Tonghuashun (同花顺) Finance Scraper

Scrapes 10jqka.com.cn — one of China's largest stock/finance platforms — for A-share quotes and market news.

Data Source

Public JSON APIs (no login, no cookies):

• Quote: d.10jqka.com.cn/v2/realhead/hs_{code}/last.js
• News: news.10jqka.com.cn/tapp/news/push/stock/
• Name→code resolution: smartbox.gtimg.cn (public Tencent suggest service)

Both verified reachable without authentication.

Modes

Mode	Inputs	Source
byStock (default)	stockCodes	realhead quote API; codes accept 600519, sh600519, 000001.SZ forms
marketNews	maxNewsItems	paginated news feed with tags, timestamps, related stocks
search	searchQuery	code → direct quote; Chinese company name → resolved to codes via the public Tencent smartbox (smartbox.gtimg.cn), with 10jqka news-feed associations as fallback, then quotes fetched
byStockUrl	stockUrls	extracts the 6-digit code from any 10jqka URL and fetches its quote

Output fields

stock records: stockCode, stockName, price, prevClose, open, high, low, volume (shares), turnover (CNY), avgPrice, bid1Price, ask1Price, change, changePct, high52w, low52w, peRatio (TTM), staticPERatio, turnoverRate (%), volumeRatio, amplitude (%), totalShares, marketCap, floatMarketCap, exchange (sh/sz), marketStatus, quoteTime, sourceUrl.

All numeric fields are cross-validated against Tencent's realtime quote API (qt.gtimg.cn); field IDs that fail validation are not shipped.

news records: newsId, title, digest, url, tags, relatedStocks ([{name, stockCode}]), publishedAt (UTC ISO), sourceUrl.

Every record carries recordType, sourceUrl and scrapedAt.

Filters

• minChangePct — drop stock records with |changePct| below the threshold (news records unaffected).

Reliability

• Exponential-backoff retries on 429/5xx and the intermittent TLS resets this Chinese CDN produces.
• On 403/429 the actor lazily engages the free Apify AUTO datacenter proxy group.
• Invalid codes / unreachable quotes produce typed recordType: "error" records.
• Duplicate inputs are deduped: repeated codes resolve once (600519 + sh600519 → one quote), repeated invalid codes produce one error record.

Limitations

• The numeric field IDs in the quote payload were cross-validated against Tencent's quotes; if 10jqka changes a field ID, that one field silently drops (omit-empty), others remain correct.
• Full-text company-name search resolves via the public Tencent smartbox (smartbox.gtimg.cn), because 10jqka's own search suggest API was retired (search.10jqka.com.cn now runs a captcha-gated AI gateway). If smartbox returns no match, the actor falls back to 10jqka's news-feed stock associations (up to 10 pages scanned) — a name only resolves via the fallback if it appears in the recent feed. Codes always resolve directly.
• News article and stock-page URLs (url, stockpage.10jqka.com.cn/...) return 403 to bare HTTP clients without a browser User-Agent; they load normally in a browser. The data URLs emitted on records (sourceUrl against d.10jqka.com.cn) answer 200 from any client.