by Edward Easton for
This talk aims to discuss problems and solutions when implementing an end-to-end Python environment for a big company.
The term ‘Enterprise’ often has negative connotations in much of the open-source world, usually along the lines of ‘expensive and bloated’. In this case I’d like to use it for describing a systems environment where the freedoms of a developer working at home with her laptop are not always available or practical. For example, coding on the holodeck of the Starship Enterprise.
Big teams and codebases need tools that make it very easy for new developers to get up and running writing code with the minimum of fuss. Large software platforms needs tools to manage complex package dependency graphs and release workflows. Also, production environments don’t always have compilers or direct internet access so many of Python’s standard packaging tools make it a bit tricky to get things going in these regards.
I’d like to present how we managed to solve a number of these problems for a recent client, a task which happily culminated in much of the software being open-sourced at https://github.com/eeaston/pkglib
Edit: The software also has numerous testing tools included which I’ll demonstrate, including py.test fixtures for mongo, redis, pyramid and generic servers, selenium webdriver and C/C++ code coverage and profiling.
