Investigate data access for MultiscopeSignalNoise data

[wbwakeman]

For internal analysts to access passive mesoscope data for the MultiscopeSignalNoise project, they need to load a session object from LIMS.

This issue is to investigate the needs an AllenSDK object with that data.

Work with scientists (Ahad and Michael Buice) to determine if ophys session with no behavior data attached is what is needed, and what effort has already been done towards this.

Put forward a plan for creating this new data access object.

This is what Michael is running that is failing for him.
'''
from allensdk.brain_observatory.behavior.behavior_project_cache import VisualBehaviorOphysProjectCache
cache = VisualBehaviorOphysProjectCache.from_lims()
experiment_id = 1033674237
dataset = cache.get_behavior_ophys_experiment(experiment_id)
'''

[danielsf]

If this is at all helpful, this code

import allensdk
print(allensdk.__version__)

from allensdk.brain_observatory.behavior.behavior_ophys_experiment import Behavi
orOphysExperiment
exp_id=1033674237
exp = BehaviorOphysExperiment.from_lims(ophys_experiment_id=exp_id)
print(exp.dff_traces)

runs on v2.12.0 of the SDK but does not run on v2.13.0 (there was some question during stand up as to whether or not this would be the case)

[danielsf]

I just spoke with Michael Buice.

In order to get himself unstuck, he is going to copy-and-paste BehaviorOphysExperiment into a class VisualCodingOphysExperiment and make the changes he needs so that VisualCodingOphysExperiment.from_lims() will just work.

We did not go into details, but it sounds like the changes will involve

• removing the dependence on BehaviorSession
• calculating the frame rate from the sync file, rather than from the pickle file
• adding a function he wrote to construct the stimulus table

When he is unstuck, he will issue a PR from his branch. Probably we won't want to merge it, but we will want to copy-and-paste/git cherry-pick his code into our final, productionized solution.