Fleur is a 100% AI-coded research platform for China A-shares that covers data collection, technical indicator computation, rule-based stock selection, strategy backtesting, and portfolio monitoring, all built and run entirely for free.

Stars

93

7-day growth

No data

Forks

18

Open issues

0

License

MIT

Last updated

2026-07-16

AI repository intelligence
FR-AI / ANALYSIS

Why it is worth attention

It is a fully open-source, end-to-end investment research platform that combines Dagster for data orchestration, dbt for ELT modeling, a Rust-based indicator engine (Furnace), and a React frontend (Racingline), all implemented by AI coding agents, with a strong emphasis on data traceability, explainable indicators, and zero-code strategy building.

Who it is for

  • Quantitative researchers and developers interested in A-share markets
  • Individual investors seeking a free, self-hosted backtesting environment
  • Data engineers learning modern data stack (Dagster, dbt, ClickHouse) on financial data
  • Financial technology enthusiasts exploring AI-assisted development workflows

Use cases

  • Backtesting custom stock selection strategies using 100+ technical indicators
  • Learning technical analysis with transparent, auditable indicator formulas
  • Building and monitoring a simulated portfolio with real-time signals
  • Practicing a modern data pipeline (ETL, data contracts, S3, ClickHouse) on open public data

Strengths

  • Fully self-contained data pipeline from public sources with no paid subscriptions
  • High-performance Rust engine for millisecond-level batch indicator computation
  • Zero-code visual strategy builder for hundreds of technical indicators
  • Comprehensive portfolio evaluation metrics including Sharpe, Sortino, and Calmar ratios

Considerations

  • Data is sourced from free public APIs and may have delays, rate limits, or quality issues
  • Designed purely for learning and research – no real-money trading or financial advice
  • Requires significant hardware (6+ vCPU, 20+ GB RAM) for full deployment
  • Setup and operation require familiarity with Python, Rust, Docker, and data engineering tools

README quick start

Fleur

[!IMPORTANT] 本项目仅研究与学习用途,不涉及资金交易,所有数据、指标、策略与 AI 输出仅供学习参考, 不保证数据真实有效,不构成任何投资建议,运行项目前请务必阅读 风险提示与免责声明

📖 项目介绍

Fleur 是一个由 harness-engineering 驱动的 100% AI Coding 的面向沪深 A 股投研平台,覆盖行情与财务数据采集、技术指标计算、规则选股、策略回测、及组合运行监控。

🎯 Fleur 聚焦的问题

  • 数据自采自洽:采集中文互联网公开数据,从源头治理字段事实,做到数据高质量、口径自洽、可重放、可回填。
  • 指标公式可解释: 实现指标计算引擎,技术指标指标算法可溯源。
  • 可视化策略构建:支持 行情涨跌\趋势指标\动量指标\形态特征\量能指标 等数上百个技术指标的可视化实时筛选,构建策略零代码。
  • 快速策略回测:实现高性能回测引擎与投资组合评价体系,快速多维度评价策略过往业绩表现。
  • 成本控制:完全免费、从数据到应用不产生任何费用。

🧩 领域与感知

领域工程感知
数据基座Pipeline(数据管道)稳定、可追溯的数据接入,确保原始数据完整留存,支持回放与审计。
ELT 分层建模清晰的数据分层,降低分析复杂度;指标口径统一,即用即查,减少重复开发。
数据契约治理数据质量有保障,字段变更可追踪;自动生成的文档让快速上手,减少沟通成本。
计算引擎与服务Furnace(指标计算)毫秒级批量计算,支持海量证券标的;指标结果可靠,可直接用于策略逻辑。
Rearview(主服务)实时信号推送、组合监控,辅助决策与风险控制。
Rearview-worker(回测引擎)快速验证投资想法,回测历史表现,获取绩效指标。
工作台Racingline(前端)一站式交互界面,无需编写代码即可构建策略、观察回测绩效、管理模拟持仓,降低使用门槛。

📷 功能截图

工作台围绕「选股 → 回测 → 模拟建仓 → 跟踪订阅」链路,把指标计算、策略编排与可视化看板串成一条交互主线。

策略看板
策略详情







指标选股
股池预览







模拟建仓
策略回测




🏗️ 架构总览

🧱 数据基座

本项目所使用数据均来自中文互联网可合法免费获取的公开数据源,所有数据采集、加工、建模均在本项目内编码实现,不产生、不收取任何费用。

顺序流向落点责任边界
1公开数据源BaoStock、EastMoney、Sina、THS、ChinaBond、JiuYan 等提供行情、会计财务报表、分红派息、涨停池、题材、图片等原始输入
2数据资产管理pipeline/scheduler采集、分区、重试、回填和运行观测
3源层数据RustFS(S3) 对象存储中的 Parquet保存可重放的源层数据对象
4数据建模与治理pipeline/elt数据加工、清洗、跨源组合、计算、可复用业务中间结构及业务口径输出
5业绩指标计算引擎engines/crates/furnace计算 KDJ、MA、RSI、BOLL、MACD、price-pattern 等技术指标
7分层模型ClickHouseraw 原始数据层、staging 字段规范化层、intermediate 可复用业务中间结构 、 calculation 业绩指标计算 、 marts 应用数据集市
8应用消费Rearview API、Portfolio worker、Racingline 工作台规则选股、策略回测、组合运行

📡 多源沪深 A 股数据采集

通过 Dagster 编排六类外部数据源,统一落盘 Parquet 文件格式(zstd压缩编码)上传至RustFS(S3 兼容),后同步至 ClickHouse raw 层。下表汇总各数据源的数据范围与注册要求:

数据源采集内容数据范围是否需要注册
BaoStock 证券宝沪深 A 股基础信息快照、日 K 线(OHLC)1990-12-19 至今✅ 全免费,无需注册,IP限流
EastMoney 东方财富资产负债表、利润表(SQ/YTD)、现金流(SQ/YTD)、分红派息、股权历史、十大流通股东1990 至今✅ 全免费,无需注册
Sina 新浪财经A 股交易日历(压缩位流解码)1990-12-19 至本年年末✅ 全免费,无需注册
THS 同花顺涨停股池(近 380 个自然日滚动)近 380

Description

Fleur 是一个由 harness-engineering 驱动 100% AI Coding 的面向沪深A股投研平台,功能覆盖行情与财务数据采集、技术指标计算、规则选股、策略回测、及组合运行监控。

Related repositories

Similar projects matched by category, topics, and programming language.

l0ng-ai
Featured
l0ng-ai GitHub avatar

tty7

tty7 is a fast, GPU-accelerated terminal workbench with persistent sessions, built-in input tools, SSH support, and coding agent awareness.

Developer ToolsCLI & Terminal
359
steelbrain
Featured
steelbrain GitHub avatar

reims-vgpu

reims-vgpu is an experimental virtual GPU for macOS guests that uses QEMU to decode the guest's GPU command stream and execute it through Metal or Vulkan, leveraging the built-in AppleParavirtGPU driver without requiring custom kexts.

Desktop Apps
143
m-novotny
Featured
m-novotny GitHub avatar

memguard-rs

A Rust library that provides secure memory handling primitives including zeroization on drop, memory locking, constant-time comparison, and compile-time guarded regions, with zero dependencies and no_std support.

Embedded & IoTSecurity
131