What is this?
QUANTAXIS is a local-first quantitative trading framework for China's stock, futures and options markets, covering data fetching, backtesting, paper trading, live trading, visualization and multi-account management, with support for distributed task scheduling.
QUANTAXIS is a local-first quantitative trading framework for China's stock, futures and options markets, covering data fetching, backtesting, paper trading, live trading, visualization and multi-account management, with support for distributed task scheduling.Is this a ready-made trading app or a framework I build on?It is a Python framework made of many modules (QAFetch, QAData, QAEngine, QAStrategy, QAWebServer, etc.) that you assemble into your own quant pipeline, not a single turnkey app.
It is a Python framework made of many modules (QAFetch, QAData, QAEngine, QAStrategy, QAWebServer, etc.) that you assemble into your own quant pipeline, not a single turnkey app.How do I get it running?Install the base Python package, then optionally add the Rust (QARS2) components for up to 100x faster accounts and backtests; a full install plus MongoDB/ClickHouse gives you the complete data-to-trading pipeline.
Install the base Python package, then optionally add the Rust (QARS2) components for up to 100x faster accounts and backtests; a full install plus MongoDB/ClickHouse gives you the complete data-to-trading pipeline.What do I need to run it besides Python?You need Python 3.9-3.12 (3.11+ recommended) and MongoDB 4.0+ for data storage; ClickHouse 20.0+ is optional, and the Rust QARS2 core is optional but recommended for the big speedups.
You need Python 3.9-3.12 (3.11+ recommended) and MongoDB 4.0+ for data storage; ClickHouse 20.0+ is optional, and the Rust QARS2 core is optional but recommended for the big speedups.Any catches before I rely on this?It is still a pre-release (v2.1.0-alpha2), the QADataSwap zero-copy data bridge is 'in progress', and v2.0 was an incompatible rewrite that dropped the older QAARP account module entirely.
It is still a pre-release (v2.1.0-alpha2), the QADataSwap zero-copy data bridge is 'in progress', and v2.0 was an incompatible rewrite that dropped the older QAARP account module entirely.Does it only run on Linux servers?No, it runs locally on Linux, macOS or Windows; it's a self-hosted Python install, not a hosted SaaS.
No, it runs locally on Linux, macOS or Windows; it's a self-hosted Python install, not a hosted SaaS.What is QARS2 and why does it matter?QARS2 is a Rust core that QUANTAXIS wraps through QARSBridge, giving up to 100x faster account operations and 10x faster backtests, with automatic fallback to pure Python if it isn't installed.
QARS2 is a Rust core that QUANTAXIS wraps through QARSBridge, giving up to 100x faster account operations and 10x faster backtests, with automatic fallback to pure Python if it isn't installed.What is the QIFI protocol?QIFI is QUANTAXIS' unified account protocol for positions, orders and trades that stays 100% consistent across the Python, Rust and C++ implementations, so you can swap in faster engines without changing your account logic.
QIFI is QUANTAXIS' unified account protocol for positions, orders and trades that stays 100% consistent across the Python, Rust and C++ implementations, so you can swap in faster engines without changing your account logic.Can I actually trade live, or is this backtest-only?It supports live and simulated trading via CTP (futures/options) and QMT (stocks) interfaces, plus parent-child account order management and risk control through an OrderGateway.
It supports live and simulated trading via CTP (futures/options) and QMT (stocks) interfaces, plus parent-child account order management and risk control through an OrderGateway.Why bother with a Rust core instead of staying pure Python?Creating 1000 accounts or sending 10000 orders takes about 50 seconds in pure Python versus roughly 0.5 seconds with the QARS2 Rust core, while also cutting memory use by around 90%, which matters for large multi-account backtests.
Creating 1000 accounts or sending 10000 orders takes about 50 seconds in pure Python versus roughly 0.5 seconds with the QARS2 Rust core, while also cutting memory use by around 90%, which matters for large multi-account backtests.What are similar projects?Vibe-Trading, abu, TradingAgents-CN, ai-hedge-fund, qlib, daily_stock_analysis, QuantDinger, vnpy
Vibe-Trading, abu, TradingAgents-CN, ai-hedge-fund, qlib, daily_stock_analysis, QuantDinger, vnpyWhich topics is it in?message-queue, backtesting, quant, clickhouse, algorithmic-trading, pandas, stock-market, quantitative-finance, trading, mongodb, microservices, finance
message-queue, backtesting, quant, clickhouse, algorithmic-trading, pandas, stock-market, quantitative-finance, trading, mongodb, microservices, finance