Run the asset pre-upload pass

// scripts/pre-upload-assets.js
import { client } from '../lib/sanity-client.js'
import { glob } from 'fs/promises'
import fs from 'fs'
import path from 'path'

// All known Webflow CDN domains — check your HTML export if you're missing assets
const CDN_PATTERNS = [
  /https:\/\/uploads-ssl\.webflow\.com\/[^\s"'<>\\]+/g,
  /https:\/\/assets-global\.website-files\.com\/[^\s"'<>\\]+/g,
  /https:\/\/cdn\.prod\.website-files\.com\/[^\s"'<>\\]+/g,
]

// Local export image paths: src="images/foo.png" → upload from webflow/export/images/foo.png
const LOCAL_IMG_PATTERN = /(?:src|href)=["'](images\/[^"']+)["']/g

const WEBFLOW_CDN_PATTERN = CDN_PATTERNS[0] // kept for backwards compat in rich-text.js

function cleanUrl(url) {
  // Strip any trailing punctuation that leaked in from HTML attribute context
  return url.replace(/[\\,;]+$/, '')
}

async function run() {
  const seen = new Set()
  const assetMap = {}

  // Scan all NDJSON item files for CDN URLs in fieldData
  const ndjsonFiles = await Array.fromAsync(glob('webflow/api/items/*.ndjson'))
  for (const filePath of ndjsonFiles) {
    const content = fs.readFileSync(filePath, 'utf8')
    for (const pattern of CDN_PATTERNS) {
      for (const match of content.matchAll(new RegExp(pattern.source, 'g'))) seen.add(cleanUrl(match[0]))
    }
  }

  // Scan HTML export files for CDN URLs and local relative paths
  const htmlFiles = await Array.fromAsync(glob('webflow/export/**/*.html'))
  const localPaths = new Set() // relative paths like "images/foo.png"
  for (const filePath of htmlFiles) {
    const content = fs.readFileSync(filePath, 'utf8')
    for (const pattern of CDN_PATTERNS) {
      for (const match of content.matchAll(new RegExp(pattern.source, 'g'))) seen.add(cleanUrl(match[0]))
    }
    // Collect local export image paths
    for (const match of content.matchAll(LOCAL_IMG_PATTERN)) localPaths.add(match[1])
  }

  console.log(`Found ${seen.size} CDN URLs + ${localPaths.size} local paths. Uploading...`)

  // Upload CDN assets
  for (const url of seen) {
    try {
      const response = await fetch(url)
      if (!response.ok) throw new Error(`HTTP ${response.status} fetching ${url}`)

      const contentType = response.headers.get('content-type') ?? 'application/octet-stream'
      const filename = decodeURIComponent(url.split('/').pop()?.split('?')[0] ?? 'asset')
      // SVGs and other non-raster files must be uploaded as 'file', not 'image'
      const assetType = /^image\/(jpeg|png|gif|webp|avif)/.test(contentType) ? 'image' : 'file'

      const buffer = Buffer.from(await response.arrayBuffer())
      const asset = await client.assets.upload(assetType, buffer, {
        filename,
        contentType,
        source: { name: 'webflow-migration', id: url }
      })
      assetMap[url] = asset._id
      console.log(`✓ ${filename}`)
    } catch (err) {
      console.warn(`✗ ${url}: ${err.message}`)
    }
  }

  // Upload local export images (e.g. images/product-hero.png)
  for (const relPath of localPaths) {
    const diskPath = path.join('webflow/export', relPath)
    if (!fs.existsSync(diskPath)) { console.warn(`✗ Local file not found: ${diskPath}`); continue }
    try {
      const buffer = fs.readFileSync(diskPath)
      const filename = path.basename(relPath)
      const asset = await client.assets.upload('image', buffer, {
        filename,
        source: { name: 'webflow-migration', id: relPath }
      })
      // Map both the relative path and a root-relative URL form so rich-text.js can look it up
      assetMap[relPath] = asset._id
      assetMap[`/${relPath}`] = asset._id
      console.log(`✓ ${filename} (local)`)
    } catch (err) {
      console.warn(`✗ ${relPath}: ${err.message}`)
    }
  }

  fs.writeFileSync('asset-map.json', JSON.stringify(assetMap, null, 2))
  console.log(`Done. asset-map.json written with ${Object.keys(assetMap).length} entries.`)
  if (Object.keys(assetMap).length === 0) {
    console.warn('⚠ asset-map.json is empty — check CDN domains in CDN_PATTERNS against your HTML export.')
  }
}

run().catch(err => { console.error(err.message); process.exit(1) })

node --env-file=.env scripts/pre-upload-assets.js