invoice.id,
        issueDate: invoice.issueDate instanceof Date ? 
          invoice.issueDate.toISOString().split('T')[0] : 
          invoice.issueDate,
        supplierName: invoice.from?.name,
        customerName: invoice.to?.name,
        lineItemCount: invoice.items?.length || 0
      };
      
      if (invoice.dueDate) {
        extractedData.dueDate = invoice.dueDate instanceof Date ?
          invoice.dueDate.toISOString().split('T')[0] :
          invoice.dueDate;
      }
      
      if (invoice.totalGross) {
        extractedData.totalAmount = invoice.totalGross;
      }
      
      console.log('  Extracted data:');
      Object.entries(extractedData).forEach(([key, value]) => {
        if (value !== undefined) {
          console.log(`    ${key}: ${value}`);
        }
      });
      
      // Verify expected data
      if (test.expectedData) {
        Object.entries(test.expectedData).forEach(([key, expectedValue]) => {
          if (extractedData[key] !== undefined) {
            expect(extractedData[key]).toEqual(expectedValue);
          }
        });
      }
      
    } catch (error) {
      console.log(`  ✗ Error: ${error.message}`);
    }
  }
});

tap.test('PARSE-08: Complex data extraction scenarios', async () => {
  console.log('\nTesting complex data extraction scenarios...\n');
  
  // Test extracting nested and repeated data
  const complexInvoice = `<?xml version="1.0" encoding="UTF-8"?>
<ubl:Invoice xmlns:ubl="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2"
             xmlns:cac="urn:oasis:names:specification:ubl:schema:xsd:CommonAggregateComponents-2"
             xmlns:cbc="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2">
  <cbc:ID>COMPLEX-001</cbc:ID>
  <cbc:IssueDate>2024-01-01</cbc:IssueDate>
  <cbc:Note>First note</cbc:Note>
  <cbc:Note>Second note</cbc:Note>
  <cbc:Note>Third note with special chars: €, ñ, 中文</cbc:Note>
  <cac:AccountingSupplierParty>
    <cac:Party>
      <cac:PartyIdentification>
        <cbc:ID schemeID="GLN">1234567890123</cbc:ID>
      </cac:PartyIdentification>
      <cac:PartyIdentification>
        <cbc:ID schemeID="DUNS">123456789</cbc:ID>
      </cac:PartyIdentification>
      <cac:PartyName>
        <cbc:Name>Complex Supplier Corp</cbc:Name>
      </cac:PartyName>
      <cac:Contact>
        <cbc:Name>John Doe</cbc:Name>
        <cbc:Telephone>+49 30 12345678</cbc:Telephone>
        <cbc:ElectronicMail>john.doe@supplier.com</cbc:ElectronicMail>
      </cac:Contact>
    </cac:Party>
  </cac:AccountingSupplierParty>
  ${Array.from({length: 5}, (_, i) => `
  <cac:InvoiceLine>
    <cbc:ID>${i + 1}</cbc:ID>
    <cbc:InvoicedQuantity unitCode="${i % 2 === 0 ? 'EA' : 'KG'}">${(i + 1) * 2}</cbc:InvoicedQuantity>
    <cbc:LineExtensionAmount currencyID="EUR">${((i + 1) * 50).toFixed(2)}</cbc:LineExtensionAmount>
    <cac:AllowanceCharge>
      <cbc:ChargeIndicator>false</cbc:ChargeIndicator>
      <cbc:Amount currencyID="EUR">${(i * 5).toFixed(2)}</cbc:Amount>
      <cbc:AllowanceChargeReason>Discount ${i + 1}</cbc:AllowanceChargeReason>
    </cac:AllowanceCharge>
    <cac:Item>
      <cbc:Name>Product ${String.fromCharCode(65 + i)}</cbc:Name>
      <cac:CommodityClassification>
        <cbc:ItemClassificationCode listID="CPV">12345678-${i}</cbc:ItemClassificationCode>
      </cac:CommodityClassification>
    </cac:Item>
  </cac:InvoiceLine>`).join('')}
</ubl:Invoice>`;
  
  try {
    const invoice = new einvoice.EInvoice();
    await invoice.fromXmlString(complexInvoice);
    
    console.log('Complex invoice extraction results:');
    console.log(`  Invoice ID: ${invoice.id}`);
    console.log(`  Notes count: ${invoice.notes?.length || 0}`);
    
    if (invoice.notes && invoice.notes.length > 0) {
      console.log('  Notes:');
      invoice.notes.forEach((note, index) => {
        console.log(`    ${index + 1}: ${note}`);
      });
    }
    
    console.log(`  Supplier identifiers: ${invoice.from?.identifiers?.length || 0}`);
    console.log(`  Line items: ${invoice.items?.length || 0}`);
    
    if (invoice.items && invoice.items.length > 0) {
      console.log('  Line item details:');
      invoice.items.forEach((item, index) => {
        console.log(`    Item ${index + 1}: ${item.name || 'Unknown'} - Qty: ${item.quantity || 0}`);
      });
    }
    
    console.log('  ✓ Complex data extraction successful');
    
  } catch (error) {
    console.log(`  ✗ Error: ${error.message}`);
  }
});

tap.test('PARSE-08: Performance of data extraction', async () => {
  console.log('\nTesting data extraction performance...\n');
  
  // Generate invoice with many fields to extract
  const generateDataRichInvoice = (complexity: string) => {
    const itemCount = complexity === 'simple' ? 5 : complexity === 'medium' ? 50 : 200;
    const noteCount = complexity === 'simple' ? 3 : complexity === 'medium' ? 10 : 30;
    
    return `<?xml version="1.0" encoding="UTF-8"?>
<ubl:Invoice xmlns:ubl="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2"
             xmlns:cac="urn:oasis:names:specification:ubl:schema:xsd:CommonAggregateComponents-2"
             xmlns:cbc="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2">
  <cbc:ID>PERF-${complexity.toUpperCase()}</cbc:ID>
  <cbc:IssueDate>2024-01-01</cbc:IssueDate>
  ${Array.from({length: noteCount}, (_, i) => `
  <cbc:Note>Note ${i + 1} with some content to extract</cbc:Note>`).join('')}
  <cac:AccountingSupplierParty>
    <cac:Party>
      <cac:PartyName>
        <cbc:Name>Performance Test Supplier</cbc:Name>
      </cac:PartyName>
    </cac:Party>
  </cac:AccountingSupplierParty>
  ${Array.from({length: itemCount}, (_, i) => `
  <cac:InvoiceLine>
    <cbc:ID>${i + 1}</cbc:ID>
    <cbc:InvoicedQuantity unitCode="EA">${i + 1}</cbc:InvoicedQuantity>
    <cbc:LineExtensionAmount currencyID="EUR">${((i + 1) * 10).toFixed(2)}</cbc:LineExtensionAmount>
    <cac:Item>
      <cbc:Name>Item ${i + 1}</cbc:Name>
    </cac:Item>
  </cac:InvoiceLine>`).join('')}
</ubl:Invoice>`;
  };
  
  const complexityLevels = ['simple', 'medium', 'complex'];
  
  for (const complexity of complexityLevels) {
    const xml = generateDataRichInvoice(complexity);
    const startTime = Date.now();
    
    try {
      const invoice = new einvoice.EInvoice();
      await invoice.fromXmlString(xml);
      
      // Extract various data points
      const extractedData = {
        id: invoice.id,
        issueDate: invoice.issueDate,
        supplierName: invoice.from?.name,
        noteCount: invoice.notes?.length || 0,
        itemCount: invoice.items?.length || 0,
        firstItemName: invoice.items?.[0]?.name,
        lastItemName: invoice.items?.[invoice.items.length - 1]?.name
      };
      
      const extractTime = Date.now() - startTime;
      
      console.log(`${complexity.charAt(0).toUpperCase() + complexity.slice(1)} invoice extraction:`);
      console.log(`  Extraction time: ${extractTime}ms`);
      console.log(`  Notes extracted: ${extractedData.noteCount}`);
      console.log(`  Items extracted: ${extractedData.itemCount}`);
      console.log(`  ✓ All data points extracted successfully`);
      
    } catch (error) {
      console.log(`  ✗ Error: ${error.message}`);
    }
  }
});

tap.test('PARSE-08: Special extraction scenarios', async () => {
  console.log('\nTesting special extraction scenarios...\n');
  
  // Test extracting data with special characters and edge cases
  const specialCases = [
    {
      name: 'Invoice with empty fields',
      xml: `<?xml version="1.0" encoding="UTF-8"?>
<ubl:Invoice xmlns:ubl="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2"
             xmlns:cbc="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2">
  <cbc:ID></cbc:ID>
  <cbc:Note></cbc:Note>
  <cbc:Note>   </cbc:Note>
</ubl:Invoice>`,
      expectedBehavior: 'Handle empty/whitespace fields gracefully'
    },
    {
      name: 'Invoice with CDATA sections',
      xml: `<?xml version="1.0" encoding="UTF-8"?>
<ubl:Invoice xmlns:ubl="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2"
             xmlns:cbc="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2">
  <cbc:ID>CDATA-001</cbc:ID>
  <cbc:Note><![CDATA[This contains <special> characters & symbols]]></cbc:Note>
</ubl:Invoice>`,
      expectedBehavior: 'Extract CDATA content correctly'
    },
    {
      name: 'Invoice with attributes',
      xml: `<?xml version="1.0" encoding="UTF-8"?>
<ubl:Invoice xmlns:ubl="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2"
             xmlns:cbc="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2">
  <cbc:ID schemeName="Invoice" schemeID="INT">ATTR-001</cbc:ID>
  <cbc:DocumentCurrencyCode listID="ISO4217">EUR</cbc:DocumentCurrencyCode>
</ubl:Invoice>`,
      expectedBehavior: 'Consider attribute values in extraction'
    }
  ];
  
  for (const testCase of specialCases) {
    console.log(`${testCase.name}:`);
    console.log(`  Expected: ${testCase.expectedBehavior}`);
    
    try {
      const invoice = new einvoice.EInvoice();
      await invoice.fromXmlString(testCase.xml);
      
      console.log(`  ID extracted: ${invoice.id || '(empty)'}`);
      console.log(`  Notes: ${invoice.notes?.length || 0} found`);
      
      if (invoice.notes && invoice.notes.length > 0) {
        invoice.notes.forEach((note, i) => {
          console.log(`    Note ${i + 1}: "${note}"`);
        });
      }
      
      console.log('  ✓ Special case handled successfully');
      
    } catch (error) {
      console.log(`  ℹ Parse result: ${error.message}`);
    }
  }
});

// Run the tests
tap.start();