Updates to user defined buckets

[jhkennedy]

I was able to deploy this branch to https://hyp3-jhk-sandbox.asf.alaska.edu/ and run a few burst jobs that write their files to a private bucket in the LAVAs account, as well as a couple autorift jobs that failed expectedly to check the failure path as well.

The example jobs:
https://hyp3-jhk-sandbox.asf.alaska.edu/jobs?name=test-user-bucket&user_id=jhkennedy

or

import hyp3_sdk as sdk

hyp3 = sdk.HyP3('https://hyp3-jhk-sandbox.asf.alaska.edu')
jobs = hyp3.find_jobs(name='test-user-bucket', user_id='jhkennedy')

This PR:

• Add the bucket-policy endpoint to the OpenAPI spec
• changes the bucken name example to "my-example-bucket" to make the bucket-policy endpoint example a little more clear
• compacts the returned bucket policy and returns it as a dict->JSON instead of string
• fixes the sts.get_caller_identity() usage for JSON output from AWS APIs
• Compacts and updates the permissions of:
  • the batch compute task for external s3 buckets
  • get-files so it can get files from external buckets
  • upload-logs so it can get logs from external buckets
• the step function to provide job_id, bucket, and bucket_prefix to get-files and upload-logs so we don't have to make a dynamo call
• updates get-files to handle user bucket download URLs and missing expiration times (user's may not have a lifecycle policy on their buckets)
• upload-log to always write the log file as {job_id}.log irrespective of the bucket and bucket_prefix provided
• make sure every job spec is using Ref::bucket Ref::bucket_prefix instead of ! Ref bucket and Ref::job_id

With these changes, I think we're ready to merge to develop and take it for a spin in a broader set of environments, particularly EDC.

[jhkennedy]

This was needed because a different version kept getting installed and uploaded, which pip was warning about...

I'd like to test removing all these after we merge to test, because we really shouldn't need to specify them.