Less strict docker dependency versions

[jcugat]

TL;DR

Change the minimum required version of the docker dependency from v5.* to v4.*

Type

• Bug Fix
• Feature
• Plugin

Are all requirements met?

• Code completed
• Smoke tested
• Unit tests added
• Code documentation added
• Any pending items have an associated Issue

Complete description

We need it because we use Flyte alongside other packages which force a docker version in the v4.* range (specifically tensorflow's tfx package).

I have tried running the tests locally while force-installing docker docker==4.0.2 and it worked without any errors, so there's no breaking change between those versions that would affect this project.

Also it would be great if this change was backported to the Flyte v1.2.* and v1.3.* branches, thanks!

Tracking Issue

NA

Follow-up issue

NA

[welcome]

Thank you for opening this pull request! 🙌

These tips will help get your PR across the finish line:

• Most of the repos have a PR template; if not, fill it out to the best of your knowledge.
• Sign off your commits (Reference: DCO Guide).

[codecov]

Codecov Report

Merging #1536 (ee52e42) into master (28da983) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #1536   +/-   ##
=======================================
  Coverage   69.70%   69.70%           
=======================================
  Files         310      310           
  Lines       28721    28721           
  Branches     2721     2721           
=======================================
  Hits        20019    20019           
  Misses       8194     8194           
  Partials      508      508           

Impacted Files	Coverage Δ
setup.py	0.00% <ø> (ø)

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[pingsutw]

@jcugat I'm just curious how do you integrate tfx with flyte. are you compiling tfx workflow to flyte workflow?

[jcugat]

@pingsutw Not exactly, we use Flyte to launch Kubeflow Pipelines, which uses TFX under the hood. There's some more details in this blog post: https://engineering.atspotify.com/2019/12/the-winding-road-to-better-machine-learning-infrastructure-through-tensorflow-extended-and-kubeflow/

[welcome]

Congrats on merging your first pull request! 🎉

[jcugat]

Thanks for merging @pingsutw!

Would it be possible to also backport this to the v1.2 and v1.3 releases? Should I do anything to help that happen?

[jcugat]

@pingsutw I've opened the following PRs to backport this fix. Hopefully it's the correct approach:

• v1.3 backport: [v1.3 Backport] Less strict docker dependency versions (#1536) #1563)
• v1.2 backport: [v1.2 Backport] Less strict docker dependency versions (#1536) #1564