Weibo Keyword Search Scraper
Pricing
$0.10 / actor start
Search Weibo posts by keyword. 7 search_type filters (综合/实时/用户/热门/视频/图片/精华) and 4 time_scope filters. Page=1 only (~10 items) due to upstream login wall.
UpstreamClient; env vars TIKHUB_API_KEY / TIKHUB_BASE_URL → UPSTREAM_API_KEY / UPSTREAM_BASE_URL; file src/tikhub_client.js → src/upstream_client.js. Per our internal SOP §100 (no upstream name in user-facing surfaces), existing 0.1.4 runs keep working until the next push.test/integration.js was the hot-search template (called fetchTrendTop + intervalSeconds); replaced with a keyword-specific test that calls fetchSearch and asserts keyword + searchType on the first item. 4/4 unit + 1/1 integration tests PASS.keyValueStore.setValue('SUMMARY', {...}) requires JSON.stringify + contentType: 'application/json', otherwise the SDK rejects (value must be String/Buffer/Stream when options.contentType is specified).JSON.stringify(current + 1) write / parse-on-read (the previous text/plain + string round-trip broke because getValue already auto-deserialises JSON).apify call SUCCEEDED, ~10 search items pushed to default dataset.rXAua7eiA6jxRogdF).TBD (next in MVP-2a).apify push build 0.1.1 → 0.1.2.rXAua7eiA6jxRogdF.searchType values (1=综合, 61=实时, 3=用户, 60=热门, 64=视频, 63=图片, 21=精华) all wired per §40.timeScope values (不限/一小时内/一天内/一周内/一个月内) wired.PAGE_HARD_CAP = 1 because page=2+ triggers upstream login wall (§39 maxItems must match upstream).extractMblogs() (handles both cards[].mblog and cards[].card_group[].mblog).