einvoice/test/suite/einvoice_encoding/test.enc-05.special-characters.ts

import { expect, tap } from '@git.zone/tstest/tapbundle';
import { EInvoice } from '../../../ts/index.js';

tap.test('ENC-05: Special Characters - should handle special XML characters correctly', async () => {
  console.log('Testing special character handling in XML content...\n');

  // Test 1: Unicode special characters
  const testUnicodeSpecialChars = async () => {
    const einvoice = new EInvoice();
    einvoice.id = 'UNICODE-SPECIAL-TEST';
    einvoice.date = Date.now();
    einvoice.currency = 'EUR';
    
    // Test various special Unicode characters
    const specialChars = {
      mathematical: '∑∏∆∇∂∞≠≤≥±∓×÷√∝∴∵∠∟⊥∥∦',
      currency: '€£¥₹₽₩₪₨₫₡₢₣₤₥₦₧₨₩₪₫',
      symbols: '™®©℗℠⁋⁌⁍⁎⁏⁐⁑⁒⁓⁔⁕⁖⁗⁘⁙⁚⁛⁜⁝⁞',
      arrows: '←→↑↓↔↕↖↗↘↙↚↛↜↝↞↟↠↡↢↣↤↥',
      punctuation: '‚„"«»‹›§¶†‡•‰‱′″‴‵‶‷‸‼⁇⁈⁉⁊⁋⁌⁍⁎⁏'
    };
    
    einvoice.subject = `Unicode test: ${specialChars.mathematical.substring(0, 10)}`;
    einvoice.notes = [
      `Math: ${specialChars.mathematical}`,
      `Currency: ${specialChars.currency}`,
      `Symbols: ${specialChars.symbols}`,
      `Arrows: ${specialChars.arrows}`,
      `Punctuation: ${specialChars.punctuation}`
    ];
    
    einvoice.from = {
      type: 'company',
      name: 'Special Characters Inc ™',
      description: 'Company with special symbols: ®©',
      address: {
        streetName: 'Unicode Street ←→',
        houseNumber: '∞',
        postalCode: '12345',
        city: 'Symbol City ≤≥',
        country: 'DE'
      },
      status: 'active',
      foundedDate: { year: 2020, month: 1, day: 1 },
      registrationDetails: {
        vatId: 'DE123456789',
        registrationId: 'HRB 12345',
        registrationName: 'Special Registry ™'
      }
    };
    
    einvoice.to = {
      type: 'company',
      name: 'Customer Ltd ©',
      description: 'Customer with currency: €£¥',
      address: {
        streetName: 'Currency Ave',
        houseNumber: '€1',
        postalCode: '54321',
        city: 'Money City',
        country: 'DE'
      },
      status: 'active',
      foundedDate: { year: 2019, month: 1, day: 1 },
      registrationDetails: {
        vatId: 'DE987654321',
        registrationId: 'HRB 54321',
        registrationName: 'Customer Registry'
      }
    };
    
    einvoice.items = [{
      position: 1,
      name: 'Product with symbols: ∑∏∆',
      unitType: 'C62',
      unitQuantity: 1,
      unitNetPrice: 100,
      vatPercentage: 19
    }];
    
    const xmlString = await einvoice.toXmlString('ubl');
    
    // Check if special characters are preserved or properly encoded
    const mathPreserved = specialChars.mathematical.split('').filter(char => 
      xmlString.includes(char) || 
      xmlString.includes(`&#${char.charCodeAt(0)};`) ||
      xmlString.includes(`&#x${char.charCodeAt(0).toString(16)};`)
    ).length;
    
    const currencyPreserved = specialChars.currency.split('').filter(char => 
      xmlString.includes(char) || 
      xmlString.includes(`&#${char.charCodeAt(0)};`) ||
      xmlString.includes(`&#x${char.charCodeAt(0).toString(16)};`)
    ).length;
    
    const symbolsPreserved = specialChars.symbols.split('').filter(char => 
      xmlString.includes(char) || 
      xmlString.includes(`&#${char.charCodeAt(0)};`) ||
      xmlString.includes(`&#x${char.charCodeAt(0).toString(16)};`)
    ).length;
    
    return {
      mathPreserved,
      currencyPreserved,
      symbolsPreserved,
      totalMath: specialChars.mathematical.length,
      totalCurrency: specialChars.currency.length,
      totalSymbols: specialChars.symbols.length,
      xmlString
    };
  };

  const unicodeResult = await testUnicodeSpecialChars();
  console.log('Test 1 - Unicode special characters:');
  console.log(`  Mathematical symbols: ${unicodeResult.mathPreserved}/${unicodeResult.totalMath} preserved`);
  console.log(`  Currency symbols: ${unicodeResult.currencyPreserved}/${unicodeResult.totalCurrency} preserved`);
  console.log(`  Other symbols: ${unicodeResult.symbolsPreserved}/${unicodeResult.totalSymbols} preserved`);

  // Test 2: Control characters and whitespace
  const testControlCharacters = async () => {
    const einvoice = new EInvoice();
    einvoice.id = 'CONTROL-CHARS-TEST';
    einvoice.date = Date.now();
    einvoice.currency = 'EUR';
    
    // Test various whitespace and control characters
    einvoice.subject = 'Control chars test:\ttab\nnewline\rcarriage return';
    einvoice.notes = [
      'Tab separated:\tvalue1\tvalue2\tvalue3',
      'Line break:\nSecond line\nThird line',
      'Mixed whitespace:  spaces  \t tabs \r\n mixed'
    ];
    
    einvoice.from = {
      type: 'company',
      name: 'Control\tCharacters\nCompany',
      description: 'Company\twith\ncontrol\rcharacters',
      address: {
        streetName: 'Control Street',
        houseNumber: '1',
        postalCode: '12345',
        city: 'Test City',
        country: 'DE'
      },
      status: 'active',
      foundedDate: { year: 2020, month: 1, day: 1 },
      registrationDetails: {
        vatId: 'DE123456789',
        registrationId: 'HRB 12345',
        registrationName: 'Registry'
      }
    };
    
    einvoice.to = {
      type: 'company',
      name: 'Customer',
      description: 'Normal customer',
      address: {
        streetName: 'Customer Street',
        houseNumber: '2',
        postalCode: '54321',
        city: 'Customer City',
        country: 'DE'
      },
      status: 'active',
      foundedDate: { year: 2019, month: 1, day: 1 },
      registrationDetails: {
        vatId: 'DE987654321',
        registrationId: 'HRB 54321',
        registrationName: 'Customer Registry'
      }
    };
    
    einvoice.items = [{
      position: 1,
      name: 'Product\twith\ncontrol\rchars',
      unitType: 'C62',
      unitQuantity: 1,
      unitNetPrice: 100,
      vatPercentage: 19
    }];
    
    const xmlString = await einvoice.toXmlString('ubl');
    
    // Check how control characters are handled
    const hasTabHandling = xmlString.includes('	') || xmlString.includes('	') || 
                          xmlString.includes('\t') || !xmlString.includes('Control\tCharacters');
    const hasNewlineHandling = xmlString.includes('
') || xmlString.includes('
') || 
                              xmlString.includes('\n') || !xmlString.includes('Characters\nCompany');
    const hasCarriageReturnHandling = xmlString.includes('
') || xmlString.includes('
') || 
                                     xmlString.includes('\r') || !xmlString.includes('control\rcharacters');
    
    return {
      hasTabHandling,
      hasNewlineHandling,
      hasCarriageReturnHandling,
      xmlString
    };
  };

  const controlResult = await testControlCharacters();
  console.log('\nTest 2 - Control characters and whitespace:');
  console.log(`  Tab handling: ${controlResult.hasTabHandling ? 'Yes' : 'No'}`);
  console.log(`  Newline handling: ${controlResult.hasNewlineHandling ? 'Yes' : 'No'}`);
  console.log(`  Carriage return handling: ${controlResult.hasCarriageReturnHandling ? 'Yes' : 'No'}`);

  // Test 3: Emojis and extended Unicode
  const testEmojisAndExtended = async () => {
    const einvoice = new EInvoice();
    einvoice.id = 'EMOJI-TEST';
    einvoice.date = Date.now();
    einvoice.currency = 'EUR';
    
    // Test emojis and extended Unicode
    const emojis = '😀😃😄😁😆😅🤣😂🙂🙃😉😊😇🥰😍🤩😘😗☺😚😙🥲😋😛😜🤪😝🤑🤗🤭🤫🤔🤐🤨😐😑😶😏😒🙄😬🤥😌😔😪🤤😴😷🤒🤕🤢🤮🤧🥵🥶🥴😵🤯🤠🥳🥸😎🤓🧐😕😟🙁☹😮😯😲😳🥺😦😧😨😰😥😢😭😱😖😣😞😓😩😫🥱😤😡😠🤬😈👿💀☠💩🤡👹👺👻👽👾🤖😺😸😹😻😼😽🙀😿😾🙈🙉🙊💋💌💘💝💖💗💓💞💕💟❣💔❤🧡💛💚💙💜🤎🖤🤍💯💢💥💫💦💨🕳💣💬👁🗨🗯💭💤';
    
    einvoice.subject = `Emoji test: ${emojis.substring(0, 20)}`;
    einvoice.notes = [
      `Faces: ${emojis.substring(0, 50)}`,
      `Hearts: 💋💌💘💝💖💗💓💞💕💟❣💔❤🧡💛💚💙💜🤎🖤🤍`,
      `Objects: 💯💢💥💫💦💨🕳💣💬👁🗨🗯💭💤`
    ];
    
    einvoice.from = {
      type: 'company',
      name: 'Emoji Company 😊',
      description: 'Company with emojis 🏢',
      address: {
        streetName: 'Happy Street 😃',
        houseNumber: '1️⃣',
        postalCode: '12345',
        city: 'Emoji City 🌆',
        country: 'DE'
      },
      status: 'active',
      foundedDate: { year: 2020, month: 1, day: 1 },
      registrationDetails: {
        vatId: 'DE123456789',
        registrationId: 'HRB 12345',
        registrationName: 'Registry 📝'
      }
    };
    
    einvoice.to = {
      type: 'company',
      name: 'Customer 🛍️',
      description: 'Shopping customer',
      address: {
        streetName: 'Customer Street',
        houseNumber: '2',
        postalCode: '54321',
        city: 'Customer City',
        country: 'DE'
      },
      status: 'active',
      foundedDate: { year: 2019, month: 1, day: 1 },
      registrationDetails: {
        vatId: 'DE987654321',
        registrationId: 'HRB 54321',
        registrationName: 'Customer Registry'
      }
    };
    
    einvoice.items = [{
      position: 1,
      name: 'Emoji Product 📦',
      unitType: 'C62',
      unitQuantity: 1,
      unitNetPrice: 100,
      vatPercentage: 19
    }];
    
    const xmlString = await einvoice.toXmlString('ubl');
    
    // Check if emojis are preserved or encoded
    const emojiCount = emojis.split('').filter(char => {
      const codePoint = char.codePointAt(0);
      return codePoint && codePoint > 0xFFFF; // Emojis are typically above the BMP
    }).length;
    
    const preservedEmojis = emojis.split('').filter(char => {
      const codePoint = char.codePointAt(0);
      if (!codePoint || codePoint <= 0xFFFF) return false;
      return xmlString.includes(char) || 
             xmlString.includes(`&#${codePoint};`) ||
             xmlString.includes(`&#x${codePoint.toString(16)};`);
    }).length;
    
    return {
      emojiCount,
      preservedEmojis,
      preservationRate: emojiCount > 0 ? (preservedEmojis / emojiCount) * 100 : 0
    };
  };

  const emojiResult = await testEmojisAndExtended();
  console.log('\nTest 3 - Emojis and extended Unicode:');
  console.log(`  Emoji preservation: ${emojiResult.preservedEmojis}/${emojiResult.emojiCount} (${emojiResult.preservationRate.toFixed(1)}%)`);

  // Test 4: XML predefined entities in content
  const testXmlPredefinedEntities = async () => {
    const xmlWithEntities = `<?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>ENTITIES-TEST</cbc:ID>
  <cbc:IssueDate>2025-01-25</cbc:IssueDate>
  <cbc:Note>Entities: &amp; &lt; &gt; &quot; &apos;</cbc:Note>
  <cbc:InvoiceTypeCode>380</cbc:InvoiceTypeCode>
  <cbc:DocumentCurrencyCode>EUR</cbc:DocumentCurrencyCode>
  <cac:AccountingSupplierParty>
    <cac:Party>
      <cac:PartyName>
        <cbc:Name>Entity &amp; Company</cbc:Name>
      </cac:PartyName>
      <cac:PostalAddress>
        <cbc:StreetName>&lt;Special&gt; Street</cbc:StreetName>
        <cbc:CityName>Entity City</cbc:CityName>
        <cbc:PostalZone>12345</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>Customer &quot;Quotes&quot;</cbc:Name>
      </cac:PartyName>
      <cac:PostalAddress>
        <cbc:StreetName>Customer Street</cbc:StreetName>
        <cbc:CityName>Customer City</cbc:CityName>
        <cbc:PostalZone>54321</cbc:PostalZone>
        <cac:Country>
          <cbc:IdentificationCode>DE</cbc:IdentificationCode>
        </cac:Country>
      </cac:PostalAddress>
    </cac:Party>
  </cac:AccountingCustomerParty>
  <cac:InvoiceLine>
    <cbc:ID>1</cbc:ID>
    <cbc:InvoicedQuantity unitCode="C62">1</cbc:InvoicedQuantity>
    <cbc:LineExtensionAmount currencyID="EUR">100.00</cbc:LineExtensionAmount>
    <cac:Item>
      <cbc:Name>Product &apos;Apostrophe&apos;</cbc:Name>
    </cac:Item>
  </cac:InvoiceLine>
</Invoice>`;
    
    try {
      const invoice = await EInvoice.fromXml(xmlWithEntities);
      
      const supplierName = invoice.from?.name || '';
      const customerName = invoice.to?.name || '';
      const itemName = invoice.items?.[0]?.name || '';
      
      const entitiesDecoded = 
        supplierName.includes('Entity & Company') &&
        customerName.includes('Customer "Quotes"') &&
        itemName.includes("Product 'Apostrophe'");
      
      return {
        success: invoice.id === 'ENTITIES-TEST',
        entitiesDecoded,
        supplierName,
        customerName,
        itemName
      };
    } catch (error) {
      return {
        success: false,
        error: error.message
      };
    }
  };

  const entitiesResult = await testXmlPredefinedEntities();
  console.log('\nTest 4 - XML predefined entities:');
  console.log(`  Invoice parsed: ${entitiesResult.success ? 'Yes' : 'No'}`);
  console.log(`  Entities decoded: ${entitiesResult.entitiesDecoded ? 'Yes' : 'No'}`);
  if (entitiesResult.error) {
    console.log(`  Error: ${entitiesResult.error}`);
  }

  // Summary
  console.log('\n=== Special Characters Test Summary ===');
  const unicodeScore = (unicodeResult.mathPreserved + unicodeResult.currencyPreserved + unicodeResult.symbolsPreserved) / 
                      (unicodeResult.totalMath + unicodeResult.totalCurrency + unicodeResult.totalSymbols) * 100;
  console.log(`Unicode symbols: ${unicodeScore.toFixed(1)}% preserved`);
  console.log(`Control characters: ${controlResult.hasTabHandling && controlResult.hasNewlineHandling ? 'Handled' : 'Issues'}`);
  console.log(`Emojis: ${emojiResult.preservationRate.toFixed(1)}% preserved`);
  console.log(`XML entities: ${entitiesResult.success && entitiesResult.entitiesDecoded ? 'Working' : 'Issues'}`);
  
  // Tests pass if basic functionality works
  expect(unicodeScore).toBeGreaterThan(50); // At least 50% of Unicode symbols preserved
  expect(entitiesResult.success).toEqual(true);
  expect(entitiesResult.entitiesDecoded).toEqual(true);
  
  console.log('\n✓ Special characters test completed');
});

tap.start();