justvitamin/scripts/run-apify-healthspan-retailer-capture.ts

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

const root = process.cwd()
const rawRoot = join(root, 'data', 'sources', 'apify', 'raw')
const token = Bun.env.APIFY_TOKEN || Bun.env.APIFY_API_TOKEN
const generatedAt = new Date().toISOString()
if (!token) throw new Error('Missing APIFY_TOKEN or APIFY_API_TOKEN')

const maxTotalChargeUsd = Number(Bun.env.APIFY_HEALTHSPAN_RETAILER_MAX_USD || 0.04)

function writeJson(path: string, data: any) {
  mkdirSync(dirname(path), { recursive: true })
  writeFileSync(path, JSON.stringify(data, null, 2) + '\n', 'utf8')
}
async function apify(path: string, options: RequestInit = {}) {
  const sep = path.includes('?') ? '&' : '?'
  const res = await fetch(`https://api.apify.com${path}${sep}token=${encodeURIComponent(token!)}`, {
    ...options,
    headers: { 'Content-Type': 'application/json', ...(options.headers || {}) }
  })
  const text = await res.text()
  let data: any
  try { data = text ? JSON.parse(text) : null } catch { data = text }
  if (!res.ok) throw new Error(`Apify ${res.status} ${res.statusText}: ${typeof data === 'string' ? data : JSON.stringify(data)}`)
  return data
}
async function waitForRun(runId: string, timeoutMs = 1000 * 60 * 6) {
  const started = Date.now()
  let run = (await apify(`/v2/actor-runs/${runId}`)).data
  while (!['SUCCEEDED', 'FAILED', 'ABORTED', 'TIMED-OUT'].includes(run.status)) {
    if (Date.now() - started > timeoutMs) throw new Error(`Timed out waiting for ${runId}`)
    await new Promise(resolve => setTimeout(resolve, 5000))
    run = (await apify(`/v2/actor-runs/${runId}`)).data
    console.log(`${runId}: ${run.status} usd=${run.usageTotalUsd || 0}`)
  }
  return run
}

const pageFunction = String.raw`
async function pageFunction(context) {
  const { request, $ } = context;
  const clean = (value) => String(value || '').replace(/\s+/g, ' ').trim();
  const text = clean($('body').text()).slice(0, 50000);
  const headings = $('h1,h2,h3').toArray().map(el => clean($(el).text())).filter(Boolean).slice(0, 80);
  const prices = $('[class*=price], [data-testid*=price], [aria-label*=price]').toArray().map(el => clean($(el).text())).filter(Boolean).slice(0, 50);
  const links = $('a[href]').toArray().map(el => ({ text: clean($(el).text()).slice(0, 160), href: $(el).attr('href') })).filter(row => row.text || row.href).slice(0, 80);
  return {
    capturedAt: new Date().toISOString(),
    sku: 'JV-VISISOFT',
    sourceType: 'competitor_retailer_pdp',
    label: 'Healthspan OptiVision retailer PDP fallback',
    competitor: 'Healthspan',
    retailer: 'Asda Pharmacy',
    requestedUrl: request.url,
    loadedUrl: request.loadedUrl || request.url,
    title: clean($('title').text()),
    h1: clean($('h1').first().text()),
    headings,
    prices,
    links,
    text,
    textChars: text.length,
    htmlChars: $.html().length,
    matchContext: 'competitor_retailer_pdp_fallback',
    note: 'Retailer PDP fallback for Healthspan OptiVision because the Healthspan direct PDP returned 0 items in Apify browser/content crawls.'
  };
}`

const input = {
  startUrls: [{ url: 'https://pharmacy.asda.com/medicines/healthspan-optivision' }],
  maxRequestsPerCrawl: 1,
  maxConcurrency: 1,
  maxRequestRetries: 0,
  requestTimeoutSecs: 45,
  pageFunction,
  proxyConfiguration: { useApifyProxy: true }
}

console.log(`Starting apify/cheerio-scraper for Healthspan OptiVision retailer fallback, maxRequests=1, cap=$${maxTotalChargeUsd}`)
const start = await apify(`/v2/acts/apify~cheerio-scraper/runs?maxTotalChargeUsd=${maxTotalChargeUsd}`, {
  method: 'POST',
  body: JSON.stringify(input)
})
const run = await waitForRun(start.data.id)
const items = run.defaultDatasetId ? await apify(`/v2/datasets/${run.defaultDatasetId}/items?clean=true`) : []
const result = {
  capturedAt: generatedAt,
  job: 'healthspan-retailer',
  actor: 'apify/cheerio-scraper',
  maxTotalChargeUsd,
  input,
  run: {
    id: run.id,
    status: run.status,
    statusMessage: run.statusMessage,
    defaultDatasetId: run.defaultDatasetId,
    startedAt: run.startedAt,
    finishedAt: run.finishedAt,
    usageTotalUsd: run.usageTotalUsd,
    usage: run.usage,
    stats: run.stats
  },
  itemCount: Array.isArray(items) ? items.length : 0,
  items
}

const stamp = generatedAt.replace(/[:.]/g, '-')
writeJson(join(rawRoot, `top3-specialist-healthspan-retailer-${stamp}.json`), result)
writeJson(join(rawRoot, 'top3-specialist-healthspan-retailer-latest.json'), result)
console.log(`Healthspan retailer fallback ${run.status}: ${result.itemCount} items, usd=${run.usageTotalUsd || 0}`)