BMTK workshop setup

[wbwakeman]

Anton Arkhipov and Kael Dai are planning a 2-day BMTK/SONATA modeling workshop in July 2022. There will be 20 - 30 participants and the expected dates are either July 21-22 or July 28-29. The infrastructure only needs to be up for the two days of the workshop (really?)

Need to set up something like we do for the SWDB Friday Harbor workshop. Recall the high-level checklist to help prepare for this. Anton has been notified to give us lots of advance notice about the data sets they intend to use.

• ready for test walkthrough by June 30, 2022

Instructions are in the repo at https://github.com/AllenInstitute/swdb_portal

Identify the repo that will contain notebooks for the workshop (e.g. swdb was at https://github.com/AllenInstitute/swdb_2021 )

Tasks:

• On or before August 6, ping Anton and Kael regarding the notebooks to be part of the course, especially about any new dependencies

• Customize AllenSDK Sagemaker conda environment to install additional packages specified in e-mail from Michael Buice.

  • The script that needs to be modified can be found at: https://github.com/AllenInstitute/swdb_portal/blob/d9c38c88d70b54897b68607c522c18981684b284/swdb_portal/scripts/lifecycle_config_on_create.sh#L37

• Retaining this note in case we need another conda environment just for BMTK Python dependencies.

  • See this example in SWDB for how to create another conda environment: https://github.com/AllenInstitute/swdb_portal/blob/d9c38c88d70b54897b68607c522c18981684b284/swdb_portal/scripts/lifecycle_config_on_create.sh#L27

• Update lifecycle script so that it also mounts the Visual Behavior Ophys bucket (?)

  • Currently only the S3 bucket for visual coding ophys/neuropixels is being mounted. See the SWDB startup script for an example of how that is mounted: https://github.com/AllenInstitute/swdb_portal/blob/d9c38c88d70b54897b68607c522c18981684b284/swdb_portal/scripts/lifecycle_config_on_start.sh#L15

• Inform scientists about new CloudCache parameter that must be added

  • See 2167 cloudcache from s3 #2171

• Set up workshop infrastructure

• Verify with course coordinators Anton/Kael that notebooks are ready

Acceptance criteria:

• Run through all the notebooks.
• Work with scientific groups to do a dry run of workshop infrastructure.

[wbwakeman]

Consider moving away from lifecycle scripts and using Docker containers instead. Need to figure out details, and also modify the swdb app to work with this. (1 - 2 sprints?)

[wbwakeman]

https://alleninstitute.org/what-we-do/brain-science/events-training/2022-modeling-workshop/

[danielsf]

This issue is probably redundant with the issues that are going to be opened over the next few months (May-July 2022). I'm leaving it here for documentation purposes.