Data science environments, for collaboration.
conda-store is an open source tool for managing data science environments in collaborative teams. It provides flexible, yet reproducible, environments while enforcing best practices throughout your environment’s life cycle.
Create and update environments quickly using a graphical UI or a YAML editor.
All environments are automatically version-controlled, and available for use.
Share environments effortlessly through the auto-generated artifacts: lockfile, docker image, YAML file, or tarball.
Exact versions of all packages and their dependencies are pinned in these artifacts.
Access admin-approved packages and channels, and request new ones when needed.
Admins have role-based access management, to allow users to share environments across (and only with) their team.
The conda packaging system works well for individual users, but it can get tricky to reproduce environments reliably to share with colleagues. On the other hand, Docker containers can ensure reproducibility, but aren’t friendly to individual users who may need new packages or versions for daily work. conda-store resolves this friction, and provides a balance of both the familiarity of conda and the robustness of containers.