Blackduck_hub.py should reports failures as separate tasks

Currently, blackduck_hub.py reports failures by emulating unit test failures by writing the pass/fail list to reports.json.

As a result, a single BFG is generated which must be manually split apart. To reduce this operation burden, each failure should be reported as a separate task that fails.

The current task will be split into two tasks: a scanner task and one or more report tasks.

1. The scanner task will be responsible for analyzing blackduck and generate a series of report files. These report files will be uploaded to S3. It will also generate a set of N report tasks with "generate.tasks" where N = the number of failures.
2. The report task will generate a single failure by downloading all reports and generating a single failure for its assigned report.