connect_hub and connect_model_output: add option to skip data checks for model-output files

[bsweger]

Background

This is a follow up to the performance investigation related to cloud-based hubs: #37

When researching the above, we found two places in connect_hub and connect_model_output that affect performance when using hubData with larger hubs based on S3:

• the check_file_format function
• using exclude_invalid_files = TRUE (of the two, this is the larger performance impact)

To ensure users of cloud-based hubs can connect in a reasonable amount of time, we decided to add an optional parameter to connect_hub and connect_model_output that will specify whether or not to perform the above checks. When not supplied by the user, the default behavior will be to skip them when working with cloud-based hubs.

Definition of done

• [ ] connect_hub has an optional parameter to skip the checks
• [ ] connect_model_output has an optional parameter to skip the checks
• [ ] when the new parameter isn't specified by the user, the code will set it skip checks when working with cloud-based data
• [ ] when there's an error opening the files when checks are skipped, the message should a note about trying again with checks enabled

[bsweger]

I'd like to tackle this one, and get some R reps!