Merge dev into master

[dougollerenshaw]

Summary of changes in this PR:

• Adds loading functionality for updated event detection files
• Moves some columns from extended_stimulus_presentations to stimulus_presentations, including epoch times, licks, rewards, reward_rate, response_latency, etc.
• adds convenience functions including get_container_response_df,
• adds plots and related functions for cell matching validation
• adjust events filtering function to account for different frame rates
• adds columns to experiments_table indicating how many times prior to the current session that the mouse was exposed to a given image set or to stimulus omissions
• adds mean pupil area for each flash to stimulus_table and downstream dfs (ex: stimulus_response_df)

To Do:

• Make columns in extended_stimulus_presentations independent of behavior model output files (especially 'engagement_state')
• Remove / update classifier validation plots in QC visualizations

[dougollerenshaw]

@matchings: note that the allensdk_head tests are failing because of the path changes noted in issue #696. I'll work on that issue in a separate PR, but if you get this PR ready to merge before that issue is solved, I think we should go ahead and merge with the one failing test.

[matchings]

@dougollerenshaw im still about halfway through addressing all the comments in the PR (primarily adding documentation), and will let you know when its ready to merge. Sorry its taking me so long.

[matchings]

@dougollerenshaw I think this PR is ready to merge in, if we are ok with the known test failure associated with #696. Should I go ahead and do that, or do you want to do another once over to double check for anything that could break downstream analysis code?

[dougollerenshaw]

Let's merge in #701 first. That should take care of the failing tests on test-allendsdk_head