einvoice/test/suite/einvoice_format-detection/test.fd-07.edge-cases.ts

import { expect, tap } from '@git.zone/tstest/tapbundle';
import { CorpusLoader } from '../../helpers/corpus.loader.js';
import { PerformanceTracker } from '../../helpers/performance.tracker.js';

tap.test('FD-07: Edge Cases - should handle malformed and edge case inputs', async () => {
  const { FormatDetector } = await import('../../../ts/formats/utils/format.detector.js');

  // Test empty input
  const { result: emptyFormat } = await PerformanceTracker.track(
    'edge-case-detection',
    async () => FormatDetector.detectFormat('')
  );
  console.log(`Empty string: ${emptyFormat}`);
  expect(emptyFormat.toString().toLowerCase()).toEqual('unknown');

  // Test non-XML content
  const { result: textFormat } = await PerformanceTracker.track(
    'edge-case-detection',
    async () => FormatDetector.detectFormat('This is not XML content')
  );
  console.log(`Non-XML text: ${textFormat}`);
  expect(textFormat.toString().toLowerCase()).toEqual('unknown');

  // Test minimal XML
  const { result: minimalFormat } = await PerformanceTracker.track(
    'edge-case-detection',
    async () => FormatDetector.detectFormat('<?xml version="1.0"?><root></root>')
  );
  console.log(`Minimal XML: ${minimalFormat}`);
  expect(minimalFormat.toString().toLowerCase()).toEqual('unknown');

  // Test with BOM
  const bomXml = '\ufeff<?xml version="1.0"?><Invoice xmlns="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2"></Invoice>';
  const { result: bomFormat } = await PerformanceTracker.track(
    'edge-case-detection',
    async () => FormatDetector.detectFormat(bomXml)
  );
  console.log(`XML with BOM: ${bomFormat}`);
  expect(bomFormat.toString().toLowerCase()).toEqual('ubl');

  // Test malformed XML
  const malformedXml = '<?xml version="1.0"?><Invoice><unclosed>';
  const { result: malformedFormat } = await PerformanceTracker.track(
    'edge-case-detection',
    async () => FormatDetector.detectFormat(malformedXml)
  );
  console.log(`Malformed XML: ${malformedFormat}`);
  expect(malformedFormat.toString().toLowerCase()).toEqual('unknown');
});

tap.test('FD-07: Encoding Handling - should handle different character encodings', async () => {
  const { FormatDetector } = await import('../../../ts/formats/utils/format.detector.js');

  const encodingTests = [
    {
      name: 'UTF-8 with special characters',
      xml: `<?xml version="1.0" encoding="UTF-8"?>
<Invoice xmlns="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2">
  <ID>Tëst-Invöice-001</ID>
  <Note>Spëcial châractërs: àáâãäåæçèéêë</Note>
</Invoice>`,
      expectedFormat: 'ubl'
    },
    {
      name: 'ISO-8859-1 encoding declaration',
      xml: `<?xml version="1.0" encoding="ISO-8859-1"?>
<Invoice xmlns="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2">
  <ID>Test-001</ID>
</Invoice>`,
      expectedFormat: 'ubl'
    },
    {
      name: 'No encoding declaration',
      xml: `<?xml version="1.0"?>
<Invoice xmlns="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2">
  <ID>Test-002</ID>
</Invoice>`,
      expectedFormat: 'ubl'
    }
  ];

  for (const test of encodingTests) {
    const { result: format } = await PerformanceTracker.track(
      'encoding-detection',
      async () => FormatDetector.detectFormat(test.xml)
    );

    console.log(`${test.name}: ${format}`);
    expect(format.toString().toLowerCase()).toEqual(test.expectedFormat);
  }
});

tap.test('FD-07: Namespace Variations - should handle different namespace patterns', async () => {
  const { FormatDetector } = await import('../../../ts/formats/utils/format.detector.js');

  const namespaceTests = [
    {
      name: 'UBL with default namespace',
      xml: `<Invoice xmlns="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2">
        <ID>UBL-001</ID>
      </Invoice>`,
      expectedFormat: 'ubl'
    },
    {
      name: 'UBL with prefixed namespace',
      xml: `<ubl:Invoice xmlns:ubl="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2">
        <ubl:ID>UBL-002</ubl:ID>
      </ubl:Invoice>`,
      expectedFormat: 'ubl'
    },
    {
      name: 'CII with default namespace',
      xml: `<rsm:CrossIndustryInvoice xmlns:rsm="urn:un:unece:uncefact:data:standard:CrossIndustryInvoice:100">
        <rsm:ExchangedDocument/>
      </rsm:CrossIndustryInvoice>`,
      expectedFormat: 'cii'
    },
    {
      name: 'Mixed namespace prefixes',
      xml: `<inv:Invoice xmlns:inv="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2">
        <inv:ID>MIX-001</inv:ID>
      </inv:Invoice>`,
      expectedFormat: 'ubl'
    }
  ];

  for (const test of namespaceTests) {
    const { result: format } = await PerformanceTracker.track(
      'namespace-variation-detection',
      async () => FormatDetector.detectFormat(test.xml)
    );

    console.log(`${test.name}: ${format}`);
    const formatStr = format.toString().toLowerCase();
    const isExpectedFormat = formatStr.includes(test.expectedFormat) || 
                           (test.expectedFormat === 'cii' && formatStr.includes('cii'));
    expect(isExpectedFormat).toEqual(true);
  }
});

tap.test('FD-07: Large Input Stress Test - should handle very large XML inputs', async () => {
  const { FormatDetector } = await import('../../../ts/formats/utils/format.detector.js');

  // Generate a large UBL invoice with many line items
  function generateLargeUBL(itemCount: number): string {
    let xml = `<?xml version="1.0" encoding="UTF-8"?>
<Invoice xmlns="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>LARGE-TEST-${Date.now()}</cbc:ID>
  <cbc:IssueDate>2024-01-01</cbc:IssueDate>`;

    for (let i = 1; i <= itemCount; i++) {
      xml += `
  <cac:InvoiceLine>
    <cbc:ID>${i}</cbc:ID>
    <cbc:InvoicedQuantity unitCode="EA">${i}</cbc:InvoicedQuantity>
    <cbc:LineExtensionAmount currencyID="EUR">${i * 100}</cbc:LineExtensionAmount>
    <cac:Item>
      <cbc:Name>Product ${i}</cbc:Name>
      <cbc:Description>Description for product ${i} with some additional text to make it longer</cbc:Description>
    </cac:Item>
  </cac:InvoiceLine>`;
    }

    xml += '\n</Invoice>';
    return xml;
  }

  const testSizes = [
    { name: 'Small (10 items)', itemCount: 10 },
    { name: 'Medium (100 items)', itemCount: 100 },
    { name: 'Large (1000 items)', itemCount: 1000 }
  ];

  for (const test of testSizes) {
    const xml = generateLargeUBL(test.itemCount);
    const sizeKB = Math.round(xml.length / 1024);
    
    console.log(`Testing ${test.name} - ${sizeKB}KB`);

    // Test multiple times for accurate measurement
    const times: number[] = [];
    let detectedFormat = '';

    for (let i = 0; i < 3; i++) {
      const { result: format, metric } = await PerformanceTracker.track(
        'large-input-detection',
        async () => FormatDetector.detectFormat(xml)
      );
      
      times.push(metric.duration);
      detectedFormat = format.toString();
    }

    const avgTime = times.reduce((a, b) => a + b, 0) / times.length;
    console.log(`  Format: ${detectedFormat}`);
    console.log(`  Average time: ${avgTime.toFixed(2)}ms`);

    // Assertions
    expect(detectedFormat.toLowerCase()).toEqual('ubl');
    expect(avgTime).toBeLessThan(100); // Should be under 100ms even for large files
  }
});

tap.test('FD-07: Invalid Format Edge Cases - should handle unknown formats gracefully', async () => {
  const { FormatDetector } = await import('../../../ts/formats/utils/format.detector.js');

  const invalidTests = [
    {
      name: 'Valid XML, unknown invoice format',
      xml: `<?xml version="1.0"?>
<SomeRandomDocument>
  <ID>123</ID>
  <Data>Some data</Data>
</SomeRandomDocument>`
    },
    {
      name: 'HTML content',
      xml: `<!DOCTYPE html>
<html>
<head><title>Not XML</title></head>
<body><p>This is HTML</p></body>
</html>`
    },
    {
      name: 'JSON content',
      xml: `{"invoice": {"id": "123", "amount": 100}}`
    },
    {
      name: 'CSV content',
      xml: `ID,Amount,Currency
123,100,EUR
124,200,USD`
    }
  ];

  for (const test of invalidTests) {
    const { result: format } = await PerformanceTracker.track(
      'invalid-format-detection',
      async () => FormatDetector.detectFormat(test.xml)
    );

    console.log(`${test.name}: ${format}`);
    expect(format.toString().toLowerCase()).toEqual('unknown');
  }
});

tap.start();