website-enchun-mgr/apps/backend/scripts/migration/csvParser.ts

/**
 * CSV Parser for Webflow Exports
 * Story 1.3: Content Migration Script
 *
 * Parses Webflow CSV export files and converts to WebflowExportData format
 */

import type { WebflowExportData, WebflowPost, WebflowCategory } from './types'
import { readFile } from 'fs/promises'
import { parse } from 'csv-parse/sync'

// ============================================================
// CSV ROW INTERFACES
// ============================================================

interface WebflowPostCsvRow {
  '文章標題': string
  'Slug': string
  'Collection ID': string
  'Item ID': string
  'Archived': string
  'Draft': string
  'Created On': string
  'Updated On': string
  'Published On': string
  '強調圖片': string
  'Open Graph 顯示圖片': string
  '文章簡述': string
  '發文日期': string
  '文章分類': string
  '發文內容': string
  '是否放在頁尾': string
}

interface WebflowCategoryCsvRow {
  name: string
  slug: string
  [key: string]: string
}

interface WebflowPortfolioCsvRow {
  Name: string
  Slug: string
  'website-link': string
  'preview-image': string
  description: string
  'website-type': string
  tags: string
  [key: string]: string
}

// ============================================================
// MAIN CSV PARSER
// ============================================================

/**
 * Parse Webflow CSV file and convert to WebflowExportData
 */
export async function parseWebflowCSV(filePath: string): Promise<WebflowExportData> {
  const content = await readFile(filePath, 'utf-8')
  const records: any[] = parse(content, {
    columns: true,
    skip_empty_lines: true,
    trim: true,
  })

  // Detect collection type from file name or headers
  if (filePath.includes('行銷放大鏡集') || records[0]?.['文章標題']) {
    return parsePostsCSV(records as WebflowPostCsvRow[])
  }

  if (filePath.includes('Categories') || filePath.includes('分類')) {
    return parseCategoriesCSV(records as WebflowCategoryCsvRow[])
  }

  if (filePath.includes('Portfolio') || filePath.includes('作品')) {
    return parsePortfolioCSV(records as WebflowPortfolioCsvRow[])
  }

  // Default: try to detect from structure
  if (records[0]?.['文章標題'] || records[0]?.['發文內容']) {
    return parsePostsCSV(records as WebflowPostCsvRow[])
  }

  return { posts: [], categories: [], portfolio: [] }
}

// ============================================================
// POSTS CSV PARSER
// ============================================================

/**
 * Parse Posts collection CSV
 * Webflow CSV headers: 文章標題, Slug, ..., 強調圖片, 發文日期, 文章分類, 發文內容, ...
 */
function parsePostsCSV(records: WebflowPostCsvRow[]): WebflowExportData {
  const posts: WebflowPost[] = []
  const categoryNames = new Set<string>()

  for (const row of records) {
    // Skip archived posts if needed
    if (row.Archived === 'true') continue

    // Extract category name
    const categoryName = row['文章分類'] || ''
    if (categoryName) {
      categoryNames.add(categoryName)
    }

    // Parse published date
    const publishedDate = parseWebflowDate(row['發文日期'] || row['Published On'] || row['Created On'])

    posts.push({
      title: row['文章標題'] || '',
      slug: row.Slug || '',
      content: row['發文內容'] || '',
      publishedDate,
      postCategory: categoryName || undefined,
      featuredImage: row['強調圖片'] || undefined,
      seoTitle: undefined, // Could be extracted from content if needed
      seoDescription: row['文章簡述'] || undefined,
      excerpt: row['文章簡述'] || undefined,
    })
  }

  // Generate categories from posts
  const categories = generateCategoriesFromPosts(Array.from(categoryNames))

  return { posts, categories, portfolio: [] }
}

// ============================================================
// CATEGORIES CSV PARSER
// ============================================================

/**
 * Parse Categories collection CSV
 */
function parseCategoriesCSV(records: WebflowCategoryCsvRow[]): WebflowExportData {
  const categories: WebflowCategory[] = []

  // Known categories with colors (from story requirements)
  const knownCategories: Record<string, string> = {
    'google-xiao-xue-tang': '#4285f4', // Google blue
    'google-workshop': '#4285f4',
    'meta-xiao-xue-tang': '#0668e1', // Meta blue
    'meta-workshop': '#0668e1',
    'xing-xiao-shi-shi-zui-qian-xian': '#34a853', // Green
    'marketing-news': '#34a853',
    'enchun-announcements': '#ea4335', // Red
    '恩群數位最新公告': '#ea4335',
  }

  for (const row of records) {
    const name = row.name || ''
    const slug = row.slug || ''

    categories.push({
      name,
      slug,
      colorHex: knownCategories[slug] || knownCategories[name] || '#0066cc',
    })
  }

  return { posts: [], categories, portfolio: [] }
}

// ============================================================
// PORTFOLIO CSV PARSER
// ============================================================

/**
 * Parse Portfolio collection CSV
 */
function parsePortfolioCSV(records: WebflowPortfolioCsvRow[]): WebflowExportData {
  const portfolio: any[] = []

  for (const row of records) {
    // Map website type strings to enum values
    const typeMapping: Record<string, 'corporate' | 'ecommerce' | 'landing' | 'brand' | 'other'> = {
      'corporate': 'corporate',
      'ecommerce': 'ecommerce',
      'landing': 'landing',
      'brand': 'brand',
    }

    const websiteType = typeMapping[row['website-type']?.toLowerCase()] || 'other'

    portfolio.push({
      name: row.Name || '',
      slug: row.Slug || '',
      websiteLink: row['website-link'] || '',
      previewImage: row['preview-image'] || '',
      description: row.description || '',
      websiteType,
      tags: row.tags || '',
    })
  }

  return { posts: [], categories: [], portfolio }
}

// ============================================================
// HELPER FUNCTIONS
// ============================================================

/**
 * Parse Webflow date format to Date object
 * Webflow dates: "Thu Jan 20 2022 00:00:00 GMT+0000 (Coordinated Universal Time)"
 */
function parseWebflowDate(dateStr: string): Date {
  if (!dateStr) return new Date()

  // Remove timezone info and parse
  const cleanDate = dateStr.replace(/\(.*\)$/, '').trim()
  const parsed = new Date(cleanDate)

  return isNaN(parsed.getTime()) ? new Date() : parsed
}

/**
 * Generate category objects from category names found in posts
 */
function generateCategoriesFromPosts(categoryNames: string[]): WebflowCategory[] {
  const nameToSlug: Record<string, string> = {
    'Google小學堂': 'google-xiao-xue-tang',
    'Meta小學堂': 'meta-xiao-xue-tang',
    '行銷時事最前線': 'xing-xiao-shi-shi-zui-qian-xian',
    '恩群數位最新公告': 'enchun-announcements',
  }

  const slugToColor: Record<string, string> = {
    'google-xiao-xue-tang': '#4285f4',
    'meta-xiao-xue-tang': '#0668e1',
    'xing-xiao-shi-shi-zui-qian-xian': '#34a853',
    'enchun-announcements': '#ea4335',
  }

  const categories: WebflowCategory[] = []
  const seen = new Set<string>()

  for (const name of categoryNames) {
    if (seen.has(name)) continue
    seen.add(name)

    const slug = nameToSlug[name] || toSlug(name)
    const colorHex = slugToColor[slug] || '#0066cc'

    categories.push({ name, slug, colorHex })
  }

  return categories
}

/**
 * Convert string to URL-friendly slug (Chinese-friendly)
 */
function toSlug(value: string): string {
  return value
    .toString()
    .toLowerCase()
    .trim()
    .normalize('NFD')
    .replace(/[\u0300-\u036f]/g, '')
    .replace(/[^a-z0-9\u4e00-\u9fa5/-]/g, '-')
    .replace(/-+/g, '-')
    .replace(/^-+|-+$/g, '')
}

// ============================================================
// BATCH CSV PARSER
// ============================================================

/**
 * Parse multiple CSV files at once
 */
export async function parseMultipleCSVs(filePaths: string[]): Promise<WebflowExportData> {
  const combined: WebflowExportData = {
    posts: [],
    categories: [],
    portfolio: [],
  }

  for (const filePath of filePaths) {
    try {
      const data = await parseWebflowCSV(filePath)

      if (data.posts) combined.posts?.push(...data.posts)
      if (data.categories) combined.categories?.push(...data.categories)
      if (data.portfolio) combined.portfolio?.push(...data.portfolio)
    } catch (error) {
      console.error(`Error parsing ${filePath}:`, error)
    }
  }

  // Deduplicate categories by slug
  if (combined.categories) {
    const seen = new Set<string>()
    combined.categories = combined.categories.filter((cat) => {
      if (seen.has(cat.slug)) return false
      seen.add(cat.slug)
      return true
    })
  }

  return combined
}