From 470629678e4184b04f546077bba48957fe9a6f83 Mon Sep 17 00:00:00 2001
From: Nulo <git@nulo.in>
Date: Thu, 21 Dec 2023 23:59:06 -0300
Subject: [PATCH] wip: dia-link-scraper

---
 dia-link-scraper/index.ts     | 51 +++++++++++++++++++++++++++++++++++
 dia-link-scraper/package.json | 18 +++++++++++++
 2 files changed, 69 insertions(+)
 create mode 100644 dia-link-scraper/index.ts
 create mode 100644 dia-link-scraper/package.json

diff --git a/dia-link-scraper/index.ts b/dia-link-scraper/index.ts
new file mode 100644
index 0000000..7eb211f
--- /dev/null
+++ b/dia-link-scraper/index.ts
@@ -0,0 +1,51 @@
+import puppeteer from "puppeteer";
+import { blockImages } from "../puppeteer-utils";
+
+(async () => {
+  const browser = await puppeteer.launch({
+    headless: false,
+    defaultViewport: { height: 4000, width: 1920 },
+  });
+  const page = await browser.newPage();
+
+  await blockImages(page);
+
+  await page.goto("https://diaonline.supermercadosdia.com.ar/almacen/");
+
+  async function getHrefs() {
+    return await page.evaluate(() =>
+      Array.from(
+        document.querySelectorAll<HTMLAnchorElement>(
+          "a.vtex-product-summary-2-x-clearLink"
+        ),
+        (a) => new URL(a.href).toString()
+      )
+    );
+  }
+  const seeMoreSel = "button ::-p-text(Ver más Productos)";
+  await page.waitForSelector(seeMoreSel);
+
+  try {
+    let prev = { n: 0, d: Date.now() };
+    while (true) {
+      const hrefs = await getHrefs();
+      console.debug(prev);
+      if (prev.n === hrefs.length && Date.now() > prev.d + 15000) break;
+      prev = { n: hrefs.length, d: Date.now() };
+
+      const seeMoreEl = await page.$eval(seeMoreSel, (el) => {
+        el.parentElement?.click();
+        el.scrollIntoView({ block: "center", behavior: "smooth" });
+      });
+      await wait(150);
+    }
+    const hrefs = await getHrefs();
+    hrefs.forEach((l) => console.log(l));
+  } finally {
+    await browser.close();
+  }
+})();
+
+function wait(ms: number) {
+  return new Promise((resolve) => setTimeout(() => resolve(void 0), ms));
+}
diff --git a/dia-link-scraper/package.json b/dia-link-scraper/package.json
new file mode 100644
index 0000000..4f9c7d6
--- /dev/null
+++ b/dia-link-scraper/package.json
@@ -0,0 +1,18 @@
+{
+  "name": "dia-link-scraper",
+  "type": "module",
+  "version": "1.0.0",
+  "description": "",
+  "main": "index.js",
+  "scripts": {
+    "test": "echo \"Error: no test specified\" && exit 1"
+  },
+  "keywords": [],
+  "author": "",
+  "license": "ISC",
+  "dependencies": {
+    "puppeteer": "^21.6.1",
+    "puppeteer-utils": "workspace:^",
+    "tsx": "^4.7.0"
+  }
+}