justvitamin/scripts/validate-dataforseo-asin-web-evidence.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 evidencePath = join(dataforseoRoot, 'asin-web-evidence.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()])))
}

if (!existsSync(evidencePath)) {
  writeFileSync(evidencePath, csv([], ['SKU', 'Product', 'EvidenceSource', 'EvidenceUrl', 'EvidenceTitle', 'EvidenceType', 'CandidateASIN', 'AmazonConfirmedUrl', 'BrandMatchesJustVitamins', 'ProductMatchesTarget', 'PackSizeMatchesTarget', 'MatchStatus', 'Confidence', 'Notes']), 'utf8')
}

const generatedAt = new Date().toISOString()
const evidenceRows = parseCsv(readFileSync(evidencePath, 'utf8'))
const rows = evidenceRows.map((row: any) => {
  const asinValid = /^[A-Z0-9]{10}$/.test(row.CandidateASIN || '')
  const amazonUrlValid = /^https:\/\/www\.amazon\.co\.uk\/dp\/[A-Z0-9]{10}/.test(row.AmazonConfirmedUrl || '')
  const amazonExact = row.MatchStatus === 'amazon_confirmed_exact_match' && row.Confidence === 'high' && row.BrandMatchesJustVitamins === 'yes' && row.ProductMatchesTarget === 'yes' && row.PackSizeMatchesTarget === 'yes' && asinValid && amazonUrlValid
  const usefulLookupAid = asinValid && row.MatchStatus === 'third_party_needs_amazon_confirmation'
  const blockers = []
  if (!asinValid && row.MatchStatus !== 'rejected_not_target') blockers.push('no valid ASIN')
  if (!amazonUrlValid && row.MatchStatus !== 'rejected_not_target') blockers.push('no canonical Amazon dp URL')
  if (row.MatchStatus !== 'amazon_confirmed_exact_match') blockers.push(row.MatchStatus === 'rejected_not_target' ? 'rejected as not target' : 'Amazon exact match not confirmed')
  if (row.Confidence !== 'high') blockers.push(`confidence=${row.Confidence || 'unknown'}`)
  if (row.BrandMatchesJustVitamins !== 'yes') blockers.push('brand not confirmed yes')
  if (row.ProductMatchesTarget !== 'yes') blockers.push('target product not confirmed yes')
  if (row.PackSizeMatchesTarget !== 'yes') blockers.push('pack size not confirmed yes')
  return {
    ...row,
    CandidateASINValid: asinValid ? 'yes' : 'no',
    AmazonUrlValid: amazonUrlValid ? 'yes' : 'no',
    UsableForPromotion: amazonExact ? 'yes' : 'no',
    LookupAid: usefulLookupAid ? 'yes' : 'no',
    Blockers: blockers.join(' | ')
  }
})

const manifest = {
  generatedAt,
  ready: rows.some((row: any) => row.UsableForPromotion === 'yes'),
  counts: {
    evidenceRows: rows.length,
    amazonConfirmedExact: rows.filter((row: any) => row.UsableForPromotion === 'yes').length,
    thirdPartyLookupAids: rows.filter((row: any) => row.LookupAid === 'yes').length,
    rejectedNotTarget: rows.filter((row: any) => row.MatchStatus === 'rejected_not_target').length,
    validAsinShape: rows.filter((row: any) => row.CandidateASINValid === 'yes').length
  },
  files: {
    evidenceCsv: 'data/sources/dataforseo/asin-web-evidence.csv',
    reportCsv: 'content_population_exports/dataforseo_asin_web_evidence.csv',
    reportJson: 'content_population_exports/dataforseo_asin_web_evidence.json',
    reportMd: 'content_population_exports/dataforseo_asin_web_evidence.md'
  },
  guardrails: [
    'Public web evidence is no-spend lookup support only.',
    'Third-party pages can suggest ASINs but cannot unlock paid DataForSEO payload rows.',
    'Only Amazon product-page exact matches with high confidence can be promoted into asin-map.csv.',
    'Rejected/not-target search results are recorded to avoid re-checking weak leads.'
  ],
  evidence: rows
}

writeFileSync(join(outRoot, 'dataforseo_asin_web_evidence.json'), JSON.stringify(manifest, null, 2) + '\n', 'utf8')
writeFileSync(join(outRoot, 'dataforseo_asin_web_evidence.csv'), csv(rows, ['SKU', 'Product', 'EvidenceSource', 'EvidenceUrl', 'EvidenceTitle', 'EvidenceType', 'CandidateASIN', 'AmazonConfirmedUrl', 'BrandMatchesJustVitamins', 'ProductMatchesTarget', 'PackSizeMatchesTarget', 'MatchStatus', 'Confidence', 'CandidateASINValid', 'AmazonUrlValid', 'LookupAid', 'UsableForPromotion', 'Blockers', 'Notes']), 'utf8')
writeFileSync(join(outRoot, 'dataforseo_asin_web_evidence.md'), [
  '# DataForSEO ASIN web evidence report',
  '',
  `Generated: ${generatedAt}`,
  '',
  'This report records no-spend public web evidence gathered before any provider subscription/API usage. It intentionally keeps third-party ASIN hints separate from Amazon-confirmed ASINs.',
  '',
  `- Evidence rows: ${manifest.counts.evidenceRows}`,
  `- Amazon-confirmed exact matches: ${manifest.counts.amazonConfirmedExact}`,
  `- Third-party lookup aids: ${manifest.counts.thirdPartyLookupAids}`,
  `- Rejected not-target leads: ${manifest.counts.rejectedNotTarget}`,
  '',
  '## Guardrails',
  ...manifest.guardrails.map(item => `- ${item}`),
  '',
  '## Evidence',
  ...rows.map((row: any) => `- [${row.UsableForPromotion === 'yes' ? 'promotable' : row.LookupAid === 'yes' ? 'lookup-aid' : 'blocked'}] ${row.SKU} — ${row.CandidateASIN || 'no ASIN'} — ${row.EvidenceSource}; blockers: ${row.Blockers || 'none'}`),
  ''
].join('\n'), 'utf8')

console.log(`DataForSEO ASIN web evidence ready: ${manifest.ready}`)
console.log(`Evidence rows: ${manifest.counts.evidenceRows}`)
console.log(`Amazon-confirmed exact matches: ${manifest.counts.amazonConfirmedExact}`)
console.log(`Third-party lookup aids: ${manifest.counts.thirdPartyLookupAids}`)