Closed corbennett closed 1 year ago
We talked a bit about this issue with @matchings at one of our backlog refinement sessions last week.
I think what would be helpful here would be see an example DataFrame of what you want the columns to look like. You can use the data of the stimulus tables I sent over as an example. As in this is before, and this is after and where I am defining the values for these columns from. Sound good @corbennett?
After a few discussions within Pika, we can change the value of the stimulus_name
column to match that of the already released ecephys data. Since re-doing the ecephys data is out of scope for this release, we won't be able to change the stimulus_block
column as this time as it would require changing that data.
Describe the bug The stimulus presentations tables for the VBN ephys and behavior-only sessions don't agree on what to represent with the
stimulus_name
column. VBN ephys lists the name of the actual stimulus (eg 'Natural_Images_Lum_Matched_set_ophys_G_2019'), while the behavior-only session indicates the type of stimulus block ('behavior').As described in https://github.com/AllenInstitute/AllenSDK/issues/2527, the ideal solution would be for the
stimulus_name
column to indicate the stimulus (eg 'Natural_Images_Lum_Matched_set_ophys_G_2019' as for the ephy sessions), while thestimulus_block
column should explicitly describe the stimulus block ('behavior' in this case), rather than an integer that requires a look-up table to interpret.This likely applies to the VBO data as well. @matchings can confirm this.