einvoice/test/suite/einvoice_validation/test.val-04.schema-validation.ts

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

tap.test('VAL-04: XSD Schema Validation - should validate against XML Schema definitions', async () => {
  // Test schema validation for different formats
  const schemaTests = [
    {
      category: 'UBL_XMLRECHNUNG',
      schemaType: 'UBL 2.1',
      description: 'UBL invoices should validate against UBL 2.1 schema'
    },
    {
      category: 'CII_XMLRECHNUNG',
      schemaType: 'UN/CEFACT CII',
      description: 'CII invoices should validate against UN/CEFACT schema'
    },
    {
      category: 'EN16931_UBL_EXAMPLES',
      schemaType: 'UBL 2.1',
      description: 'EN16931 UBL examples should be schema-valid'
    }
  ] as const;

  console.log('Testing XSD schema validation across formats');

  const { EInvoice } = await import('../../../ts/index.js');
  let totalFiles = 0;
  let validFiles = 0;
  let invalidFiles = 0;
  let errorFiles = 0;

  for (const test of schemaTests) {
    try {
      const files = await CorpusLoader.getFiles(test.category);
      const xmlFiles = files.filter(f => f.endsWith('.xml')).slice(0, 3); // Test 3 per category

      if (xmlFiles.length === 0) {
        console.log(`\n${test.category}: No XML files found, skipping`);
        continue;
      }

      console.log(`\n${test.category} (${test.schemaType}): Testing ${xmlFiles.length} files`);

      for (const filePath of xmlFiles) {
        const fileName = path.basename(filePath);
        totalFiles++;

        try {
          const xmlContent = await fs.readFile(filePath, 'utf-8');

          const { result: einvoice } = await PerformanceTracker.track(
            'schema-xml-loading',
            async () => await EInvoice.fromXml(xmlContent)
          );

          // Perform schema validation (if available)
          const { result: validation } = await PerformanceTracker.track(
            'xsd-schema-validation',
            async () => {
              // Try to validate with schema validation level
              return await einvoice.validate(/* ValidationLevel.SCHEMA */);
            },
            { 
              category: test.category,
              file: fileName,
              schemaType: test.schemaType
            }
          );

          if (validation.valid) {
            validFiles++;
            console.log(`    ✓ ${fileName}: Schema valid`);
          } else {
            invalidFiles++;
            console.log(`    ○ ${fileName}: Schema validation failed`);
            if (validation.errors && validation.errors.length > 0) {
              const schemaErrors = validation.errors.filter(e => 
                e.message && (
                  e.message.toLowerCase().includes('schema') ||
                  e.message.toLowerCase().includes('xsd') ||
                  e.message.toLowerCase().includes('element')
                )
              );
              console.log(`      Schema errors: ${schemaErrors.length}`);
              schemaErrors.slice(0, 2).forEach(err => {
                console.log(`        - ${err.code}: ${err.message}`);
              });
            }
          }

        } catch (error) {
          errorFiles++;
          console.log(`    ✗ ${fileName}: Error - ${error.message}`);
        }
      }

    } catch (error) {
      console.log(`Error testing ${test.category}: ${error.message}`);
    }
  }

  console.log('\n=== XSD SCHEMA VALIDATION SUMMARY ===');
  console.log(`Total files tested: ${totalFiles}`);
  console.log(`Schema valid: ${validFiles}`);
  console.log(`Schema invalid: ${invalidFiles}`);
  console.log(`Errors: ${errorFiles}`);

  if (totalFiles > 0) {
    const validationRate = (validFiles / totalFiles * 100).toFixed(1);
    console.log(`Validation rate: ${validationRate}%`);

    // Performance summary
    const perfSummary = await PerformanceTracker.getSummary('xsd-schema-validation');
    if (perfSummary) {
      console.log(`\nSchema Validation Performance:`);
      console.log(`  Average: ${perfSummary.average.toFixed(2)}ms`);
      console.log(`  Min: ${perfSummary.min.toFixed(2)}ms`);
      console.log(`  Max: ${perfSummary.max.toFixed(2)}ms`);
      console.log(`  P95: ${perfSummary.p95.toFixed(2)}ms`);
    }

    // Expect most files to process successfully (valid or invalid, but not error)
    expect((validFiles + invalidFiles) / totalFiles).toBeGreaterThan(0.8);
  }
});

tap.test('VAL-04: Schema Validation Error Types - should identify different types of schema violations', async () => {
  const { EInvoice } = await import('../../../ts/index.js');

  const schemaViolationTests = [
    {
      name: 'Missing required element',
      xml: `<?xml version="1.0"?>
<Invoice xmlns="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2"
         xmlns:cbc="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2">
  <!-- Missing required ID element -->
  <cbc:IssueDate>2024-01-01</cbc:IssueDate>
</Invoice>`,
      violationType: 'missing-element'
    },
    {
      name: 'Invalid element order',
      xml: `<?xml version="1.0"?>
<Invoice xmlns="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2"
         xmlns:cbc="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2">
  <cbc:IssueDate>2024-01-01</cbc:IssueDate>
  <cbc:ID>WRONG-ORDER</cbc:ID> <!-- ID should come before IssueDate -->
</Invoice>`,
      violationType: 'element-order'
    },
    {
      name: 'Invalid data type',
      xml: `<?xml version="1.0"?>
<Invoice xmlns="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2"
         xmlns:cbc="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2">
  <cbc:ID>VALID-ID</cbc:ID>
  <cbc:IssueDate>not-a-date</cbc:IssueDate> <!-- Invalid date format -->
</Invoice>`,
      violationType: 'data-type'
    },
    {
      name: 'Unexpected element',
      xml: `<?xml version="1.0"?>
<Invoice xmlns="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2"
         xmlns:cbc="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2">
  <cbc:ID>VALID-ID</cbc:ID>
  <cbc:IssueDate>2024-01-01</cbc:IssueDate>
  <UnknownElement>Not allowed</UnknownElement> <!-- Not in schema -->
</Invoice>`,
      violationType: 'unexpected-element'
    }
  ];

  for (const test of schemaViolationTests) {
    try {
      const { result: validation } = await PerformanceTracker.track(
        'schema-violation-test',
        async () => {
          const einvoice = await EInvoice.fromXml(test.xml);
          return await einvoice.validate();
        }
      );

      console.log(`${test.name}: ${validation.valid ? 'VALID' : 'INVALID'}`);
      
      if (!validation.valid && validation.errors) {
        const schemaErrors = validation.errors.filter(e => 
          e.message && (
            e.message.toLowerCase().includes('schema') ||
            e.message.toLowerCase().includes('element') ||
            e.message.toLowerCase().includes('type')
          )
        );
        
        console.log(`  Schema errors detected: ${schemaErrors.length}`);
        schemaErrors.slice(0, 1).forEach(err => {
          console.log(`    - ${err.code}: ${err.message}`);
        });
        
        // Should detect schema violations
        expect(schemaErrors.length).toBeGreaterThan(0);
      } else {
        console.log(`  ○ No schema violations detected (may need stricter validation)`);
      }

    } catch (error) {
      console.log(`${test.name}: Error - ${error.message}`);
      // Parsing errors are also a form of schema violation
      console.log(`  ✓ Error during parsing indicates schema violation`);
    }
  }
});

tap.test('VAL-04: Schema Validation Performance - should validate schemas efficiently', async () => {
  const { EInvoice } = await import('../../../ts/index.js');

  // Generate test XMLs of different sizes
  function generateUBLInvoice(lineItems: 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>PERF-${Date.now()}</cbc:ID>
  <cbc:IssueDate>2024-01-01</cbc:IssueDate>
  <cbc:DocumentCurrencyCode>EUR</cbc:DocumentCurrencyCode>`;

    for (let i = 1; i <= lineItems; 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:InvoiceLine>`;
    }

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

  const performanceTests = [
    { name: 'Small invoice (5 lines)', lineItems: 5, threshold: 50 },
    { name: 'Medium invoice (25 lines)', lineItems: 25, threshold: 100 },
    { name: 'Large invoice (100 lines)', lineItems: 100, threshold: 200 }
  ];

  console.log('Testing schema validation performance');

  for (const test of performanceTests) {
    const xml = generateUBLInvoice(test.lineItems);
    console.log(`\n${test.name} (${Math.round(xml.length/1024)}KB)`);

    const { metric } = await PerformanceTracker.track(
      'schema-performance-test',
      async () => {
        const einvoice = await EInvoice.fromXml(xml);
        return await einvoice.validate();
      }
    );

    console.log(`  Validation time: ${metric.duration.toFixed(2)}ms`);
    console.log(`  Memory used: ${metric.memory ? (metric.memory.used / 1024 / 1024).toFixed(2) : 'N/A'}MB`);

    // Performance assertions
    expect(metric.duration).toBeLessThan(test.threshold);
    
    if (metric.memory && metric.memory.used > 0) {
      const memoryMB = metric.memory.used / 1024 / 1024;
      expect(memoryMB).toBeLessThan(100); // Should not use more than 100MB
    }
  }
});

tap.test('VAL-04: Schema Validation Caching - should cache schema validation results', async () => {
  const { EInvoice } = await import('../../../ts/index.js');

  const testXml = `<?xml version="1.0"?>
<Invoice xmlns="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2"
         xmlns:cbc="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2">
  <cbc:ID>CACHE-TEST</cbc:ID>
  <cbc:IssueDate>2024-01-01</cbc:IssueDate>
  <cbc:DocumentCurrencyCode>EUR</cbc:DocumentCurrencyCode>
</Invoice>`;

  console.log('Testing schema validation caching');

  const einvoice = await EInvoice.fromXml(testXml);

  // First validation (cold)
  const { metric: coldMetric } = await PerformanceTracker.track(
    'schema-validation-cold',
    async () => await einvoice.validate()
  );

  // Second validation (potentially cached)
  const { metric: warmMetric } = await PerformanceTracker.track(
    'schema-validation-warm',
    async () => await einvoice.validate()
  );

  console.log(`Cold validation: ${coldMetric.duration.toFixed(2)}ms`);
  console.log(`Warm validation: ${warmMetric.duration.toFixed(2)}ms`);

  // Warm validation should not be significantly slower
  const speedupRatio = coldMetric.duration / warmMetric.duration;
  console.log(`Speedup ratio: ${speedupRatio.toFixed(2)}x`);

  // Either caching helps (speedup) or both are fast
  const bothFast = coldMetric.duration < 20 && warmMetric.duration < 20;
  const cachingHelps = speedupRatio > 1.2;
  
  if (cachingHelps) {
    console.log('✓ Caching appears to improve performance');
  } else if (bothFast) {
    console.log('✓ Both validations are fast (caching may not be needed)');
  } else {
    console.log('○ Caching behavior unclear');
  }

  expect(bothFast || cachingHelps).toEqual(true);
});

tap.start();