Phandas

<div align="center"> <img src="https://raw.githubusercontent.com/quantbai/phandas/main/assets/PHANDAS2.png" alt="Phandas" width="500">

 

</div>

English

A multi-factor quantitative trading framework for cryptocurrency markets.

Overview

Phandas is a streamlined toolkit for alpha factor research and backtesting in cryptocurrency markets. Design factors with 60+ operators, test with dollar-neutral backtesting, and analyze with professional metrics.

Try it now

Web Demo - Experience Phandas directly in your browser. No installation required.

Key Features

• Data Fetching: Multi-source OHLCV data (Binance, OKX)
• Factor Engine: 60+ time-series and cross-sectional operators
• Neutralization: Vector projection & regression-based orthogonalization
• Backtesting: Dollar-neutral strategies with full/partial rebalancing
• Performance Metrics: Sharpe, Sortino, Calmar, Max Drawdown, VaR, PSR
• Factor Analysis: IC, IR, correlation, coverage, turnover
• MCP Integration: AI agents (Claude) can directly access Phandas

Installation

pip install phandas

Quick Start

from phandas import *

# Fetch market data
panel = fetch_data(
    symbols=['ETH', 'SOL', 'ARB', 'OP', 'POL', 'SUI'],
    timeframe='1d',
    start_date='2023-01-01',
    sources=['binance'],
)

# Extract factors
close = panel['close']
volume = panel['volume']
open = panel['open']

# Construct momentum factor
momentum_20 = (close / close.ts_delay(20)) - 1

# Neutralize against volume
factor = vector_neut(rank(momentum_20), rank(-volume))

# Backtest strategy
result = backtest(
    entry_price_factor=open, 
    strategy_factor=factor,
    transaction_cost=(0.0003, 0.0003)
)

result.plot_equity()

AI Integration via MCP

Use Phandas with AI IDEs (Cursor, Claude Desktop) directly—no coding required.

Setup for Cursor (Recommended)

1. pip install phandas
2. Open Cursor → Settings → Tools & MCP → New MCP Server
3. Paste the JSON config below, save and restart

{
  "mcpServers": {
    "phandas": {
      "command": "python",
      "args": ["-m", "phandas.mcp_server"]
    }
  }
}

Available Tools (4 Functions)

• fetch_market_data: Get OHLCV data for symbols
• list_operators: Browse all 50+ factor operators
• read_source: View source code of any function
• execute_factor_backtest: Backtest custom factor expressions

---

繁體中文

一個專為加密貨幣市場設計的多因子量化交易框架。

概述

Phandas 是一個精簡的加密貨幣因子研究與回測工具。提供 60+ 運算子設計因子、美元中性回測、專業績效指標分析。

立即體驗

網頁演示 - 直接在瀏覽器中體驗 Phandas，無需安裝。

核心功能

• 資料獲取：多源 OHLCV 資料（Binance、OKX）
• 因子引擎：60+ 時間序列與橫截面運算子
• 因子中性化：向量投影與迴歸正交化
• 回測引擎：美元中性策略、全/部分調倉
• 績效指標：夏普比、Sortino、Calmar、最大回撤、VaR、PSR
• 因子分析：IC、IR、相關性、覆蓋率、換手率
• MCP 集成：AI 代理（Claude）可直接調用 Phandas

安裝

pip install phandas

快速開始

from phandas import *

# 獲取市場資料
panel = fetch_data(
    symbols=['ETH', 'SOL', 'ARB', 'OP', 'POL', 'SUI'],
    timeframe='1d',
    start_date='2023-01-01',
    sources=['binance'],
)

# 提取因子
close = panel['close']
volume = panel['volume']
open = panel['open']

# 構建動量因子
momentum_20 = (close / close.ts_delay(20)) - 1

# 對成交量進行中性化
factor = vector_neut(rank(momentum_20), rank(-volume))

# 回測策略
result = backtest(
    entry_price_factor=open, 
    strategy_factor=factor,
    transaction_cost=(0.0003, 0.0003)
)

result.plot_equity()

AI 集成（MCP 支援）

在 AI IDE（Cursor、Claude Desktop）中直接使用 Phandas—無需編碼。

Cursor 設定（推薦）

1. pip install phandas
2. 開啟 Cursor → Settings → Tools & MCP → New MCP Server
3. 貼上下方 JSON 配置，儲存並重啟

{
  "mcpServers": {
    "phandas": {
      "command": "python",
      "args": ["-m", "phandas.mcp_server"]
    }
  }
}

可用工具（4 個函數）

• fetch_market_data: 獲取代幣 OHLCV 資料
• list_operators: 瀏覽 50+ 因子運算子
• read_source: 查看任何函數的源代碼
• execute_factor_backtest: 回測自訂因子表達式

---

Documentation | 文檔

• Full Docs - Complete API reference
• Operators Guide - 50+ operators
• MCP Setup - AI IDE integration

---

Community & Support | 社群與支持

• Discord: Join us - Phantom Management
• GitHub Issues: Report bugs or request features

License

This project is licensed under the BSD 3-Clause License - see LICENSE file for details.