Bing SERP scraper v0.3 adds technical isolation (new BrowserContext, index-lock query params, strip pq/cvid, international toggle, language:en last resort) so CN session bleed is not treated as a long_gap.

Date: 2026-09-01 10:45 Summary: Bing SERP scraper v0.3 adds technical isolation (new BrowserContext, index-lock query params, strip pq/cvid, international toggle, language:en last resort) so CN session bleed is not treated as a long_gap.

[question] 技术也针对bing serp进行策略优化处理

[try to solve] User-habit query rewriting (v0.2) was not enough: search-box submits often drop ensearch=1 and attach pq= from the previous keyword, and one Playwright page still shares cookies (MUID, conversation id). Technical layer now:

New BrowserContext per query with Accept-Language: en-US, US timezone/geolocation.
URL lock: setmkt + ensearch=1 + qs=n + sp=-1 + lq=0 (does not change the user's query text).
After the search box submits, reload without pq/cvid and restore the lock.
Click the on-page 国际版 control when present; skip ad packs; CJK-heavy titles or captcha → unusable.
If habit variants are still polluted, one extra fetch with language:en (research only, never an H1).

Documented as strategy §3.3 I.2. Flags: --reuse-page, --no-lang-lock.

[actions]

Updated ops/seo/bing_serp/{browser_util,bing_parse,query_strategy,run_bing_serp,analyze} to v0.3.0
Extended docs/seo/2026-08-20-long-tail-gap-strategy.md §3.3 I.2 and bing_serp README

Bing SERP scraper v0.3 adds technical isolation (new BrowserContext, index-lock query params, strip pq/cvid, international toggle, language:en last resort) so CN session bleed is not treated as a long_gap. | Dev Logs