amazon-po-parser/scrape-payee.mjs

import { chromium } from "playwright";
import { readFileSync, writeFileSync, existsSync } from "fs";

const PO_LIST = "output/unknown-pos.json";
const OUTPUT_FILE = "output/payee-results.json";
const PROGRESS_FILE = "output/payee-progress.json";
const DELAY_MS = 800;

const allPOs = JSON.parse(readFileSync(PO_LIST, "utf-8"));
console.log(`Loaded ${allPOs.length} POs to scrape from Payee Central`);

let completed = {};
if (existsSync(PROGRESS_FILE)) {
  completed = JSON.parse(readFileSync(PROGRESS_FILE, "utf-8"));
  console.log(`Resuming: ${Object.keys(completed).length} already scraped`);
}

const remaining = allPOs.filter((po) => !completed[po]);
console.log(`Remaining: ${remaining.length} POs\n`);

if (remaining.length === 0) {
  console.log("All POs already scraped!");
  process.exit(0);
}

function poUrl(po) {
  return `https://payeecentral.amazon.com/PurchaseOrder/Details?poNumber=${po}`;
}

const userDataDir = new URL("./chrome-profile", import.meta.url).pathname;
const context = await chromium.launchPersistentContext(userDataDir, {
  headless: false,
  channel: "chrome",
  args: ["--disable-blink-features=AutomationControlled"],
  ignoreDefaultArgs: ["--enable-automation"],
});

const page = context.pages()[0] || (await context.newPage());

console.log("Opening Payee Central...");
await page.goto(poUrl(remaining[0]), { waitUntil: "networkidle", timeout: 30000 });

if (page.url().includes("signin") || page.url().includes("/ap/")) {
  console.log("\n🔐 Login required. Please sign in in the browser window...");
  await page.waitForURL(
    (u) => {
      const s = u.toString();
      return s.includes("payeecentral.amazon.com") && !s.includes("signin") && !s.includes("/ap/");
    },
    { timeout: 180000 }
  );
  console.log("✅ Logged in!\n");
}

let results = [];
if (existsSync(OUTPUT_FILE)) {
  results = JSON.parse(readFileSync(OUTPUT_FILE, "utf-8"));
}

let scraped = 0;
let errors = 0;
let consecutiveErrors = 0;
const startTime = Date.now();

for (const po of remaining) {
  try {
    await page.goto(poUrl(po), { waitUntil: "networkidle", timeout: 20000 });

    if (page.url().includes("signin") || page.url().includes("/ap/")) {
      console.log("\n🔐 Session expired. Please log in again...");
      await page.waitForURL(
        (u) => !u.toString().includes("signin") && !u.toString().includes("/ap/"),
        { timeout: 180000 }
      );
      await page.goto(poUrl(po), { waitUntil: "networkidle", timeout: 20000 });
    }

    await page.waitForTimeout(500);
    const bodyText = await page.innerText("body");

    // Extract Ship To with site code
    const shipToMatch = bodyText.match(/Ship To\s+([\s\S]*?)(?=Payee Central Account)/);
    const shipTo = shipToMatch ? shipToMatch[1].trim() : null;

    let siteCode = null;
    if (shipTo) {
      const m = shipTo.match(/\(([A-Z0-9]{3,5})\)/);
      if (m) siteCode = m[1];
      if (!siteCode) {
        const m2 = shipTo.match(/(?:LLC|Inc)\s*-\s*([A-Z0-9]{3,5})\b/);
        if (m2) siteCode = m2[1];
      }
    }

    // Extract line items from table
    const lineItems = await page.$$eval(
      "table tbody tr",
      (rows) =>
        rows.map((row) => {
          const cells = row.querySelectorAll("td");
          if (cells.length < 7) return null;
          return {
            description: cells[1]?.textContent?.trim() || null,
            need_by: cells[2]?.textContent?.trim() || null,
            quantity: cells[3]?.textContent?.trim() || null,
            unit: cells[4]?.textContent?.trim() || null,
            price: cells[5]?.textContent?.trim() || null,
            amount: cells[6]?.textContent?.trim() || null,
          };
        }).filter(Boolean)
    ).catch(() => []);

    const result = {
      po_number: po,
      site_code: siteCode,
      ship_to_raw: shipTo,
      line_items: lineItems,
    };

    results.push(result);
    completed[po] = true;
    scraped++;
    consecutiveErrors = 0;

    const elapsed = (Date.now() - startTime) / 1000;
    const rate = scraped / elapsed;
    const eta = Math.round((remaining.length - scraped) / rate);
    const etaMin = Math.floor(eta / 60);
    const etaSec = eta % 60;

    process.stdout.write(
      `\r[${scraped}/${remaining.length}] ${po} | ${siteCode || "?"} | ETA: ${etaMin}m${etaSec}s    `
    );

    if (scraped % 10 === 0) {
      writeFileSync(OUTPUT_FILE, JSON.stringify(results, null, 2));
      writeFileSync(PROGRESS_FILE, JSON.stringify(completed));
    }
  } catch (err) {
    console.error(`\n  [${po}] Error: ${err.message}`);
    errors++;
    consecutiveErrors++;
    if (consecutiveErrors > 10) {
      console.error("Too many consecutive errors, stopping.");
      break;
    }
  }

  await page.waitForTimeout(DELAY_MS);
}

writeFileSync(OUTPUT_FILE, JSON.stringify(results, null, 2));
writeFileSync(PROGRESS_FILE, JSON.stringify(completed));

console.log(`\n\nDone! Scraped: ${scraped}, Errors: ${errors}`);
console.log(`Total in output: ${results.length} POs`);
console.log(`Results saved to ${OUTPUT_FILE}`);

await context.close();