# Changelog of Baidu Baike Scraper (`searchapi/baidu-baike-scraper`) Actor

- **URL**: https://apify.com/searchapi/baidu-baike-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/searchapi/baidu-baike-scraper.md

## Changelog

### 2.3.1

- Broadened Baike result-link extraction to tolerate changed wrappers and tracking parameters while accepting only visible official `/item/` links.
- Added fail-closed selector-drift handling; failed runs retain their bounded status/count summary before the terminal error is surfaced.
- Expanded verification-page detection without storing page bodies or challenge content.

### 2.3.0

- Fixed Baidu Baike search pagination when the live page omits a visible next link by preserving the page's search parameters and advancing its observed `pn` offset.
- Added bounded pagination tracking to the `OUTPUT` summary.
- Stopped pagination when a page contributes no new entry requests, preventing duplicate-page loops.
- Added a Browserslist security override; the remaining audit findings are upstream moderate `stream-json` issues without an available fix.

### 2.2.0

- Added title, summary, content, section, infobox, media, reference, and related-entry metrics.
- Added breadcrumb category paths, structured section details, related Baike item links, inclusion flags, and OpenGraph/page metadata.
- Expanded the dataset contract and live validation coverage for the richer DOM output.

### 2.0.0

- Added single-query, multiple-query, direct-entry, and mixed modes with exact global and per-query limits.
- Reduced representative three-entry runtime from about 99 seconds to about 31 seconds through two-phase discovery, bounded entry concurrency, and removal of repeated optional-selector waits.
- Added stable IDs, canonical URL normalization, search metadata, extraction provenance, and atomic writes.
- Added configurable content/image/reference detail modes and preserved rich default extraction.
- Added consistent browser fingerprints, sessions, authorized proxy support, strict block detection, bounded retries/timeouts, schemas, docs, fixtures, and regression tests.
- Pinned deterministic dependencies and remediated production dependency advisories.