Method compilePattern

apps/sim/lib/chunkers/regex-chunker.ts:70–108

(pattern: string)

  }

  private compilePattern(pattern: string): RegExp {
    if (!pattern) {
      throw new Error('Regex pattern is required')
    }

    if (pattern.length > MAX_PATTERN_LENGTH) {
      throw new Error(`Regex pattern exceeds maximum length of ${MAX_PATTERN_LENGTH} characters`)
    }

    try {
      const regex = new RegExp(toNonCapturing(pattern), 'g')

      const testStrings = [
        'a'.repeat(10000),
        ' '.repeat(10000),
        'a '.repeat(5000),
        'aB1 xY2\n'.repeat(1250),
        `${'a'.repeat(30)}!`,
        `${'a b '.repeat(25)}!`,
      ]
      for (const testStr of testStrings) {
        regex.lastIndex = 0
        const start = Date.now()
        regex.test(testStr)
        const elapsed = Date.now() - start
        if (elapsed > 50) {
          throw new Error('Regex pattern appears to have catastrophic backtracking')
        }
      }

      regex.lastIndex = 0
      return regex
    } catch (error) {
      if (error instanceof Error && error.message.includes('catastrophic')) {
        throw error
      }
      throw new Error(`Invalid regex pattern "${pattern}": ${toError(error).message}`)
    }
  }

  async chunk(content: string): Promise<Chunk[]> {
    if (!content?.trim()) {