justvitamin/scripts/promote-dataforseo-asin-candidates.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 candidatePath = join(dataforseoRoot, 'asin-candidates.csv')
const confirmationReportPath = join(outRoot, 'dataforseo_asin_confirmation_report.json')
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 normalizeAmazonDpUrl(asin: string, productUrl: string) {
  const url = String(productUrl || '').trim()
  if (/^https:\/\/www\.amazon\.co\.uk\/dp\/[A-Z0-9]{10}/.test(url)) return `https://www.amazon.co.uk/dp/${asin}`
  if (/^https:\/\/amazon\.co\.uk\/dp\/[A-Z0-9]{10}/.test(url)) return `https://www.amazon.co.uk/dp/${asin}`
  return url
}

const generatedAt = new Date().toISOString()
const asinRows = existsSync(asinMapPath) ? parseCsv(readFileSync(asinMapPath, 'utf8')) : []
const candidateRows = existsSync(candidatePath) ? parseCsv(readFileSync(candidatePath, 'utf8')) : []
const confirmationReport = existsSync(confirmationReportPath) ? JSON.parse(readFileSync(confirmationReportPath, 'utf8').replace(/^\uFEFF/, '')) : { confirmations: [] }
const candidateBySku = new Map(candidateRows.map((row: any) => [row.SKU, row]))
const confirmationBySku = new Map((confirmationReport.confirmations || []).map((row: any) => [row.SKU, row]))
const asinMapHeaders = ['SKU', 'Product', 'Priority', 'Provider', 'AmazonDomain', 'ASIN', 'ProductUrl', 'Depth', 'LocationName', 'LanguageName', 'Status']

const promotionRows = asinRows.map((row: any) => {
  const candidate: any = candidateBySku.get(row.SKU) || {}
  const candidateAsin = String(candidate.CandidateASIN || '').trim().toUpperCase()
  const candidateUrl = normalizeAmazonDpUrl(candidateAsin, candidate.CandidateProductUrl || '')
  const asinLooksValid = /^[A-Z0-9]{10}$/.test(candidateAsin)
  const amazonUrlValid = candidateUrl === `https://www.amazon.co.uk/dp/${candidateAsin}`
  const alreadyFilled = Boolean(row.ASIN)
  const confirmed = candidate.MatchStatus === 'amazon_confirmed_exact_match'
  const highConfidence = candidate.Confidence === 'high'
  const confirmation: any = confirmationBySku.get(row.SKU) || {}
  const confirmationReady = confirmation.ReadyForPromotion === 'yes'
  const promotable = !alreadyFilled && asinLooksValid && amazonUrlValid && confirmed && highConfidence && confirmationReady
  const blockers = []
  if (!candidate.CandidateASIN) blockers.push('no candidate')
  if (alreadyFilled) blockers.push('asin-map already has ASIN')
  if (candidate.CandidateASIN && !asinLooksValid) blockers.push('candidate ASIN shape invalid')
  if (candidate.CandidateASIN && !amazonUrlValid) blockers.push('candidate ProductUrl is not canonical amazon.co.uk/dp/ASIN')
  if (candidate.CandidateASIN && !confirmed) blockers.push('MatchStatus is not amazon_confirmed_exact_match')
  if (candidate.CandidateASIN && !highConfidence) blockers.push(`Confidence is ${candidate.Confidence || 'blank'}`)
  if (candidate.CandidateASIN && !confirmationReady) blockers.push(`confirmation gate not ready${confirmation.Blockers ? `: ${confirmation.Blockers}` : ''}`)
  return {
    SKU: row.SKU,
    Product: row.Product,
    CurrentASIN: row.ASIN || '',
    CandidateASIN: candidateAsin,
    CandidateProductUrl: candidateUrl,
    MatchStatus: candidate.MatchStatus || '',
    Confidence: candidate.Confidence || '',
    ConfirmationReady: confirmationReady ? 'yes' : 'no',
    Promote: promotable ? 'yes' : 'no',
    Blockers: blockers.join(' | ')
  }
})

const promoted = promotionRows.filter(row => row.Promote === 'yes')
const promotedBySku = new Map(promoted.map(row => [row.SKU, row]))
let updatedAsinRows = asinRows
if (promoted.length) {
  updatedAsinRows = asinRows.map((row: any) => {
    const promotion: any = promotedBySku.get(row.SKU)
    if (!promotion) return row
    return {
      ...row,
      ASIN: promotion.CandidateASIN,
      ProductUrl: promotion.CandidateProductUrl,
      Status: 'ready_for_task_payload'
    }
  })
  writeFileSync(asinMapPath, csv(updatedAsinRows, asinMapHeaders), 'utf8')
}

const manifest = {
  generatedAt,
  ready: promoted.length > 0,
  applied: promoted.length > 0,
  counts: {
    asinMapRows: asinRows.length,
    candidateRows: candidateRows.length,
    promotable: promoted.length,
    blocked: promotionRows.filter(row => row.Promote !== 'yes').length,
    asinMapReadyAfterPromotion: updatedAsinRows.filter((row: any) => row.ASIN).length,
    asinMapMissingAfterPromotion: updatedAsinRows.filter((row: any) => !row.ASIN).length
  },
  files: {
    asinMap: 'data/sources/dataforseo/asin-map.csv',
    candidates: 'data/sources/dataforseo/asin-candidates.csv',
    confirmations: 'data/sources/dataforseo/asin-confirmations.csv',
    confirmationReport: 'content_population_exports/dataforseo_asin_confirmation_report.json',
    promotionReportMd: 'content_population_exports/dataforseo_asin_promotion_report.md',
    promotionReportCsv: 'content_population_exports/dataforseo_asin_promotion_report.csv',
    promotionReportJson: 'content_population_exports/dataforseo_asin_promotion_report.json'
  },
  guardrails: [
    'Only candidates with MatchStatus=amazon_confirmed_exact_match and Confidence=high are promoted.',
    'The separate ASIN confirmation report must mark the row ReadyForPromotion=yes.',
    'Candidate ProductUrl must be canonical https://www.amazon.co.uk/dp/ASIN.',
    'Existing ASIN map rows are never overwritten by this script.',
    'Promotion only prepares a payload; source:spend-gate still blocks paid provider use until budget and approval gates pass.'
  ],
  promotions: promotionRows
}

writeFileSync(join(outRoot, 'dataforseo_asin_promotion_report.json'), JSON.stringify(manifest, null, 2) + '\n', 'utf8')
writeFileSync(join(outRoot, 'dataforseo_asin_promotion_report.csv'), csv(promotionRows, ['SKU', 'Product', 'CurrentASIN', 'CandidateASIN', 'CandidateProductUrl', 'MatchStatus', 'Confidence', 'ConfirmationReady', 'Promote', 'Blockers']), 'utf8')
writeFileSync(join(outRoot, 'dataforseo_asin_promotion_report.md'), [
  '# DataForSEO ASIN promotion report',
  '',
  `Generated: ${generatedAt}`,
  '',
  'This no-spend step safely copies only Amazon-confirmed, high-confidence candidate ASINs into `data/sources/dataforseo/asin-map.csv`.',
  '',
  `- Candidate rows: ${manifest.counts.candidateRows}`,
  `- Promoted rows: ${manifest.counts.promotable}`,
  `- ASIN map ready after promotion: ${manifest.counts.asinMapReadyAfterPromotion}`,
  `- ASIN map missing after promotion: ${manifest.counts.asinMapMissingAfterPromotion}`,
  '',
  '## Guardrails',
  ...manifest.guardrails.map(item => `- ${item}`),
  '',
  '## Promotion rows',
  ...promotionRows.map(row => `- [${row.Promote === 'yes' ? 'promoted' : 'blocked'}] ${row.SKU} — ${row.CandidateASIN || 'no candidate'} — ${row.Blockers || 'all checks passed'}`),
  ''
].join('\n'), 'utf8')

console.log(`DataForSEO ASIN promotion ready: ${manifest.ready}`)
console.log(`Promoted rows: ${manifest.counts.promotable}`)
console.log(`ASINs ready after promotion: ${manifest.counts.asinMapReadyAfterPromotion}`)
console.log(`ASINs missing after promotion: ${manifest.counts.asinMapMissingAfterPromotion}`)