feat: implement automated data scraping and history generation pipeline for PVL reservoir levels

scratch_get_storage.ts

@@ -0,0 +1,28 @@
+import axios from 'axios';
+import * as cheerio from 'cheerio';
+import https from 'https';
+import { lakesConfig } from './scripts/lakesConfig';
+
+async function run() {
+  const agent = new https.Agent({ rejectUnauthorized: false });
+  for (const lake of lakesConfig) {
+    const [internalId, oid] = lake.id.split('|');
+    const URL = `https://www.pvl.cz/portal/nadrze/cz/pc/Mereni.aspx?oid=${oid}&id=${internalId}`;
+    try {
+      const res = await axios.get(URL, { httpsAgent: agent, headers: { 'User-Agent': 'Mozilla/5.0' } });
+      const $ = cheerio.load(res.data);
+      let storage = 0;
+      $('table').each((i, tbl) => {
+        const text = $(tbl).text();
+        const match = text.match(/Hladina z[aá]sobn[ií]ho prostoru:\s*([\d,]+)/i);
+        if (match) {
+          storage = parseFloat(match[1].replace(',', '.'));
+        }
+      });
+      console.log(`{ id: "${lake.id}", storageLevel: ${storage} },`);
+    } catch (e) {
+      console.error(e.message);
+    }
+  }
+}
+run();