Fixing #317 Invalid Eigenvalues

[skfegan]

Summary

Resolving Issue #317 by correcting the looping over frames.

Changes

Fixing level_dag.py

• looping over trajectory frames in _run_frame_stage

Tests

• update regression baselines

Impact

• By looping over frames (not just frame index) in level_dag.py, the code is now collecting data from all the relevant frames not just the first one, so the sampling is correct.
• This will reduce the number of invalid (zero, negative, or imaginary) eigenvalues and improve the vibrational entropy results.

[ioanaapapa]

It looks good to me and the results match what I was getting before.
dna_example_output.txt

[ioanaapapa]

It looks good to me and the results match what I was getting before.

[jimboid]

An important bug fix. This looks like the origin was the quirky behaviour of how frames from MDA are dealt with in some types of iterables. We've had a valuable report from @ioanaapapa that the numbers are now indeed functioning correctly as per previous behaviour. I'm wholly satisfied that this PR fixes the issue reported with invalid eigen values.