← Back to Skills Marketplace
ugpoor

DolphinDB 量化金融技能

by superStupidBear · GitHub ↗ · v1.4.0 · MIT-0
cross-platform ⚠ suspicious
239
Downloads
0
Stars
2
Active Installs
6
Versions
Install in OpenClaw
/install dolphindb-quant-finance
Description
基于 DolphinDB 提供因子计算、策略回测、行情处理、绩效归因及投资组合优化的量化金融专业解决方案。
README (SKILL.md)

DolphinDB 量化金融技能 v1.1.3

🚨 强制流程:使用前必须加载环境

无论在何种场景下调用此技能(单独运行或被引用),必须先执行环境检测:

# 方法 1: 在技能目录内运行(推荐)
cd ~/.jvs/.openclaw/workspace/skills/\x3Cskill-name>
source ../dolphindb-skills/scripts/dolphin_wrapper.sh

# 方法 2: 在任何位置运行(推荐)
source ~/.jvs/.openclaw/workspace/skills/dolphindb-skills/scripts/dolphin_global.sh

# 方法 3: 手动检测
python3 ~/.jvs/.openclaw/workspace/skills/dolphindb-skills/scripts/init_dolphindb_env.py

验证环境:

$DOLPHINDB_PYTHON_BIN -c "import dolphindb; print(dolphindb.__version__)"

# 或使用包装器命令
dolphin_python -c "import dolphindb; print(dolphindb.__version__)"

重要: 详见 dolphindb-skills/USAGE_GUIDE.md


⚠️ 前置依赖

本技能依赖 dolphindb-basic 技能,请先安装:

clawhub install dolphindb-basic

运行前初始化需确保 Python 环境有 DolphinDB SDK,详细方法可参见 dolphindb-skills 技能。

# 加载环境检测器(相对路径,技能安装后自动可用)
source ../dolphindb-skills/scripts/load_dolphindb_env.sh

# 查看环境信息
dolphin_env_info

# 验证 SDK 已安装
dolphin_python -c "import dolphindb; print('SDK 版本:', dolphindb.__version__)"

统一调用接口:

dolphin_python script.py    # 运行 Python 脚本
dolphin_pip install pkg     # 安装包

重要:所有 DolphinDB 脚本在 Python 中的调用方式

import dolphindb as ddb

# 1. 建立连接
s = ddb.session()
s.connect(host="localhost", port=8848, userid="admin", password="123456")

# 2. 执行 DolphinDB 脚本(所有数据库操作都通过 s.run())
result = s.run('''
    // DolphinDB 脚本
    use mytt
    select 
        sym,
        RSI(close, 24) as rsi,
        MA(close, 5) as ma5
    from loadTable("dfs://bars_db.bars_minute")
    context by sym
''')

# 3. 转换为 pandas DataFrame(可选)
df = result.toDF()

# 4. 关闭连接
s.close()

描述

提供基于 DolphinDB 的量化金融场景解决方案,包括因子计算、策略回测、行情处理、绩效归因等金融专业能力。

触发条件

当用户提到以下关键词时触发此技能:

  • "量化因子"、"因子计算"、"多因子模型"
  • "策略回测"、"backtest"、"回测引擎"
  • "K 线合成"、"OHLC"、"行情数据"
  • "Level-2 数据"、"逐笔数据"、"订单簿"
  • "绩效归因"、"Brinson"、"Campisi"
  • "投资组合"、"权重优化"、"MVO"
  • "实时计算"、"流式因子"、"高频因子"
  • "Alphalens"、"因子分析"、"IC 分析"

能力范围

1. 因子计算

  • 日频/分钟频因子计算
  • 高频因子实时计算
  • Alpha101 因子库
  • TA-Lib 技术指标
  • 流批一体因子计算平台

2. 策略回测

  • 股票中低频策略回测
  • 期货回测
  • 期权回测
  • 多资产组合回测
  • 模拟撮合引擎

3. 行情处理

  • K 线合成(分钟/小时/日)
  • Level-2 数据处理
  • 订单簿合成
  • 行情回放
  • 复权计算

4. 绩效归因

  • Brinson 绩效归因
  • Campisi 绩效归因
  • 因子归因分析
  • 多因子风险模型

5. 投资组合优化

  • 均值方差优化(MVO)
  • 约束优化(SOCP)
  • 权重分配
  • 风险控制

6. 实时计算

  • 实时高频因子
  • 实时资金流计算
  • 实时涨幅榜
  • 实时波动率预测

使用示例

1. 因子计算最佳实践

import dolphindb as ddb

s = ddb.session()
s.connect(host="localhost", port=8848, userid="admin", password="123456")

# 加载 MyTT 模块计算技术指标
s.run('use mytt')

# 计算均线、RSI、波动率
result = s.run('''
    // 计算均线
    def getMA(close, period=5){
        return MA(close, period)
    }
    
    // 计算 RSI
    def getRSI(close, N=24){
        return RSI(close, N)
    }
    
    // 计算波动率
    def getVolatility(close, period=10){
        returns = DIFF(close) / REF(close, 1)
        return STD(returns, period)
    }
    
    // 批量计算因子
    select 
        sym,
        getMA(close, 5) as ma5,
        getMA(close, 20) as ma20,
        getRSI(close, 24) as rsi,
        getVolatility(close, 10) as vol
    from loadTable("dfs://bars_db.bars_minute")
    where date = 2024.01.01
    context by sym
''')

df = result.toDF()
s.close()

2. 股票回测配置

import dolphindb as ddb

s = ddb.session()
s.connect(host="localhost", port=8848, userid="admin", password="123456")

# 回测参数配置
s.run('''
    config = dict(STRING, ANY)
    config["startDate"] = 2024.01.01
    config["endDate"] = 2024.12.31
    config["strategyGroup"] = "stock"
    config["frequency"] = 0
    config["cash"] = 10000000.0  // 初始资金 1000 万
    config["commission"] = 0.0005  // 交易佣金
    config["tax"] = 0.0
    config["dataType"] = 3  // 分钟频
    config["matchingMode"] = 3  // 委托价格成交
    config["msgAsTable"] = false
    
    // 上下文配置
    context = dict(STRING, ANY)
    context["activeTime"] = 14:30m  // 每日交易时间
    context["maxPosition"] = 0.1    // 单只股票最大持仓 10%
    context["minPosition"] = 0.001  // 单只股票最小持仓 0.1%
    config["context"] = context
''')

s.close()

3. K 线合成

import dolphindb as ddb

s = ddb.session()
s.connect(host="localhost", port=8848, userid="admin", password="123456")

# 基于 tick 数据合成分钟 K 线
result = s.run('''
    def synthesizeOHLC(tickData, interval=60000){
        return select 
            first(time) as openTime,
            last(time) as closeTime,
            first(open) as open,
            max(high) as high,
            min(low) as low,
            last(close) as close,
            sum(volume) as volume,
            sum(turnover) as turnover
        from tickData
        group by sym, time_bar(interval, time) as barTime
    }
''')

# 使用内置 OHLC 引擎
s.run('''
    engine = createOHLEngine(
        name=`ohlcEngine,
        windowSize=50000,
        timeColumn=`time,
        groupingColumn=`sym,
        buckets=100,
        freq=60000,  // 1 分钟
        metrics=`open`high`low`close`volume
    )
''')

s.close()

4. Level-2 订单簿合成

import dolphindb as ddb

s = ddb.session()
s.connect(host="localhost", port=8848, userid="admin", password="123456")

# 基于逐笔数据合成订单簿
s.run('''
    use orderbook
    
    engine = createOrderBookEngine(
        name=`obEngine,
        windowSize=100000,
        timeColumn=`time,
        symColumn=`sym,
        priceColumn=`price,
        volumeColumn=`volume,
        bsFlagColumn=`bsFlag,  // B=买,S=卖
        level=10  // 10 档行情
    )
    
    // 订阅订单簿输出
    subscribeTable(
        tableName=`obEngine,
        actionName=`processOB,
        offset=0,
        handler=processOrderBook
    )
''')

s.close()

5. Brinson 绩效归因

import dolphindb as ddb

s = ddb.session()
s.connect(host="localhost", port=8848, userid="admin", password="123456")

# Brinson 归因分析
s.run('use brinson')

# 配置归因并执行
result = s.run('''
    // Brinson 归因分析
    brinsonConfig = dict(STRING, ANY)
    brinsonConfig["portfolioReturn"] = portfolioRet  // 组合收益
    brinsonConfig["benchmarkReturn"] = benchmarkRet  // 基准收益
    brinsonConfig["portfolioWeights"] = portWeights  // 组合权重
    brinsonConfig["benchmarkWeights"] = benchWeights // 基准权重
    
    // 执行归因
    brinsonResult = brinsonAttribution(brinsonConfig)
    // 输出:配置效应、选股效应、交互效应
''')

df = result.toDF()
s.close()

6. 投资组合优化(MVO)

import dolphindb as ddb

s = ddb.session()
s.connect(host="localhost", port=8848, userid="admin", password="123456")

# 均值方差优化
s.run('use mvo')

# 准备数据并执行优化
result = s.run('''
    // 准备数据
    returns = loadReturns()  // 历史收益率矩阵
    covMatrix = cov(returns) // 协方差矩阵
    expectedReturns = mean(returns) // 预期收益
    
    // 配置优化参数
    mvoConfig = dict(STRING, ANY)
    mvoConfig["expectedReturns"] = expectedReturns
    mvoConfig["covariance"] = covMatrix
    mvoConfig["riskFreeRate"] = 0.02  // 无风险利率 2%
    mvoConfig["constraints"] = dict(
        "sumWeights" = 1.0,      // 权重和为 1
        "minWeight" = 0.0,       // 最小权重 0
        "maxWeight" = 0.2        // 最大权重 20%
    )
    
    // 执行优化
    optimalWeights = mvoOptimize(mvoConfig)
''')

s.close()

7. 实时高频因子计算

import dolphindb as ddb

s = ddb.session()
s.connect(host="localhost", port=8848, userid="admin", password="123456")

# 创建流数据表
s.run('streamTable(10000:0, `time`sym`price`volume, [TIMESTAMP,SYMBOL,DOUBLE,LONG])')

# 定义实时因子计算函数并创建引擎
s.run('''
    def calcRealtimeFactor(data){
        return select 
            sym,
            wap(price, volume) as vwap,  // 成交量加权均价
            (max(price) - min(price)) / last(price) as range,  // 价格区间
            sum(volume) as totalVolume,
            count(*) as tradeCount
        from data
        group by sym, time_bar(60000, time) as minute
    }
    
    engine = createStreamEngine(
        name=`realtimeFactorEngine,
        handler=calcRealtimeFactor,
        streamTableNames=`tickStream,
        windowSize=1000,
        timeColumn=`time,
        groupingColumn=`sym
    )
''')

s.close()

8. Alphalens 因子分析

import dolphindb as ddb

s = ddb.session()
s.connect(host="localhost", port=8848, userid="admin", password="123456")

# 使用 Alphalens 框架分析因子
s.run('use alphalens')

# 准备因子数据和收益数据并执行分析
result = s.run('''
    // 准备因子数据和收益数据
    factorData = loadFactorData()  // 因子值
    returnsData = loadReturnsData() // 未来收益
    
    // 配置分析参数
    alphalensConfig = dict(STRING, ANY)
    alphalensConfig["factor"] = factorData
    alphalensConfig["returns"] = returnsData
    alphalensConfig["periods"] = [1, 5, 10, 20]  // 分析周期
    
    // 执行分析
    result = alphalensAnalysis(alphalensConfig)
    // 输出:IC 分析、分层回测、换手率等
''')

s.close()

9. 实时资金流计算

import dolphindb as ddb

s = ddb.session()
s.connect(host="localhost", port=8848, userid="admin", password="123456")

# 实时计算分钟资金流
s.run('''
    def calcCapitalFlow(tickData){
        // 定义主买主卖
        update tickData set 
            isBuy = iif(bsFlag=='B' or (price>prev(price)), 1, 0),
            isSell = iif(bsFlag=='S' or (price\x3Cprev(price)), 1, 0)
        
        // 计算资金流
        return select 
            sym,
            sum(iif(isBuy, price*volume, 0)) as buyFlow,
            sum(iif(isSell, price*volume, 0)) as sellFlow,
            sum(price*volume) as totalFlow
        from tickData
        group by sym, time_bar(60000, time) as minute
    }
''')

s.close()

金融数据模型

统一数据模型(Instrument & MarketData)

import dolphindb as ddb

s = ddb.session()
s.connect(host="localhost", port=8848, userid="admin", password="123456")

# 证券主表
s.run('''
    instrument = table(
        `AAPL`GOOG`MSFT as sym,
        `Apple Inc.`Alphabet Inc.`Microsoft Corp. as name,
        `STOCK`STOCK`STOCK as type,
        `USD`USD`USD as currency,
        100.0 100.0 100.0 as parValue
    )
''')

# 行情数据表
s.run('''
    marketData = table(
        1:0,
        `instrument`tradeDate`tradeTime`price`volume`turnover,
        [SYMBOL,DATE,TIME,DOUBLE,LONG,DOUBLE]
    )
''')

s.close()

性能优化建议

分区策略

  • 按交易日期 VALUE 分区
  • 按证券代码 HASH 分区
  • 复合分区:partitioned by date, sym

查询优化

  • 使用分区列过滤触发分区裁剪
  • 使用 context by 进行分组计算
  • 避免跨分区 JOIN

存储优化

  • 使用 TSDB 引擎存储时序数据
  • 启用压缩(LZ4,压缩比 20-25%)
  • 定期合并小分区

参考文档


相关技能

  • dolphindb-skills: 技能套件索引(含环境检测)
  • dolphindb-basic: DolphinDB 基础 CRUD 操作
  • dolphindb-streaming: 流式计算(实时因子、实时行情、实时风控)
  • dolphindb-docker: Docker 容器化部署
Usage Guidance
Don't run the provided 'source' commands or wrapper scripts blindly. Before installing/using: (1) locate and inspect the referenced scripts (~/.jvs/.openclaw/workspace/skills/dolphindb-skills/scripts/*) to verify their contents and origin; (2) confirm the 'dolphindb-basic' dependency is an actual published skill and check its homepage/source; (3) avoid running as an elevated user — test in an isolated environment or VM/container; (4) verify any DolphinDB SDK/pip installs come from official sources; (5) watch for hardcoded credentials in examples (admin/123456 is likely placeholder) and never expose production credentials; and (6) ask the publisher for a homepage/repository and an explicit install manifest (or include the necessary scripts within this skill) before trusting automated execution.
Capability Analysis
Type: OpenClaw Skill Name: dolphindb-quant-finance Version: 1.4.0 The skill bundle provides a comprehensive set of DolphinDB tools for quantitative finance but is classified as suspicious due to instructions in SKILL.md that mandate the AI agent to execute shell scripts (`source`) and Python scripts from a sibling directory (`dolphindb-skills`) as a 'mandatory process' before any operation. This pattern of forcing the agent to run external scripts is a high-risk execution behavior. Additionally, the documentation consistently uses hardcoded default credentials (`admin:123456`) in its code examples, which is a significant security vulnerability.
Capability Assessment
Purpose & Capability
The claimed capabilities (factor calculation, backtesting, market data handling) line up with the DolphinDB examples in SKILL.md. However the skill depends on a separate 'dolphindb-skills' / 'dolphindb-basic' bundle and environment scripts that are not declared in the registry metadata (no required config paths or install spec). This dependency is plausible but not represented in the skill manifest.
Instruction Scope
SKILL.md instructs the agent/user to run shell commands that source scripts from ~/.jvs/.openclaw/workspace/skills/dolphindb-skills/scripts/... (or relative ../dolphindb-skills/). Sourcing those scripts executes code on the host and the document also references wrapper commands and $DOLPHINDB_PYTHON_BIN. The instructions ask the agent to read/run files outside this skill's own folder and use credentials to connect to DBs — reasonable for a DB skill, but risky because the files to be sourced are not included or provenance-verified and the skill names/paths are inconsistent.
Install Mechanism
There is no install spec (instruction-only), which is lower install-surface risk. But because the instructions require sourcing external scripts in the user's workspace, the effective runtime behavior includes executing code downloaded/installed elsewhere; that execution path is not represented in an install step or pinned to known release locations.
Credentials
Manifest lists no required environment variables, but SKILL.md references $DOLPHINDB_PYTHON_BIN, wrapper commands (dolphin_python/dolphin_pip), and shows database userid/password usage in examples. The skill does not declare these env vars or the dependency on the other skill that would provide them — omission reduces transparency and could hide credential access.
Persistence & Privilege
The skill is not always: true and does not declare persistent modifications. However the runtime instructions direct sourcing scripts that may modify the environment or create persistent wrappers; because those scripts live outside this skill and are not present in the manifest, their effects are unknown.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install dolphindb-quant-finance
  3. After installation, invoke the skill by name or use /dolphindb-quant-finance
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.4.0
v1.4.0:添加强制环境检测流程,支持全局包装器调用
v1.1.3
环境检测脚本路径修复,添加 dolphindb-basic 依赖提示
v1.1.2
- 增加了 Python 环境检测和 DolphinDB SDK 前置依赖说明,提供环境验证和统一调用接口示例。 - 所有 DolphinDB 脚本和用例均替换为 Python SDK 方式调用,便于直接在 Python 环境集成。 - 完善了参数配置和数据处理流程,使用更结构化的 Python + DolphinDB 演示。 - 保持原有金融场景能力描述和案例覆盖,结构未做大调整。
v1.1.1
v1.1.1 - 集成环境自动检测
v1.1.0
Fix: Replace local file paths with official DolphinDB documentation links
v1.0.0
Initial release: Quant finance skills with factor calculation, backtesting, and performance attribution
Metadata
Slug dolphindb-quant-finance
Version 1.4.0
License MIT-0
All-time Installs 2
Active Installs 2
Total Versions 6
Frequently Asked Questions

What is DolphinDB 量化金融技能?

基于 DolphinDB 提供因子计算、策略回测、行情处理、绩效归因及投资组合优化的量化金融专业解决方案。 It is an AI Agent Skill for Claude Code / OpenClaw, with 239 downloads so far.

How do I install DolphinDB 量化金融技能?

Run "/install dolphindb-quant-finance" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is DolphinDB 量化金融技能 free?

Yes, DolphinDB 量化金融技能 is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does DolphinDB 量化金融技能 support?

DolphinDB 量化金融技能 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created DolphinDB 量化金融技能?

It is built and maintained by superStupidBear (@ugpoor); the current version is v1.4.0.

💬 Comments