Skip to content

knowledge: 同步 FinMind-Doc 8/20~9/29 dataset 新增與說明更新 - #23

Merged
linsamtw merged 1 commit into
masterfrom
feature/catalog-sync-sep
Sep 29, 2026
Merged

linsamtw merged 1 commit into
masterfrom
feature/catalog-sync-sep

Conversation

@linsamtw

Copy link
Copy Markdown
Contributor

Summary

上次 catalog 同步是 #22(2026-08-20)。本次對齊 FinMind-Doc origin/master(至 #219),只改 knowledge/datasets.md,catalog 106 → 107。

新增

  • TaiwanFuturesKBar(Sponsor;必帶 data_id、單日;整日全商品走 storage_objects,Sponsor Pro,2011-01-03 起)— doc-sync gate 原本會紅在這個

分類 / 區間修正

  • TaiwanAssetSwapFixedIncomeDaily、TaiwanAssetSwapOptionDaily:搬到「可轉債」分類(對齊 Doc),資料區間 2026-06-01 → 2011-05-03

既有 dataset 說明更新(gate 抓不到的能力/口徑漂移)

  • TaiwanStockPriceTick:區間 2018-12-07 起、已知缺漏日、TickType 全區間修正(0=無法判定)、整日檔排序規則
  • TaiwanStockKBar:缺漏日更新;大盤分 K 用 TAIEX(001 已移除)
  • TaiwanFuturesTick / TaiwanOptionTick:必帶 data_id,整日改 storage_objects;期貨逐筆雙邊計量+議價申報鉅額;移除 OptionTick 2019 上半年不完整註記
  • 三大法人(長/寬表):法人別新舊制,Foreign_Dealer_Self 上市 2017-12-18 起
  • TaiwanStockTradingDailyReport:分點買≠賣為來源特性
  • TaiwanStockMonthRevenue:create_time 語意
  • TaiwanStockStatisticsOfOrderBookAndTrade、TaiwanStockEvery5SecondsIndex:統計範圍、13:30 收盤、名稱沿革、已知缺漏

Test plan

  • scripts/check_doc_sync.py 對 FinMind-Doc 最新 master:OK
  • uv run --extra dev pytest -q:32 passed;smoke.py:SMOKE OK
  • build_instructions.py:instructions.txt 7987/8000
  • live API ?dataset=__INVALID__ enum 確認 TaiwanFuturesKBar、兩個 AssetSwap 拼字正確;TaiwanFuturesKBar 免費帳號回 level 限制(非 422)

merge 後需下新 tag 才會發到 PyPI 使用者端;Custom GPT 需重跑 build_instructions.py 重貼 knowledge bundle。

修改檔案:knowledge/datasets.md(catalog 106 → 107)

- 新增 TaiwanFuturesKBar(Sponsor,必帶 data_id、單日;整日全商品走
  storage_objects,Sponsor Pro,2011-01-03 起)
- TaiwanAssetSwapFixedIncomeDaily / TaiwanAssetSwapOptionDaily:
  從「衍生性金融商品」搬到「可轉債」分類(對齊 Doc),資料區間
  2026-06-01 → 2011-05-03
- TaiwanStockPriceTick:資料區間 2018-12-07 起、已知缺漏日期、
  TickType 全區間修正與 0=無法判定說明、整日檔排序規則
- TaiwanStockKBar:已知缺漏日期更新;大盤分 K 用 TAIEX(001 已移除)
- TaiwanFuturesTick / TaiwanOptionTick:必帶 data_id、整日改用
  storage_objects;期貨逐筆雙邊計量與議價申報鉅額說明;移除
  OptionTick 2019 上半年不完整註記(已補齊)
- TaiwanStockInstitutionalInvestorsBuySell / Wide:法人別新舊制
  (Foreign_Dealer_Self 上市 2017-12-18 起)
- TaiwanStockTradingDailyReport:分點買進合計≠賣出合計為來源特性
- TaiwanStockMonthRevenue:create_time 語意
- TaiwanStockStatisticsOfOrderBookAndTrade:累積成交統計範圍與資料間隔
- TaiwanStockEvery5SecondsIndex:13:30 收盤、化學類名稱沿革、已知缺漏

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@linsamtw
linsamtw merged commit 06313e8 into master Sep 29, 2026
4 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant