Open
Milestone
Enhance deployability and packaging
Milestone ID: 3401
As a IT team deploying Fed-BioMed, I want to deploy and configure the software more easily.
(implicit: divide the "user" role between (1) the IT team deploying configuring and maintaining a Fed-BioMed instance, and (2) the end users on the researcher side + node side using it to make data + models available and conduct experiments with them).
Phase 1:
-
CI slave(s) (update, add slaves for new OS) for better testing before issuing releasesee #531 (closed) #532 #533 -
#provide backward compatibilty or awareness when upgrading Fed-BioMeddone as #506 (closed) #534 (closed) #535 (closed) - #448 revise versions of packages included
- #266 VPN/container env script and process enhance
- misc robustify:
Phase 2:
- modular deployment (eg: by function: scikit-learn, pytorch, secagg, flamby, etc.; by component)
- packaging: pip, conda or other, etc.
- write and support deployement scenario (
dev
,vpn
, maybe others)