Code and data on Code Ocean have been open access from day one, and a key component of that commitment is the ability to export code and data for use elsewhere. We are excited to announce that you can now export a compute capsule as a whole, computational environment included. This new functionality will enable researchers to take published and private compute capsules out of Code Ocean and run the code on machines of their choosing, to reproduce published results or build on top of them.
On Code Ocean, authors retain ownership of their code and data and can specify whatever reuse license they choose. This means that they can export compute capsules from Code Ocean and freely share or store them in alternative repositories to their liking. To facilitate this, Code Ocean’s curated set of base environments is now publicly available on Code Ocean’s public docker registry and can be accessed by following the instructions provided in the exported archive.
The exported capsule archive will contain the following:
- Code and (optionally) data folders;
- A metadata folder with a file in YAML format containing the complete capsule metadata;
- An environment folder containing a recipe for building the capsule’s computational environment, in the form of a Dockerfile and any supporting files required.
To reproduce the results locally, follow the instructions in REPRODUCING.md, found at the root level (or see this knowledge base article). These will walk you through installing the necessary prerequisites (such as Docker), downloading additional public datasets (if applicable), fetching the Docker image or building it locally, and finally using the image to execute the code within a Docker container. Base images containing commercial software such as MATLAB and Stata will require the user to provide a license (whether a pre-existing personal license or a trial license) and mount it into the container during runtime.
We would like to take this opportunity to commit to an open access compute capsule pledge. Published compute capsules will always remain open access. Our intention is to build a sustainable platform for researchers and their institutions to make their work reproducible and reusable. We hope to leave a legacy for generations to come, serving the scientific and research community in a fair and enduring manner.
We’re open and receptive to comments and feedback, and will adjust accordingly to make things clearer or make improvements to this functionality. You can write us privately at email@example.com, comment below, or tweet to us @codeoceanHQ.
Originally published at codeocean.com.