Reshuffle the data after reading from cache if shuffle_rows is true

[kashishmittal55]

If shuffle_rows is true, we do want to shuffle the rows after reading it from cache.
In order to do so, once we read the data, we can reshuffle it. This means, the first time when we are warming up the cache, we will be doing double shuffling.

[codecov]

Codecov Report

❌ Patch coverage is 92.85714% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 86.44%. Comparing base (b6fbf92) to head (740c524).
⚠️ Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
petastorm/arrow_reader_worker.py	92.85%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #817      +/-   ##
==========================================
+ Coverage   86.39%   86.44%   +0.05%     
==========================================
  Files          84       84              
  Lines        5144     5158      +14     
  Branches      808      813       +5     
==========================================
+ Hits         4444     4459      +15     
+ Misses        558      557       -1     
  Partials      142      142              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[arushi297]

@kashishmittal55 will the current reproducibility/randomness behavior remain intact with this change?

[kashishmittal55]

@kashishmittal55 will the current reproducibility/randomness behavior remain intact with this change?

Yes correct, it should not affect reproducibility since we will use the same random variable for different training runs.