# Changelog of Baidu Zhidao Scraper (`searchapi/baidu-zhidao-scraper`) Actor

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

## Changelog

### Unreleased

- Report every refused question page with bounded evidence (observed document title, title kind, body length, and a redacted body excerpt), logged per page and echoed in `OUTPUT.detailFailureSamples`.
- A question page that yields no usable record now fails closed with the observed evidence instead of returning a null record that silently shrank the result set.
- Separate interactive verification, access notices, rate limits, transport failures, and extraction failures in run status diagnostics.
- Redact query strings from failure reasons and include the primary failure class, code, and safe reason in the run summary.
- Parse result cards before checking page text for access markers, and stop before fallback navigation when a target refusal is identified.

### 2.0.0 - 2026-09-06

- Replaced the fixed Crawlee flow with a persistent Playwright browser runner that uses the observed Zhidao homepage search form, public-DNS host mapping when system DNS is unavailable, and DOM extraction with fail-closed challenge detection.
- Added rich normalized question and answer fields, canonical URLs, counts, categories, tags, related questions, provenance, and search-result metadata.
- Added multi-query fair selection, bounded pagination, direct public question mode, filters, sorting, request/detail concurrency, retries, timeouts, request delay, media blocking, proxy configuration, and safe run summaries.
- Added dataset, output, and key-value-store schemas plus deterministic input, README, dataset, and run validators.