justvitamin/scripts/generate-dataforseo-asin-lookup.ts

#!/usr/bin/env bun
import { existsSync, mkdirSync, readFileSync, writeFileSync } from 'fs'
import { join } from 'path'

const root = process.cwd()
const outRoot = join(root, 'content_population_exports')
const dataforseoRoot = join(root, 'data', 'sources', 'dataforseo')
const asinMapPath = join(dataforseoRoot, 'asin-map.csv')
const asinCandidatesPath = join(dataforseoRoot, 'asin-candidates.csv')

mkdirSync(outRoot, { recursive: true })
mkdirSync(dataforseoRoot, { recursive: true })

function csv(rows: any[], headers: string[]) {
  const esc = (value: any) => {
    if (value === undefined || value === null) value = ''
    if (Array.isArray(value)) value = value.join(' | ')
    value = String(value)
    return /[",\n\r]/.test(value) ? `"${value.replace(/"/g, '""')}"` : value
  }
  return [headers.join(','), ...rows.map(row => headers.map(header => esc(row[header])).join(','))].join('\n') + '\n'
}

function parseCsv(text: string) {
  const rows: string[][] = []
  let row: string[] = []
  let cell = ''
  let quoted = false
  for (let index = 0; index < text.length; index += 1) {
    const char = text[index]
    const next = text[index + 1]
    if (quoted) {
      if (char === '"' && next === '"') { cell += '"'; index += 1 }
      else if (char === '"') quoted = false
      else cell += char
    } else if (char === '"') quoted = true
    else if (char === ',') { row.push(cell); cell = '' }
    else if (char === '\n') { row.push(cell); rows.push(row); row = []; cell = '' }
    else if (char !== '\r') cell += char
  }
  if (cell || row.length) { row.push(cell); rows.push(row) }
  const headers = rows.shift()?.map(header => header.trim()) || []
  return rows.filter(values => values.some(value => value.trim())).map(values => Object.fromEntries(headers.map((header, index) => [header, (values[index] || '').trim()])))
}

function url(base: string, params: Record<string, string>) {
  const search = new URLSearchParams(params)
  return `${base}?${search.toString()}`
}

const generatedAt = new Date().toISOString()
const asinRows = existsSync(asinMapPath) ? parseCsv(readFileSync(asinMapPath, 'utf8')) : []
const candidateRows = existsSync(asinCandidatesPath) ? parseCsv(readFileSync(asinCandidatesPath, 'utf8')) : []
const candidateBySku = new Map(candidateRows.map((row: any) => [row.SKU, row]))
const queue = asinRows.map((row: any) => {
  const product = row.Product || row.SKU
  const query = `Just Vitamins ${product}`.replace(/\s+/g, ' ').trim()
  const asin = row.ASIN || ''
  const productUrl = row.ProductUrl || ''
  const candidate: any = candidateBySku.get(row.SKU) || {}
  const status = asin ? 'asin_present' : 'needs_manual_lookup'
  return {
    SKU: row.SKU,
    Product: product,
    Priority: row.Priority,
    Status: status,
    CurrentASIN: asin,
    CurrentProductUrl: productUrl,
    CandidateASIN: candidate.CandidateASIN || '',
    CandidateConfidence: candidate.Confidence || '',
    CandidateMatchStatus: candidate.MatchStatus || '',
    CandidateEvidenceUrl: candidate.EvidenceUrl || '',
    AmazonSearchUrl: url('https://www.amazon.co.uk/s', { k: query }),
    AmazonBrandSearchUrl: url('https://www.amazon.co.uk/s', { k: `Just Vitamins ${row.SKU}` }),
    WebSearchQuery: `${query} site:amazon.co.uk/dp OR site:amazon.co.uk/Just-Vitamins`,
    SuggestedAction: asin
      ? 'Verify ASIN/ProductUrl still match the target product before paid task creation.'
      : candidate.CandidateASIN
        ? 'Open CandidateEvidenceUrl and CandidateProductUrl, confirm exact Amazon product match, then paste ASIN/ProductUrl into asin-map.csv only if confidence is high.'
        : 'Open AmazonSearchUrl, confirm the exact Just Vitamins product, then paste ASIN and ProductUrl into data/sources/dataforseo/asin-map.csv.'
  }
})

const ready = queue.length > 0 && queue.every(row => row.CurrentASIN)
const manifest = {
  generatedAt,
  ready,
  counts: {
    targets: queue.length,
    asinPresent: queue.filter(row => row.CurrentASIN).length,
    missingAsins: queue.filter(row => !row.CurrentASIN).length
  },
  guardrails: [
    'Manual lookup only; this script does not call Amazon, DataForSEO, or Apify.',
    'Only accept an ASIN when the product title/brand clearly matches Just Vitamins and the target SKU/product.',
    'Do not create a DataForSEO paid task until every Stage 1 ASIN has been reviewed.',
    'Candidate ASINs from third-party pages are lookup aids only; they do not count as ready until confirmed on Amazon.'
  ],
  files: {
    asinMap: 'data/sources/dataforseo/asin-map.csv',
    asinCandidates: 'data/sources/dataforseo/asin-candidates.csv',
    asinCandidateReport: 'content_population_exports/dataforseo_asin_candidate_report.md',
    lookupQueueCsv: 'content_population_exports/dataforseo_asin_lookup_queue.csv',
    lookupQueueJson: 'content_population_exports/dataforseo_asin_lookup_queue.json'
  },
  queue
}

writeFileSync(join(outRoot, 'dataforseo_asin_lookup_queue.json'), JSON.stringify(manifest, null, 2) + '\n', 'utf8')
writeFileSync(join(outRoot, 'dataforseo_asin_lookup_queue.csv'), csv(queue, ['SKU', 'Product', 'Priority', 'Status', 'CurrentASIN', 'CurrentProductUrl', 'CandidateASIN', 'CandidateConfidence', 'CandidateMatchStatus', 'CandidateEvidenceUrl', 'AmazonSearchUrl', 'AmazonBrandSearchUrl', 'WebSearchQuery', 'SuggestedAction']), 'utf8')
writeFileSync(join(outRoot, 'dataforseo_asin_lookup_queue.md'), [
  '# DataForSEO ASIN lookup queue',
  '',
  `Generated: ${generatedAt}`,
  '',
  'This is a no-spend prep queue for the Stage 1 DataForSEO pilot. It does not call Amazon, DataForSEO, or Apify.',
  '',
  `- Targets: ${manifest.counts.targets}`,
  `- ASINs present: ${manifest.counts.asinPresent}`,
  `- Missing ASINs: ${manifest.counts.missingAsins}`,
  '',
  '## Workflow',
  '1. Open `dataforseo_asin_lookup_queue.csv`.',
  '2. For each missing row, open the Amazon search URL. If a CandidateASIN is present, use it only as a shortcut to find/verify the Amazon product page.',
  '3. Paste confirmed ASIN and product URL into `data/sources/dataforseo/asin-map.csv` only after exact Amazon product match is confirmed.',
  '4. Rerun `bun run dataforseo:pilot`; the task payload should then show ready rows.',
  '5. Review the payload before making any paid DataForSEO call.',
  '',
  '## Guardrails',
  ...manifest.guardrails.map(item => `- ${item}`),
  '',
  '## Targets',
  ...queue.map(row => `- ${row.CurrentASIN ? '[ready]' : '[missing]'} ${row.SKU} - ${row.Product} (${row.Priority})`),
  ''
].join('\n'), 'utf8')

console.log(`DataForSEO ASIN lookup ready: ${manifest.ready}`)
console.log(`Targets: ${manifest.counts.targets}`)
console.log(`ASINs present: ${manifest.counts.asinPresent}`)
console.log(`Missing ASINs: ${manifest.counts.missingAsins}`)