einvoice/test/suite/einvoice_parsing/test.parse-05.namespace-resolution.ts

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

tap.test('PARSE-05: Namespace Resolution - Handle XML namespaces correctly', async (t) => {
  const performanceTracker = new PerformanceTracker('PARSE-05');
  
  await t.test('Basic namespace declarations', async () => {
    performanceTracker.startOperation('basic-namespaces');
    
    const namespaceTests = [
      {
        name: 'Default namespace',
        xml: `<?xml version="1.0"?>
<Invoice xmlns="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2">
  <ID>TEST-001</ID>
  <IssueDate>2024-01-01</IssueDate>
</Invoice>`,
        expectedNamespaces: [{
          prefix: '',
          uri: 'urn:oasis:names:specification:ubl:schema:xsd:Invoice-2'
        }]
      },
      {
        name: 'Prefixed namespace',
        xml: `<?xml version="1.0"?>
<ubl:Invoice xmlns:ubl="urn:oasis:names:specification:ubl:schema:xsd:Invoice-2">
  <ubl:ID>TEST-002</ubl:ID>
  <ubl:IssueDate>2024-01-01</ubl:IssueDate>
</ubl:Invoice>`,
        expectedNamespaces: [{
          prefix: 'ubl',
          uri: 'urn:oasis:names:specification:ubl:schema:xsd:Invoice-2'
        }]
      },
      {
        name: 'Multiple namespaces',
        xml: `<?xml version="1.0"?>
<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>TEST-003</cbc:ID>
  <cac:AccountingSupplierParty>
    <cac:Party>
      <cbc:Name>Test Supplier</cbc:Name>
    </cac:Party>
  </cac:AccountingSupplierParty>
</ubl:Invoice>`,
        expectedNamespaces: [
          { prefix: 'ubl', uri: 'urn:oasis:names:specification:ubl:schema:xsd:Invoice-2' },
          { prefix: 'cac', uri: 'urn:oasis:names:specification:ubl:schema:xsd:CommonAggregateComponents-2' },
          { prefix: 'cbc', uri: 'urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2' }
        ]
      },
      {
        name: 'Namespace with schema location',
        xml: `<?xml version="1.0"?>
<Invoice 
  xmlns="http://www.example.com/invoice"
  xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
  xsi:schemaLocation="http://www.example.com/invoice invoice.xsd">
  <ID>TEST-004</ID>
</Invoice>`,
        expectedNamespaces: [
          { prefix: '', uri: 'http://www.example.com/invoice' },
          { prefix: 'xsi', uri: 'http://www.w3.org/2001/XMLSchema-instance' }
        ]
      }
    ];
    
    for (const test of namespaceTests) {
      const startTime = performance.now();
      
      console.log(`${test.name}:`);
      
      // Extract namespace declarations
      const namespaceMatches = test.xml.matchAll(/xmlns(?::([^=]+))?="([^"]+)"/g);
      const foundNamespaces = Array.from(namespaceMatches).map(match => ({
        prefix: match[1] || '',
        uri: match[2]
      }));
      
      console.log(`  Expected: ${test.expectedNamespaces.length} namespaces`);
      console.log(`  Found: ${foundNamespaces.length} namespaces`);
      
      for (const ns of foundNamespaces) {
        console.log(`    ${ns.prefix ? `${ns.prefix}:` : '(default)'} ${ns.uri}`);
      }
      
      // Verify parsing
      try {
        const invoice = new einvoice.EInvoice();
        if (invoice.fromXmlString) {
          await invoice.fromXmlString(test.xml);
          console.log('  ✓ Parsed successfully with namespaces');
        }
      } catch (error) {
        console.log(`  ✗ Parse error: ${error.message}`);
      }
      
      performanceTracker.recordMetric('namespace-declaration', performance.now() - startTime);
    }
    
    performanceTracker.endOperation('basic-namespaces');
  });
  
  await t.test('Namespace scope and inheritance', async () => {
    performanceTracker.startOperation('namespace-scope');
    
    const scopeTests = [
      {
        name: 'Namespace inheritance',
        xml: `<?xml version="1.0"?>
<root xmlns="http://example.com/default">
  <parent>
    <child>Inherits default namespace</child>
  </parent>
</root>`,
        description: 'Child elements inherit parent namespace'
      },
      {
        name: 'Namespace override',
        xml: `<?xml version="1.0"?>
<root xmlns="http://example.com/default">
  <parent>
    <child xmlns="http://example.com/child">Different namespace</child>
  </parent>
</root>`,
        description: 'Child can override inherited namespace'
      },
      {
        name: 'Mixed namespace scopes',
        xml: `<?xml version="1.0"?>
<root xmlns:a="http://example.com/a" xmlns:b="http://example.com/b">
  <a:element1>
    <a:child>Same namespace as parent</a:child>
    <b:child>Different namespace prefix</b:child>
    <unqualified>No namespace prefix</unqualified>
  </a:element1>
</root>`,
        description: 'Multiple namespace prefixes in scope'
      },
      {
        name: 'Namespace undeclaration',
        xml: `<?xml version="1.0"?>
<root xmlns="http://example.com/default">
  <parent>
    <child xmlns="">No namespace</child>
  </parent>
</root>`,
        description: 'Empty xmlns removes default namespace'
      }
    ];
    
    for (const test of scopeTests) {
      const startTime = performance.now();
      
      console.log(`${test.name}:`);
      console.log(`  Description: ${test.description}`);
      
      try {
        const invoice = new einvoice.EInvoice();
        if (invoice.fromXmlString) {
          await invoice.fromXmlString(test.xml);
          console.log('  ✓ Namespace scope handled correctly');
        }
      } catch (error) {
        console.log(`  ✗ Error: ${error.message}`);
      }
      
      performanceTracker.recordMetric('namespace-scope', performance.now() - startTime);
    }
    
    performanceTracker.endOperation('namespace-scope');
  });
  
  await t.test('Namespace prefix conflicts', async () => {
    performanceTracker.startOperation('namespace-conflicts');
    
    const conflictTests = [
      {
        name: 'Duplicate prefix - different URIs',
        xml: `<?xml version="1.0"?>
<root>
  <parent xmlns:ns="http://example.com/ns1">
    <ns:element1>Namespace 1</ns:element1>
    <child xmlns:ns="http://example.com/ns2">
      <ns:element2>Namespace 2 (redefined)</ns:element2>
    </child>
  </parent>
</root>`,
        issue: 'Same prefix maps to different URIs in nested scopes'
      },
      {
        name: 'Multiple prefixes - same URI',
        xml: `<?xml version="1.0"?>
<root xmlns:ns1="http://example.com/common" 
      xmlns:ns2="http://example.com/common">
  <ns1:element>Using ns1</ns1:element>
  <ns2:element>Using ns2 (same namespace)</ns2:element>
</root>`,
        issue: 'Different prefixes for the same namespace URI'
      },
      {
        name: 'Prefix collision with attributes',
        xml: `<?xml version="1.0"?>
<root xmlns:attr="http://example.com/attributes">
  <element attr:id="123" xmlns:attr="http://example.com/different">
    <attr:child>Which namespace?</attr:child>
  </element>
</root>`,
        issue: 'Attribute uses prefix before redefinition'
      }
    ];
    
    for (const test of conflictTests) {
      const startTime = performance.now();
      
      console.log(`${test.name}:`);
      console.log(`  Issue: ${test.issue}`);
      
      try {
        const invoice = new einvoice.EInvoice();
        if (invoice.fromXmlString) {
          await invoice.fromXmlString(test.xml);
          console.log('  ✓ Conflict handled gracefully');
        }
      } catch (error) {
        console.log(`  ⚠️  Parser warning: ${error.message}`);
      }
      
      performanceTracker.recordMetric('namespace-conflict', performance.now() - startTime);
    }
    
    performanceTracker.endOperation('namespace-conflicts');
  });
  
  await t.test('Common e-invoice namespace patterns', async () => {
    performanceTracker.startOperation('einvoice-namespaces');
    
    const einvoiceNamespaces = [
      {
        name: 'UBL Invoice',
        namespaces: {
          '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',
          'xmlns:ext': 'urn:oasis:names:specification:ubl:schema:xsd:CommonExtensionComponents-2'
        },
        rootElement: 'Invoice'
      },
      {
        name: 'Cross Industry Invoice (CII)',
        namespaces: {
          'xmlns:rsm': 'urn:un:unece:uncefact:data:standard:CrossIndustryInvoice:100',
          'xmlns:ram': 'urn:un:unece:uncefact:data:standard:ReusableAggregateBusinessInformationEntity:100',
          'xmlns:qdt': 'urn:un:unece:uncefact:data:standard:QualifiedDataType:100',
          'xmlns:udt': 'urn:un:unece:uncefact:data:standard:UnqualifiedDataType:100'
        },
        rootElement: 'rsm:CrossIndustryInvoice'
      },
      {
        name: 'FatturaPA',
        namespaces: {
          'xmlns:p': 'http://ivaservizi.agenziaentrate.gov.it/docs/xsd/fatture/v1.2',
          'xmlns:xsi': 'http://www.w3.org/2001/XMLSchema-instance'
        },
        rootElement: 'p:FatturaElettronica'
      },
      {
        name: 'PEPPOL BIS',
        namespaces: {
          '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'
        },
        rootElement: 'Invoice',
        profile: 'PEPPOL BIS Billing 3.0'
      }
    ];
    
    for (const format of einvoiceNamespaces) {
      console.log(`\n${format.name}:`);
      console.log(`  Root element: ${format.rootElement}`);
      if (format.profile) {
        console.log(`  Profile: ${format.profile}`);
      }
      console.log('  Namespaces:');
      
      for (const [attr, uri] of Object.entries(format.namespaces)) {
        const prefix = attr === 'xmlns' ? '(default)' : attr.replace('xmlns:', '');
        console.log(`    ${prefix}: ${uri}`);
      }
      
      // Generate sample XML
      const sampleXml = generateSampleXml(format);
      
      try {
        const invoice = new einvoice.EInvoice();
        if (invoice.fromXmlString) {
          await invoice.fromXmlString(sampleXml);
          console.log('  ✓ Sample parsed successfully');
        }
      } catch (error) {
        console.log(`  ⚠️  Parse issue: ${error.message}`);
      }
    }
    
    performanceTracker.endOperation('einvoice-namespaces');
  });
  
  await t.test('Namespace validation and well-formedness', async () => {
    performanceTracker.startOperation('namespace-validation');
    
    const validationTests = [
      {
        name: 'Undefined namespace prefix',
        xml: `<?xml version="1.0"?>
<root>
  <undefined:element>No namespace declaration for 'undefined'</undefined:element>
</root>`,
        valid: false,
        error: 'Undefined namespace prefix'
      },
      {
        name: 'Invalid namespace URI',
        xml: `<?xml version="1.0"?>
<root xmlns="not a valid URI">
  <element>Invalid namespace URI</element>
</root>`,
        valid: true, // XML parsers typically don't validate URI format
        error: null
      },
      {
        name: 'Reserved namespace prefix',
        xml: `<?xml version="1.0"?>
<root xmlns:xml="http://wrong.uri/xml">
  <xml:element>Wrong URI for xml prefix</xml:element>
</root>`,
        valid: false,
        error: 'xml prefix must be bound to http://www.w3.org/XML/1998/namespace'
      },
      {
        name: 'Circular namespace reference',
        xml: `<?xml version="1.0"?>
<ns1:root xmlns:ns1="http://example.com/ns1" xmlns:ns2="http://example.com/ns2">
  <ns2:element xmlns:ns1="http://example.com/different">
    <ns1:child>Which namespace?</ns1:child>
  </ns2:element>
</ns1:root>`,
        valid: true,
        error: null // Valid but potentially confusing
      }
    ];
    
    for (const test of validationTests) {
      const startTime = performance.now();
      
      console.log(`${test.name}:`);
      console.log(`  Expected: ${test.valid ? 'Valid' : 'Invalid'}`);
      if (test.error) {
        console.log(`  Expected error: ${test.error}`);
      }
      
      try {
        const invoice = new einvoice.EInvoice();
        if (invoice.fromXmlString) {
          await invoice.fromXmlString(test.xml);
          if (test.valid) {
            console.log('  ✓ Parsed as expected');
          } else {
            console.log('  ✗ Should have failed validation');
          }
        }
      } catch (error) {
        if (!test.valid) {
          console.log(`  ✓ Validation failed as expected: ${error.message}`);
        } else {
          console.log(`  ✗ Unexpected error: ${error.message}`);
        }
      }
      
      performanceTracker.recordMetric('namespace-validation', performance.now() - startTime);
    }
    
    performanceTracker.endOperation('namespace-validation');
  });
  
  await t.test('Corpus namespace analysis', async () => {
    performanceTracker.startOperation('corpus-namespaces');
    
    const corpusLoader = new CorpusLoader();
    const xmlFiles = await corpusLoader.getFiles(/\.(xml|ubl|cii)$/);
    
    console.log(`\nAnalyzing namespaces in ${xmlFiles.length} corpus files...`);
    
    const namespaceStats = {
      total: 0,
      byFormat: new Map<string, number>(),
      prefixUsage: new Map<string, number>(),
      uniqueURIs: new Set<string>(),
      avgNamespacesPerFile: 0,
      errors: 0
    };
    
    const sampleSize = Math.min(100, xmlFiles.length);
    const sampledFiles = xmlFiles.slice(0, sampleSize);
    let totalNamespaces = 0;
    
    for (const file of sampledFiles) {
      namespaceStats.total++;
      
      try {
        const content = await plugins.fs.readFile(file.path, 'utf8');
        
        // Extract all namespace declarations
        const namespaceMatches = content.matchAll(/xmlns(?::([^=]+))?="([^"]+)"/g);
        const namespaces = Array.from(namespaceMatches);
        
        totalNamespaces += namespaces.length;
        
        for (const match of namespaces) {
          const prefix = match[1] || '(default)';
          const uri = match[2];
          
          // Track prefix usage
          namespaceStats.prefixUsage.set(
            prefix,
            (namespaceStats.prefixUsage.get(prefix) || 0) + 1
          );
          
          // Track unique URIs
          namespaceStats.uniqueURIs.add(uri);
          
          // Detect format by namespace
          if (uri.includes('ubl:schema:xsd')) {
            namespaceStats.byFormat.set(
              'UBL',
              (namespaceStats.byFormat.get('UBL') || 0) + 1
            );
          } else if (uri.includes('uncefact:data:standard')) {
            namespaceStats.byFormat.set(
              'CII',
              (namespaceStats.byFormat.get('CII') || 0) + 1
            );
          } else if (uri.includes('agenziaentrate.gov.it')) {
            namespaceStats.byFormat.set(
              'FatturaPA',
              (namespaceStats.byFormat.get('FatturaPA') || 0) + 1
            );
          }
        }
      } catch (error) {
        namespaceStats.errors++;
      }
    }
    
    namespaceStats.avgNamespacesPerFile = totalNamespaces / namespaceStats.total;
    
    console.log('\nNamespace Statistics:');
    console.log(`Files analyzed: ${namespaceStats.total}`);
    console.log(`Average namespaces per file: ${namespaceStats.avgNamespacesPerFile.toFixed(2)}`);
    console.log(`Unique namespace URIs: ${namespaceStats.uniqueURIs.size}`);
    
    console.log('\nFormat detection by namespace:');
    for (const [format, count] of namespaceStats.byFormat.entries()) {
      console.log(`  ${format}: ${count} files`);
    }
    
    console.log('\nMost common prefixes:');
    const sortedPrefixes = Array.from(namespaceStats.prefixUsage.entries())
      .sort((a, b) => b[1] - a[1])
      .slice(0, 10);
    
    for (const [prefix, count] of sortedPrefixes) {
      console.log(`  ${prefix}: ${count} occurrences`);
    }
    
    console.log(`\nErrors: ${namespaceStats.errors}`);
    
    performanceTracker.endOperation('corpus-namespaces');
  });
  
  await t.test('Namespace resolution performance', async () => {
    performanceTracker.startOperation('namespace-performance');
    
    // Generate XML with varying namespace complexity
    const complexityLevels = [
      { namespaces: 1, elements: 10 },
      { namespaces: 5, elements: 50 },
      { namespaces: 10, elements: 100 },
      { namespaces: 20, elements: 200 }
    ];
    
    for (const level of complexityLevels) {
      const xml = generateComplexNamespaceXml(level.namespaces, level.elements);
      
      const startTime = performance.now();
      
      try {
        const invoice = new einvoice.EInvoice();
        if (invoice.fromXmlString) {
          await invoice.fromXmlString(xml);
        }
        
        const parseTime = performance.now() - startTime;
        
        console.log(`Complexity: ${level.namespaces} namespaces, ${level.elements} elements`);
        console.log(`  Parse time: ${parseTime.toFixed(2)}ms`);
        console.log(`  Time per element: ${(parseTime / level.elements).toFixed(3)}ms`);
        
        performanceTracker.recordMetric(`ns-complexity-${level.namespaces}`, parseTime);
      } catch (error) {
        console.log(`  Error: ${error.message}`);
      }
    }
    
    performanceTracker.endOperation('namespace-performance');
  });
  
  // Helper functions
  function generateSampleXml(format: any): string {
    const namespaceAttrs = Object.entries(format.namespaces)
      .map(([attr, uri]) => `${attr}="${uri}"`)
      .join('\n  ');
    
    return `<?xml version="1.0"?>
<${format.rootElement} ${namespaceAttrs}>
  <!-- Sample ${format.name} document -->
</${format.rootElement}>`;
  }
  
  function generateComplexNamespaceXml(nsCount: number, elemCount: number): string {
    let xml = '<?xml version="1.0"?>\n<root';
    
    // Add namespace declarations
    for (let i = 0; i < nsCount; i++) {
      xml += `\n  xmlns:ns${i}="http://example.com/namespace${i}"`;
    }
    xml += '>\n';
    
    // Add elements using various namespaces
    for (let i = 0; i < elemCount; i++) {
      const nsIndex = i % nsCount;
      xml += `  <ns${nsIndex}:element${i}>Content ${i}</ns${nsIndex}:element${i}>\n`;
    }
    
    xml += '</root>';
    return xml;
  }
  
  // Performance summary
  console.log('\n' + performanceTracker.getSummary());
  
  // Namespace resolution best practices
  console.log('\nNamespace Resolution Best Practices:');
  console.log('1. Always declare namespaces before use');
  console.log('2. Use consistent prefixes across documents');
  console.log('3. Avoid redefining prefixes in nested scopes');
  console.log('4. Validate namespace URIs match expected schemas');
  console.log('5. Handle both default and prefixed namespaces');
  console.log('6. Preserve namespace context for accurate processing');
  console.log('7. Support all common e-invoice namespace patterns');
  console.log('8. Optimize namespace resolution for large documents');
});

tap.start();