Files
justvitamin/scripts/validate-dataforseo-task-preflight.ts
T
Omair Saleh 056c47581f feat: editorial review dashboard + elite-grade pilot batch (5 SKUs)
Ships the second dashboard surface — a Pattern Library + Preview Theatre — that
presents the 4-section PDP pilot batch back to Umar, compliance, and the board
in an editorial format. Adds the full data layer that drives it: 5 source-backed
per-SKU drafts at QA 100/100, 15 competitor PDP semantic extracts, PubMed
evidence packs, EFSA claims library extension, JV brand voice guide, hand-curated
product FAQs, and the Matrixify-ready CSV exports for Lewis.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-06-02 18:50:09 +08:00

176 lines
8.7 KiB
TypeScript

#!/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 defaultEnvPath = 'C:\\Users\\uldvs\\OneDrive\\Desktop\\work\\pi-fresh-foundation\\.env'
mkdirSync(outRoot, { recursive: true })
function loadEnvFile(path: string) {
if (!existsSync(path)) return false
for (const line of readFileSync(path, 'utf8').split(/\r?\n/)) {
const match = line.match(/^\s*(DATAFORSEO_[A-Za-z0-9_]+)\s*=\s*(.*)\s*$/)
if (match && !process.env[match[1]]) process.env[match[1]] = match[2].trim().replace(/^['"]|['"]$/g, '')
}
return true
}
function readJson(path: string, fallback: any = null) {
if (!existsSync(path)) return fallback
return JSON.parse(readFileSync(path, 'utf8').replace(/^\uFEFF/, ''))
}
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'
}
const generatedAt = new Date().toISOString()
const credentialEnvFile = process.env.DATAFORSEO_ENV_FILE || defaultEnvPath
const credentialEnvLoaded = loadEnvFile(credentialEnvFile)
const payloadPath = join(outRoot, 'dataforseo_amazon_reviews_task_payload.json')
const pilot = readJson(join(outRoot, 'dataforseo_pilot_manifest.json'), {})
const spendGate = readJson(join(outRoot, 'source_acquisition_spend_gate.json'), {})
const supportResponseGate = readJson(join(outRoot, 'dataforseo_support_response_gate.json'), {})
const reviewsDocsStatus = readJson(join(outRoot, 'dataforseo_reviews_docs_status.json'), {})
const budgetGate = readJson(join(outRoot, 'source_acquisition_budget_gate.json'), {})
const budget = budgetGate.budget || {}
const taskPayload = readJson(payloadPath, [])
const allowFlag = process.env.JV_SOURCE_ALLOW_PAID || ''
const loginPresent = Boolean(process.env.DATAFORSEO_LOGIN)
const passwordPresent = Boolean(process.env.DATAFORSEO_PASSWORD)
const explicitAllow = allowFlag === 'dataforseo-stage1'
const maxTasks = Number(budget.dataforseoStage1?.maxTasks || 0)
const maxDepth = Number(budget.dataforseoStage1?.maxReviewsPerAsin || 0)
const payloadRows = Array.isArray(taskPayload) ? taskPayload : []
const payloadWithinTaskCap = payloadRows.length > 0 && maxTasks > 0 && payloadRows.length <= maxTasks
const payloadWithinDepthCap = payloadRows.length > 0 && maxDepth > 0 && payloadRows.every((row: any) => Number(row.depth || 0) <= maxDepth)
const uniqueTags = new Set(payloadRows.map((row: any) => row.tag).filter(Boolean))
const checks = [
{
Check: 'Explicit paid-run flag set',
Status: explicitAllow ? 'pass' : 'blocked',
Blocking: 'yes',
Evidence: explicitAllow ? 'JV_SOURCE_ALLOW_PAID=dataforseo-stage1' : 'set JV_SOURCE_ALLOW_PAID=dataforseo-stage1 only when intentionally running paid Stage 1'
},
{
Check: 'DataForSEO credentials present in environment',
Status: loginPresent && passwordPresent ? 'pass' : 'blocked',
Blocking: 'yes',
Evidence: `DATAFORSEO_LOGIN=${loginPresent ? 'present' : 'missing'}, DATAFORSEO_PASSWORD=${passwordPresent ? 'present' : 'missing'}`
},
{
Check: 'DataForSEO Reviews endpoint cleared by support',
Status: supportResponseGate.ready === true ? 'pass' : 'blocked',
Blocking: 'yes',
Evidence: supportResponseGate.ready === true ? 'support response gate ready' : `${supportResponseGate.counts?.passing || 0}/${supportResponseGate.counts?.checks || 0} support-response checks passing; do not retry Reviews endpoint yet`
},
{
Check: 'Official DataForSEO Reviews docs no longer mark endpoint unavailable',
Status: reviewsDocsStatus.ready === true ? 'pass' : 'blocked',
Blocking: 'yes',
Evidence: reviewsDocsStatus.ready === true
? 'official docs status gate ready'
: `endpointTemporarilyUnavailable=${reviewsDocsStatus.endpointTemporarilyUnavailable === true}; ${reviewsDocsStatus.counts?.passing || 0}/${reviewsDocsStatus.counts?.checks || 0} docs-status checks passing`
},
{
Check: 'Spend gate allows DataForSEO Stage 1',
Status: spendGate.spendAllowed?.dataforseoStage1 === true ? 'pass' : 'blocked',
Blocking: 'yes',
Evidence: `${spendGate.counts?.passing || 0}/${spendGate.counts?.checks || 0} spend checks passing`
},
{
Check: 'Budget gate allows DataForSEO Stage 1',
Status: budgetGate.budgetAllowed?.dataforseoStage1 === true ? 'pass' : 'blocked',
Blocking: 'yes',
Evidence: `${budgetGate.counts?.passing || 0}/${budgetGate.counts?.checks || 0} budget checks passing`
},
{
Check: 'Task payload has rows',
Status: payloadRows.length > 0 ? 'pass' : 'blocked',
Blocking: 'yes',
Evidence: `${payloadRows.length} task rows`
},
{
Check: 'Payload stays within task cap',
Status: payloadWithinTaskCap ? 'pass' : 'blocked',
Blocking: 'yes',
Evidence: `${payloadRows.length} task rows / maxTasks=${maxTasks || 'unset'}`
},
{
Check: 'Payload stays within depth cap',
Status: payloadWithinDepthCap ? 'pass' : 'blocked',
Blocking: 'yes',
Evidence: `${payloadRows.length ? Math.max(...payloadRows.map((row: any) => Number(row.depth || 0))) : 0} max payload depth / maxReviewsPerAsin=${maxDepth || 'unset'}`
},
{
Check: 'Payload tags are unique and traceable',
Status: payloadRows.length > 0 && uniqueTags.size === payloadRows.length && payloadRows.every((row: any) => String(row.tag || '').startsWith('jv:')) ? 'pass' : 'blocked',
Blocking: 'yes',
Evidence: `${uniqueTags.size}/${payloadRows.length} unique tags`
}
]
const ready = checks.every(row => row.Status === 'pass')
const manifest = {
generatedAt,
ready,
endpoint: pilot.endpoint || 'POST https://api.dataforseo.com/v3/merchant/amazon/reviews/task_post',
command: 'bun run dataforseo:task-preflight',
paidRunCommand: `Credentials load from shell variables or ${credentialEnvFile}; set only JV_SOURCE_ALLOW_PAID=dataforseo-stage1 for the paid-run window, then run the guarded reviews-live command after this preflight is ready.`,
counts: {
checks: checks.length,
passing: checks.filter(row => row.Status === 'pass').length,
blocked: checks.filter(row => row.Status !== 'pass').length,
taskRows: payloadRows.length
},
files: {
payload: 'content_population_exports/dataforseo_amazon_reviews_task_payload.json',
curlTemplate: 'content_population_exports/dataforseo_amazon_reviews_curl_template.sh',
preflightMd: 'content_population_exports/dataforseo_task_preflight.md',
preflightCsv: 'content_population_exports/dataforseo_task_preflight.csv',
preflightJson: 'content_population_exports/dataforseo_task_preflight.json',
credentialEnvFile
},
credentialEnvLoaded,
checks,
blockers: checks.filter(row => row.Status !== 'pass').map(row => `${row.Check}: ${row.Evidence}`),
payloadPreview: payloadRows.slice(0, 12).map((row: any) => ({ asin: row.asin, depth: row.depth, tag: row.tag }))
}
writeFileSync(join(outRoot, 'dataforseo_task_preflight.json'), JSON.stringify(manifest, null, 2) + '\n', 'utf8')
writeFileSync(join(outRoot, 'dataforseo_task_preflight.csv'), csv(checks, ['Check', 'Status', 'Blocking', 'Evidence']), 'utf8')
writeFileSync(join(outRoot, 'dataforseo_task_preflight.md'), [
'# DataForSEO paid task preflight',
'',
`Generated: ${generatedAt}`,
'',
`Ready to run paid Stage 1 task: ${ready ? 'YES' : 'NO'}`,
'',
'This preflight is a final local guard before any DataForSEO paid task creation. It does not call DataForSEO.',
'',
`Endpoint: ${manifest.endpoint}`,
'',
'## Checks',
...checks.map(row => `- ${row.Status === 'pass' ? '[pass]' : '[blocked]'} ${row.Check}${row.Evidence}`),
'',
'## Paid-run rule',
'- All checks must pass.',
`- Credentials are loaded from shell environment variables or \`${credentialEnvFile}\`; secrets are not written to the report.`,
'- `JV_SOURCE_ALLOW_PAID` must equal `dataforseo-stage1` for the paid run window.',
'- The official DataForSEO docs status gate must no longer show the Amazon Reviews endpoint as temporarily unavailable.',
'- The DataForSEO support response gate must confirm endpoint availability or a replacement endpoint before retrying after the archived 50304.',
'- Run only the generated Stage 1 payload, then stop and inspect raw task_get output/cost before any scale-up.',
''
].join('\n'), 'utf8')
console.log(`DataForSEO task preflight ready: ${ready}`)
console.log(`Checks passing: ${manifest.counts.passing}/${manifest.counts.checks}`)
console.log(`Task rows: ${manifest.counts.taskRows}`)