Since 2017, Data Central has evolved from a single monolithic application to a science platform. A significant part of this evolution has been due to large structural changes in the core web service known internally as "dcapi", written in django, with djangorestframework and react being key supporting components. These changes include the introduction of private data releases (implying a significant change in access control and how it is tracked internally, including the merging of databases), the move from hadoop to postgresql, and the splitting out of non-survey-based components (including the telescope archive, user registration and management, and user support). In this talk, we will discuss the technical reasons behind these changes, and reflect on how choices at the start of the project still inform the choices we make now about future directions of this service.

In this focus demo, I will give a short introduction to the hypothesis Python library (https://hypothesis.readthedocs.io/), which provides a property-based testing framework that integrates into the existing Python testing frameworks of pytest and unittest. I'll provide some examples of how I've used hypothesis in the past, and show how effective it is at finding edge and corner cases in your code.