Using Docker to Reproduce and Run pysystemtrade Workflows

Summary

The document explains how Docker can package a Python trading research environment with specific library versions and project code. The motivation is reproducibility: a legacy system may depend on older software versions, while its host machine still needs operating system tools, network drives, data, and brokerage software. The described workflow builds or obtains an image, runs a container with a host directory mounted for results, executes a futures-system calculation, and saves cached output for later inspection.

It also discusses possible uses beyond a local example, including backtesting servers, clusters, cloud computing, production microservices, data placement, messaging, scheduling, security, and release practices. The example reports querying a portfolio Sharpe ratio but gives no resulting value or comparison. The text is an infrastructure overview, not evidence about strategy performance; operational details such as dependency versions, data management, security controls, and deployment behavior would need to be assessed for a particular system.

Key ideas

• Docker images can package Python dependencies and trading-system code for repeatable runs.
• A container can mount a host directory to preserve research outputs outside the image.
• Cached calculations can be saved and reused when inspecting results.
• The document identifies backtesting infrastructure and production services as possible Docker applications.
• The example does not report strategy results or establish the quality of the trading system.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.