Repository navigation
Data sources & plugins: source rights, access and public-safe examples / 数据共建 #164
Replies: 1 comment
From links to a working example: CCXT + Kraken -> dsh-quantThanks to everyone sharing data sources and practical integration ideas. We Code: example and install guide Run from a source checkoutUse a maintained Node24 release. Review Kraken's terms first, then from the npm ci
npm ci --prefix demos/ccxt-public --ignore-scripts
node --import tsx demos/ccxt-public/run.ts --live --personal-researchNo account, API key, paid endpoint, model inference or trade is used. Optional Actual benchmark, not a speedup claimAccepted live run:2026-11-14 15:47:34–15:47:44, Asia/Shanghai.
Latency includes scheduled rate limiting, response hashing and initial market Protocol, comparisons and limitations License and use boundaryCCXT4.5.92 is MIT code; that does not license Kraken's market data. Public Keep building data plugins togetherPlease share a source/plugin/MCP/SDK, one concrete use, access/fee limits, code Thanks for the collaboration. We welcome newcomers and continued PRs, give |
Uh oh!
There was an error while loading. Please reload this page.
Data sources & plugins: where does your research data come from?
欢迎分享 Data 相关的 plugin、MCP、SDK、来源和工程项目。我们这次先做了一轮
有边界的一手来源调研,再来邀请共建,而不是只发一个空泛征集帖。
First review batch / 本轮实际整理
首批11个项目:LLMQuant/data-mcp、AKShare、AKTools、OpenBB、CCXT、TuShare、
yfinance、fredapi、EdgarTools、DuckDB 和 dlt。它们分别属于 MCP/桥接、数据 SDK、
集成平台或存储/摄取工程,不全是原生 DSH 插件,也不全是数据供应商。
完整数据来源与接入报告
逐项列出用途、固定 revision/许可证证据、账号/数据权利边界与待验证方向。
本轮引用状态为 LINK_ONLY / NOT_RUN,没有安装上游项目、创建账号或购买服务。
不是全网穷尽、优劣排名、数据质量认证或真实行情性能 benchmark。
LLMQuant Data 已经收录,不再重复新增。我们修正了现有渠道说明:MIT MCP
代码与后台数据服务不是一回事,接入有账号/凭据/credit 边界。文档提到 beta
试用,不代表无限免费;pricing 页面本轮返回403,实际现金价格、注册与付费均未
验证。检查与用户自主验证流程
写明了哪些步骤有一手依据、哪些还没跑。不要在这里发送 API key 或完整托管 MCP URL。
也只读查看了 PDAT(proprietary/private repo)。本轮选出的公开材料是
通用来源契约和新合成样例,
不是私有源码镜像:不复制内部采集器、客户/策略、日志、凭据或第三方真实数据。
What to share / 怎么参与
在本帖回复项目链接与一个具体用例即可;需要收录/实现跟踪时沿用
Issue #2。FICC 专题仍用
#152 和
#151,不再造平行投稿队列。
自己的项目、第三方资源、小型教学例子和失败复盘都欢迎。不要提交未经许可的数据、
密钥、客户资料、私有持仓或策略。代码许可、技术可访问和数据再分发权分别核验。
Collaboration and review
我们按
source/rights review -> scoped issue/PR -> tests/CI -> merge -> credit推进。收录不等于集成,凭据就绪不等于 API 实测,不为未知模型/权限问题硬贴新人
标签。已经整理好的新人任务仍见 #162。
Thanks to upstream maintainers and everyone who shares a reproducible example,
limitation or correction. We appreciate your time, give specific review feedback,
preserve attribution and thank contributors after accepted merges. Keep going;
we welcome continued collaboration, without lowering the review standard.
All reactions