Please enable JavaScript.
Coggle requires JavaScript to display documents.
Datasciences environement cross-platform with Docker (Context -…
Datasciences environement cross-platform with Docker
Context - Introduction
Problems
jupyter-lab with docker we doesn't find persistence system
We work with different Os (Mac and linux)
We want to use use the Conda env because of the recent studie that prove this intallation optimize computer calculation.
source studies
Why use Docker to containerised ?
The solution
bind folder between host and container
bind user between host and container ( beceause if not you can't modify the host from the container)
Don't reinvent the wheel (Docker images exist with the big part of the works already done, use it)
Conclusion
Why use this solution
Local solution for multi-platform team
Easy run to test Gist of code
Alternatives
Deploy Jupyter on a server with connection tiers ( ldap, token, etc..)
directly on the host (but all team need the same environement like packages and dependencies, etc..)
Good solution easy to run we use it for all our personal datascience projects.