Fix df_with_chain2array() silently recycling data with unequal chain lengths

[utkarshpawade]

Fixes #498

Summary

• Validate that all chains have the same number of iterations before constructing the 3-D array
• Replace ceiling(nrow(a) / n_chain) with the validated per-chain count
• Add regression test with unequal chain lengths

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.70%. Comparing base (d15d028) to head (d1b7c77).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #499   +/-   ##
=======================================
  Coverage   98.70%   98.70%           
=======================================
  Files          35       35           
  Lines        5857     5864    +7     
=======================================
+ Hits         5781     5788    +7     
  Misses         76       76           

☔ 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.

[jgabry]

Thanks. I think this is good but the check and error should happen in validate_df_with_chain instead of df_with_chain2array. That's more consistent with how we handle checks for other input types.

[jgabry]

Thanks, looks good! I just added a couple more things to make this a little more robust. If all tests pass we can merge.