test(artifacts): replace sleeps with flush when waiting on a file to write

[moredatarequired]

Description

Replace flaky calls to time.sleep(0.1) with calls to flush when waiting on a file to write.

For some reason on my machine the calls to time.sleep often do nothing. Regardless, this is a faster and more reliable way to ensure files are written with sequential c/m/a times.

Testing

[test only PR, no behavior change]

Checklist

• Include reference to internal ticket "Fixes WB-NNNN" and/or GitHub issue "Fixes #NNNN" (if applicable)
• Ensure PR title compliance with the conventional commits standards

[codecov]

Codecov Report

Merging #4523 (e7c5aea) into main (d911344) will increase coverage by 0.03%.
The diff coverage is n/a.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4523      +/-   ##
==========================================
+ Coverage   83.08%   83.12%   +0.03%     
==========================================
  Files         259      259              
  Lines       32979    32979              
==========================================
+ Hits        27401    27414      +13     
+ Misses       5578     5565      -13     

Flag	Coverage Δ
functest	56.95% <ø> (+0.21%)	⬆️
unittest	73.30% <ø> (-0.02%)	⬇️

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

Impacted Files	Coverage Δ
wandb/sdk/internal/internal_api.py	89.17% <0.00%> (-0.52%)	⬇️
wandb/sdk/wandb_setup.py	88.44% <0.00%> (-0.51%)	⬇️
wandb/sdk/lib/git.py	79.01% <0.00%> (ø)
wandb/sdk/wandb_run.py	90.72% <0.00%> (+0.05%)	⬆️
wandb/sdk/wandb_settings.py	94.73% <0.00%> (+0.27%)	⬆️
wandb/sdk/interface/artifacts.py	79.03% <0.00%> (+0.34%)	⬆️
wandb/sdk/lib/printer.py	89.24% <0.00%> (+0.63%)	⬆️
wandb/sdk/internal/file_stream.py	89.86% <0.00%> (+1.01%)	⬆️
wandb/sdk/lib/sock_client.py	94.17% <0.00%> (+1.05%)	⬆️
wandb/sdk/internal/system/system_info.py	91.55% <0.00%> (+3.24%)	⬆️
... and 1 more

[kptkin]

Nice! Thanks for the fix!
getting away from time based tests is def an improvement :)

[vwrj]

lgtm!

Nit: from browsing online, looks like the os.fsync method guarantees writes to actual disk, I'm seeing a few folks recommend os.flush() followed by the fysnc.

[moredatarequired]

lgtm!

Nit: from browsing online, looks like the os.fsync method guarantees writes to actual disk, I'm seeing a few folks recommend os.flush() followed by the fysnc.

Ha, I mentioned that in the first PR where I had to add it. I suppose I can be rigorous here instead of stopping once I have it working.