GH-1289: Partial fix for data access

[kschelonka]

There is currently a TODO to unit-test this function, which has multiple undocumented dependencies. I think that's beyond the scope of this fix, but I was able to confirm that this works by getting the stimulus presentations via the Session object for a previously failed ID.

[codecov-io]

Codecov Report

Merging #1347 into rc/1.5.0 will not change coverage.
The diff coverage is 0%.

@@            Coverage Diff            @@
##           rc/1.5.0    #1347   +/-   ##
=========================================
  Coverage     34.35%   34.35%           
=========================================
  Files           340      340           
  Lines         33004    33004           
=========================================
  Hits          11340    11340           
  Misses        21664    21664

Impacted Files	Coverage Δ
.../brain_observatory/behavior/stimulus_processing.py	27.04% <0%> (ø)	⬆️

---

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 05bdbb6...bc8fd2d. Read the comment docs.

[njmei]

This seems fine to me, I would double check that the start frames always end with *.0. There should be another issue created to also address this problem upstream since start_frame and end_frame shouldn't even be float in the first place...

[kschelonka]

Thanks Nick, I've added an issue to the visual behavior epic to check on this data source. This function and it's dependencies are unfortunately undocumented and untested. I'd like to dig into the requirements in the future.