SUBMARINE-1389. restrict scipy < 1.11.0 for python 3.9 and tensorflow 2.6

[cdmikechen]

What is this PR for?

From scipy 1.11.0 (https://github.com/scipy/scipy/releases/tag/v1.11.0), scipy need numpy 1.21.6 or geater in python 3.9, so that tensorflow 2.6.5 CI will not succeed. Link: https://github.com/apache/submarine/actions/runs/5486277299/jobs/9996172343

This is because in tensorflow 2.6.5 dependency list, the dependency version of numpy is 1.19 (required: ~=1.19.2), so that we need restrict scipy < 1.11.0.

Also, we will remove support for python 3.7 and TensorFlow 2.6 in subsequent releases (issue: https://issues.apache.org/jira/browse/SUBMARINE-1391). The scipy restrictions may be removed at that time.

What type of PR is it?

Hot Fix

Todos

• [x] - scipy < 1.11.0

What is the Jira issue?

https://issues.apache.org/jira/browse/SUBMARINE-1389

How should this be tested?

Python CI test

Screenshots (if appropriate)

NA

Questions:

• Do the license files need updating? No
• Are there breaking changes for older versions? No
• Does this need new documentation? No

[codecov[bot]]

Codecov Report

Merging #1082 (c0fc99f) into master (f7c2fe5) will not change coverage. The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #1082   +/-   ##
=======================================
  Coverage   67.27%   67.27%           
=======================================
  Files         128      128           
  Lines        6224     6224           
=======================================
  Hits         4187     4187           
  Misses       2037     2037           

Flag	Coverage Δ
python-integration	54.20% <ø> (ø)
python-unit	48.00% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

:mega: We’re building smart automated test selection to slash your CI/CD build times. Learn more