Locker: A User-Friendly Tool to Run and Manage Interactive Docker Containers Supporting Reproducible Research

Reproducible computational analysis requires careful tracking and retention of all relevant artifacts, including code, input data, results, run logs, and files that define the computational environment. A Docker image is an excellent method of encapsulating an environment that can be both run as a container to perform analyses and easily archived for future use and reproduction of the analyses if necessary. However, many users would benefit from a simple, intuitive way to develop code and run analyses using Docker images, with a mechanism that includes strong safeguards against accidental loss of work. We have created and publicly released “Locker” to meet this need. Locker provides a web-based graphical user interface (GUI) for users to easily start, use, and stop containers with common data science IDEs such as RStudio and Jupyter. Locker can be used on a local computer or on a remote instance. To facilitate Locker’s use on remote machines, we have included “Locker Services”, an intuitive, web-based portal that can create and manage Amazon Elastic Compute Cloud (EC2) instances pre-configured for running Locker. Locker and Locker Services provide a flexible, easy-to-use interface for performing reproducible computational analyses locally or in the cloud.