Glue crawler Athena table creation fails silently on invalid yml input

[asparke2]

Describe the bug
If the aws > s3 >prefix entry in the yml file has a trailing /, the s3 upload will work, and the glue crawler will be created, but no Athena table will be created because the s3 location will be invalid.

To Reproduce
Steps to reproduce the behavior:

1. Add a trailing / to the aws:s3:prefix entry in the yml file as shown below

postprocessing:
  aws:
    region_name: 'us-west-2'
    s3:
      bucket: eulp    
      prefix: simulation_output/regional_runs/comstock/region-1-fort-collins-3500-ts/
    athena:
      glue_service_role: service-role/AWSGlueServiceRole-default
      database_name: testing
      max_crawling_time: 300 #time to wait for the crawler to complete before aborting it

2. run a post-process only run

Expected behavior
First, sanitize the input to strip trailing /. Second, check that files exist inside the target s3 bucket before initializing the glue crawler.

Logs
Root issue was identified by noticing // in the following line of postprocessing.out:

INFO:2020-05-22 15:00:22:buildstockbatch.postprocessing:Upload to S3 completed. The files are uploaded to: eulp/simulation_output/regional_runs/comstock/region-1-fort-collins-3500-ts//region-1-fort-collins-3500-ts/

Platform (please complete the following information):

• Eagle
• BuildStockBatch 0.18
• OpenStudio-BuildStock version, branch, or sha: comstockv1.8
• Local Desktop OS: Windows

[nmerket]

You gonna open up a PR for this or what?

[asparke2]

If you'd prefer a PR in the current state I'll do it, but I figured I'd wait until I had fixed added/fixed unit tests for it first since there is an easy workaround. Right now the unit tests are failing on new code which checks the s3 location, and the test s3 location doesn't really exist.

[nmerket]

I'd be open to a draft PR.