Only deploying from linux is working since adding the cryptography package.

I've tried many things, but none worked.

It seems like etting the lambda architecture to arm by setting arch to arm64 in serverless.yml should work, but it doesn't.

• use Flask app.config
• create Cognito user pool
  • verify email to use as sender in Cognito communications (like when sending user invitation)
    - [ ] AWS SSM Parameter Store or AWS Secret Manager
• integrate with Flask app

Some thoughts:

• add db entry for failed job and show pane of upload messages for user (like in Minerva Cloud)
• log error details for later debugging
• give user-friendly message about errors

I tried to redirect inside the promise resolve and reject, but it didn't work. Need to look into doing this correctly, but not sure where we should be going yet, so I'm just making this ticket so I don't forget.

Currently giving all s3 permissions and all we're doing is listing

https://flask-wtf.readthedocs.io/en/0.15.x/csrf/

just a reminder this will need to be done when all the endpoints are in place.

From Jeremy's suggestion:

the Minerva Cloud image import design is to use bulk S3 upload to a staging bucket, then an import task extracts the content into a cloud-friendly form (zarr) in a different bucket and archives the original file in glacier.

He also mentioned that it should be generalized and then make image import a specific instance of that.

Need a unified exception catching scheme and need to send users to endpoints that alert them to specific problems.
It would be nice to also give users suggestions for correcting or at least who to contact to correct each kind of issue.

Each user will be setting min and max values for each color channel, so we'll have to store that.

URLs need to be in links with the filename as the list items.

We'll also need sliders for each color channel for users to set preferences.

I thought this would only be the case for local test environment, but I'm seeing it in production too.

At the very least we should have a cron job that tries the secure endpoints to make sure unauthenticated users can't get in and sends mail or slack message if it fails.
Could do something nicer with Selenium.

We need to have a working image and user table before this can be added because we'll user userID and imageID as primary key.

first pass page:

• let user select from a range of images
• get channels based on image selected
• if preferences for image already exist, show them
• let user select min/max for each channel
• save the above min/max preferences to db

This seems better than the current method of storing parameters in a yaml file.