Korean Bank SMS JSON Parser

sms-parser.ts

interface TransactionData {
  when: string;
  place: string;
  amount: number;
  balance: number;
}

class KoreanBankSMSParser {
  private static readonly AMOUNT_REGEX = /(?:출금|체크카드출금)\s*([\d,]+)원?/;
  private static readonly BALANCE_REGEX = /잔액\s*([\d,]+)원?/;
  private static readonly DATE_REGEX = /(\d{1,2}\/\d{1,2}(?:\s+\d{1,2}:\d{1,2})?|\d{4}\/\d{1,2}\/\d{1,2}\s+\d{1,2}:\d{1,2})/;
  private static readonly TIME_REGEX = /\d{1,2}:\d{2}/;
  private static readonly BANK_PATTERNS = /^\[?(KB|우리|농협|기업)\]?$/;
  private static readonly CARD_NUMBER_PATTERN = /^[\d\*\-]+$/;
  private static readonly TRANSACTION_PATTERN = /(?:출금|체크카드출금)\s*[\d,]+원?/;
  
  private static cleanNumber(str: string): number {
    return parseInt(str.replace(/[,원]/g, ''));
  }

  private static extractDateTime(text: string): string {
    const match = text.match(this.DATE_REGEX);
    if (!match) return '';
    
    // Normalize date format
    let dateStr = match[1];
    if (dateStr.includes('/')) {
      const parts = dateStr.split(' ');
      // If it's a full date (2024/11/15), take just month/day
      if (parts[0].split('/').length === 3) {
        dateStr = parts[0].split('/').slice(1).join('/') + (parts[1] ? ' ' + parts[1] : '');
      }
    }
    return dateStr;
  }

  private static isKoreanOrAlphanumeric(str: string): boolean {
    return /[\uAC00-\uD7AF\w]+/.test(str);
  }

  private static shouldSkipWord(word: string): boolean {
    if (!word || word === '[Web발신]') return true;
    
    const patterns = [
      this.DATE_REGEX,
      this.TIME_REGEX,
      this.BANK_PATTERNS,
      this.CARD_NUMBER_PATTERN,
      /^[\d,]+원?$/,
      /잔액/,
      /출금/,
      /^Web발신$/,
      /^\d+$/,
      /^[\d\*\-]+$/,
      /체크카드출금/
    ];
    
    return patterns.some(pattern => pattern.test(word));
  }

  private static extractPlace(text: string): string {
    // Split into lines and words, filtering out empty strings
    const lines = text.split('\n')
      .map(line => line.trim())
      .filter(line => line);
    
    // Look for merchant names in each line
    for (const line of lines) {
      // Skip lines that are purely transaction-related
      if (/^(?:출금|체크카드출금|잔액)\s*[\d,]+원?$/.test(line.trim())) {
        continue;
      }

      const words = line.split(/\s+/);
      
      for (const word of words) {
        if (this.shouldSkipWord(word)) continue;
        
        // Valid merchant name should contain Korean or alphanumeric characters
        // and be at least 2 characters long to avoid single-character noise
        if (this.isKoreanOrAlphanumeric(word) && word.length >= 2) {
          return word;
        }
      }
    }
    
    return '';
  }

  private static extractAmount(text: string): number {
    const match = text.match(this.AMOUNT_REGEX);
    if (!match) {
      // Try alternative pattern for simple "출금 amount원" format
      const altMatch = text.match(/출금\s*([\d,]+)원/);
      if (altMatch) {
        return this.cleanNumber(altMatch[1]);
      }
      return 0;
    }
    return this.cleanNumber(match[1]);
  }

  private static extractBalance(text: string): number {
    const match = text.match(this.BALANCE_REGEX);
    if (!match) return 0;
    return this.cleanNumber(match[1]);
  }

  public static parse(smsText: string): TransactionData {
    return {
      when: this.extractDateTime(smsText),
      place: this.extractPlace(smsText),
      amount: this.extractAmount(smsText),
      balance: this.extractBalance(smsText)
    };
  }
}

// Example usage:
const testSMSMessages = [
  `[Web발신]
[KB]11/18 14:58
279801**027
과일놀이터
체크카드출금
18,000
잔액238,281`,

  `[Web발신]
우리 11/17 18:17
*250284
출금 3,000원
서울특별시－현
잔액 214,164원`,

  `[Web발신]
2024/11/15 12:57
출금 354,594원
잔액 50,622원
현대카드
469***03801011
기업`,

  `[Web발신]
농협 출금2,500원
11/03 17:43 301-****-2640-41 파우PC 잔액5,428원`
];

// Process each message
const results = testSMSMessages.map(sms => KoreanBankSMSParser.parse(sms));

// Output results
console.log(JSON.stringify(results, null, 2));

🏦 Korean Bank SMS Parser

Transforms Korean bank transaction SMS into structured JSON. Handles multiple banks (KB, Woori, NH, IBK), various date formats, and merchant names with smart pattern matching.

// Input: Raw bank SMS
[Web발신]
[KB]11/18 14:58
과일놀이터
체크카드출금18,000
잔액238,281

// Output: Clean JSON
{
  "when": "11/18 14:58",
  "place": "과일놀이터",
  "amount": 18000,
  "balance": 238281
}

Features

• Zero dependencies
• Smart Korean text filtering
• Flexible date/amount parsing
• Merchant name extraction
• TypeScript ready