maxun-core 0.0.14 → 0.0.15

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (2) hide show
  1. package/build/interpret.js +32 -6
  2. package/package.json +1 -1
@@ -477,6 +477,7 @@ class Interpreter extends events_1.EventEmitter {
477
477
  let visitedUrls = new Set();
478
478
  const MAX_RETRIES = 3;
479
479
  const RETRY_DELAY = 1000; // 1 second delay between retries
480
+ const MAX_UNCHANGED_RESULTS = 5;
480
481
  const debugLog = (message, ...args) => {
481
482
  console.log(`[Page ${visitedUrls.size}] [URL: ${page.url()}] ${message}`, ...args);
482
483
  };
@@ -555,30 +556,55 @@ class Interpreter extends events_1.EventEmitter {
555
556
  }
556
557
  });
557
558
  let availableSelectors = config.pagination.selector.split(',');
559
+ let unchangedResultCounter = 0;
558
560
  try {
559
561
  while (true) {
560
- // Reduced timeout for faster performance
561
- yield page.waitForLoadState('networkidle', { timeout: 10000 }).catch(() => { });
562
562
  switch (config.pagination.type) {
563
563
  case 'scrollDown': {
564
+ let previousResultCount = allResults.length;
565
+ yield scrapeCurrentPage();
566
+ if (checkLimit()) {
567
+ return allResults;
568
+ }
564
569
  yield page.evaluate(() => window.scrollTo(0, document.body.scrollHeight));
565
570
  yield page.waitForTimeout(2000);
566
571
  const currentHeight = yield page.evaluate(() => document.body.scrollHeight);
572
+ const currentResultCount = allResults.length;
573
+ if (currentResultCount === previousResultCount) {
574
+ unchangedResultCounter++;
575
+ if (unchangedResultCounter >= MAX_UNCHANGED_RESULTS) {
576
+ return allResults;
577
+ }
578
+ }
579
+ else {
580
+ unchangedResultCounter = 0;
581
+ }
567
582
  if (currentHeight === previousHeight) {
568
- const finalResults = yield page.evaluate((cfg) => window.scrapeList(cfg), config);
569
- allResults = allResults.concat(finalResults);
570
583
  return allResults;
571
584
  }
572
585
  previousHeight = currentHeight;
573
586
  break;
574
587
  }
575
588
  case 'scrollUp': {
589
+ let previousResultCount = allResults.length;
590
+ yield scrapeCurrentPage();
591
+ if (checkLimit()) {
592
+ return allResults;
593
+ }
576
594
  yield page.evaluate(() => window.scrollTo(0, 0));
577
595
  yield page.waitForTimeout(2000);
578
596
  const currentTopHeight = yield page.evaluate(() => document.documentElement.scrollTop);
597
+ const currentResultCount = allResults.length;
598
+ if (currentResultCount === previousResultCount) {
599
+ unchangedResultCounter++;
600
+ if (unchangedResultCounter >= MAX_UNCHANGED_RESULTS) {
601
+ return allResults;
602
+ }
603
+ }
604
+ else {
605
+ unchangedResultCounter = 0;
606
+ }
579
607
  if (currentTopHeight === 0) {
580
- const finalResults = yield page.evaluate((cfg) => window.scrapeList(cfg), config);
581
- allResults = allResults.concat(finalResults);
582
608
  return allResults;
583
609
  }
584
610
  previousHeight = currentTopHeight;
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "maxun-core",
3
- "version": "0.0.14",
3
+ "version": "0.0.15",
4
4
  "description": "Core package for Maxun, responsible for data extraction",
5
5
  "main": "build/index.js",
6
6
  "typings": "build/index.d.ts",