init

parser/scripts/debug-fetch.js

@@ -0,0 +1,71 @@
+const fs = require('fs/promises');
+const path = require('path');
+const cheerio = require('cheerio');
+const config = require('../src/config');
+const { launchBrowser, createContext } = require('../src/session');
+const { fetchTargetHtmlWithSession, buildForecastUrl } = require('../src/fetcher');
+const { parseItemsFromHtml, countForecastRows } = require('../src/parser');
+const { sanitizeFilePart } = require('../src/utils');
+
+async function main() {
+  const botArg = process.argv[2] || 'raketafon';
+  const activeTab = Number(process.argv[3] || 1);
+  const currentPage = Number(process.argv[4] || 1);
+  const botUrl = botArg.startsWith('http') ? botArg : `https://alpinbet.com/dispatch/antigol/${botArg}`;
+  const targetUrl = buildForecastUrl(botUrl, {
+    activeTab,
+    perPage: config.forecastPerPage,
+    page: currentPage
+  });
+
+  const browser = await launchBrowser({ headless: true });
+  const context = await createContext({ browser, sessionFile: config.sessionFile });
+
+  try {
+    const result = await fetchTargetHtmlWithSession({ context, config, targetUrl });
+    const items = parseItemsFromHtml(result.html, config.selectors);
+    const rows = countForecastRows(result.html, config.selectors);
+    const $ = cheerio.load(result.html);
+    const htmlFile = path.join(
+      config.htmlSnapshotDir,
+      `${sanitizeFilePart(botArg)}_tab-${activeTab}_page-${currentPage}_manual-debug.html`
+    );
+
+    await fs.mkdir(config.htmlSnapshotDir, { recursive: true });
+    await fs.writeFile(htmlFile, result.html, 'utf8');
+
+    console.log('status:', result.status);
+    console.log('title:', result.title);
+    console.log('url:', result.finalUrl);
+    console.log('rows:', rows);
+    console.log('parsedItems:', items.length);
+    console.log('htmlFile:', htmlFile);
+
+    const rowCandidates = $('.table-link')
+      .slice(0, 5)
+      .toArray()
+      .map((element, index) => {
+        const $row = $(element);
+        return {
+          index,
+          classes: $row.attr('class') || '',
+          text: $row.text().replace(/\s+/g, ' ').trim().slice(0, 200),
+          hasLink: $row.find('.cell-team-title a').length > 0,
+          hasFavorite: $row.find('.rating-mailings__favorite').length > 0,
+          hasCoefficient: $row.find('.cell-coefficient__total, .cell-coefficient').length > 0
+        };
+      });
+
+    console.log('tableLinkCandidates:', JSON.stringify(rowCandidates, null, 2));
+    if (items[0]) {
+      console.log('firstItem:', JSON.stringify(items[0], null, 2));
+    }
+  } finally {
+    await browser.close();
+  }
+}
+
+main().catch((error) => {
+  console.error(error);
+  process.exit(1);
+});