[Fix] Try to fix data pipeline scheduled test (attempt 2)

[sxjscience]

Description

It seems that the test is still failing so I'm submitting this PR.

Checklist

Essentials

• PR's title starts with a category (e.g. [BUGFIX], [MODEL], [TUTORIAL], [FEATURE], [DOC], etc)
• Changes are complete (i.e. I finished coding on this PR)
• All changes have test coverage
• Code is well-documented

Changes

• Feature1, tests, (and when applicable, API doc)
• Feature2, tests, (and when applicable, API doc)

Comments

• If this change is a backward incompatible change, why must this change be made.
• Interesting edge cases to note here

cc @dmlc/gluon-nlp-team

[codecov]

Codecov Report

Merging #1483 (7cbf74b) into master (52da8ab) will decrease coverage by 0.20%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #1483      +/-   ##
==========================================
- Coverage   85.87%   85.67%   -0.21%     
==========================================
  Files          52       52              
  Lines        6909     6909              
==========================================
- Hits         5933     5919      -14     
- Misses        976      990      +14     

Impacted Files	Coverage Δ
src/gluonnlp/data/loading.py	78.11% <0.00%> (-5.29%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 52da8ab...7cbf74b. Read the comment docs.

[github-actions]

The documentation website for preview: http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR1483/fix_data_pipeline_test2/index.html

[szha]

What's the error? Is there a way to validate the fix short of committing?

[sxjscience]

https://github.com/dmlc/gluon-nlp/runs/1710545425?check_suite_focus=true

[szha]

The log doesn't seem to show where the encoding error happens so it's a bit unclear whether the patch is the right fix.

[sxjscience]

I'm also not sure. But currently do not have enough time to take a look. My guess is this python encoding issue.