BNL, Upton, New York, USA
Python provides broad design freedom to programmers and a low barrier of entry for new software developers. These aspects have proven that unless standardized, a Python codebase will tend to diverge from a common style and architecture, becoming unmaintainable across the scope of a large controls system. Mitigating these effects requires a set of tools, standards, and procedures developed to assert boundaries on certain aspects of Python development – namely project organization, version management, and deployment procedures. Common tools like Git, GitLab, and virtual environments form a basis for development, with in-house utilities presenting their capabilities in a clear, developer-focused way. This paper describes the necessary constraints needed for development and deployment of large-scale Python applications, the function of the tools which comprise the development environment, and how these tools are leveraged to create simple and effective procedures to guide development.
