src/text-normalizer.ts

/**
 * TEXT NORMALIZATION UTILITIES
 *
 * Consolidates string cleanup operations into a single, maintainable module.
 * Replaces fragile .replace().replace().replace() chains throughout the codebase.
 */

/**
 * Normalize review language by removing weak phrases and standardizing terminology
 */
export function normalizeReviewLanguage(text: string): string {
  if (!text) return ''

  let normalized = text

  // Remove only the most problematic weak/hedging phrases
  const weakPhrases = [
    /\bmay cause (?:unexpected )?(?:issues?|problems?|behavior)\b/gi,
    /\bplease note,?\s*/gi,
    /\bit's recommended that\s*/gi,
    /\bit is recommended that\s*/gi
  ]

  for (const phrase of weakPhrases) {
    normalized = normalized.replace(phrase, '')
  }

  // Standardize terminology
  normalized = normalized.replace(
    /\binconsistent indentation\b/gi,
    'incorrect indentation'
  )

  // Remove excessive whitespace
  normalized = normalized.replace(/\s+/g, ' ').trim()

  return normalized
}

/**
 * Remove markdown formatting for plain text comparison
 */
export function stripMarkdown(text: string): string {
  if (!text) return ''

  return text
    .replace(/```[\s\S]*?```/g, '') // Remove code blocks
    .replace(/\*\*/g, '') // Remove bold
    .replace(/\*/g, '') // Remove italic
    .replace(/`/g, '') // Remove inline code
    .replace(/\[([^\]]+)\]\([^)]+\)/g, '$1') // Remove links
    .trim()
}

/**
 * Extract key terms from text (first 5 meaningful words)
 */
export function extractKeyTerms(text: string, limit: number = 5): string {
  if (!text) return ''

  const normalized = normalizeReviewLanguage(text)
  const words = normalized.split(' ').filter(w => w.length > 3)

  return words.slice(0, limit).join(' ')
}

/**
 * Clean description for display (removes code blocks, weak phrases)
 */
export function cleanDescription(text: string): string {
  if (!text) return ''

  let cleaned = stripMarkdown(text)
  cleaned = normalizeReviewLanguage(cleaned)

  // Remove confidence statements (handled separately)
  cleaned = cleaned.replace(/confidence\s*:\s*(high|medium|low)/gi, '')

  return cleaned.trim()
}

/**
 * Generate minimal, actionable description
 */
export function generateMinimalDescription(
  description: string,
  type: string,
  context?: {variableName?: string; functionName?: string}
): string {
  let minimal = cleanDescription(description)

  // Add context if available
  if (context?.functionName) {
    minimal = `In \`${context.functionName}\`, ${minimal.charAt(0).toLowerCase() + minimal.slice(1)}`
  }

  // Ensure it starts with capital letter
  minimal = minimal.charAt(0).toUpperCase() + minimal.slice(1)

  // Remove trailing period if present
  minimal = minimal.replace(/\.$/, '')

  return minimal
}