Pyfgaws

This is a minor bug fix release.

The following bug fixes have been made

* The added `--delay` option from [Release 0.1.4](https://github.com/fulcrumgenomics/pyfgaws/releases/tag/0.1.4) introduced a bug with `run-job --print-log` would fail. Fixed in #19.

This is a minor feature release.

The following features have been added:

* expose the delay between API requests via the `--delay` command line option for the `run-job` and `watch-job` tools (18)
* a small random jitter (+/-2 seconds) to the delay to help avoid AWS batch API limits for monitoring batch jobs in the cases of many requests across concurrent jobs. A minimum delay of 1 second is subsequently applied.

This is a bug fix release.

The following bugs have been fixed:
* Support command-line parsing the batch Status (15)

Note: release 0.1.1 and release 0.1.2 are not available, as they were inadvertently published from a non-master branch, and so are deleted on pypi and not released here.

This is a backwards API incompatible release.

The following changes were made:
* do not log if no batch logs are present (6)
* refactor batch.Status to store which statuses have CloudWatch logs; this introduces a backwards compatible API change (11)
* `watch_job` does not skip over cloud logs (7)
* better handling of exceptions while print logs while watching a batch job (11)
* Update mypy_boto3 so type discovery is not needed (13)
* Batch submit jobs needs to filter the input keyword arguments (14)
* set the job id after submission (14)

This is the first release of `pyfgaws` version 0.0.1.

The following tools are new:
- `run-job`: runs an AWS batch job and optionally waits for it to reach a given state (ex. succeeded)
- `watch-job`: watches an AWS batch job until it completes
- `watch-logs`: watches an AWS CloudWatch log stream

The following APIs are new for AWS batch:
- `batch.BatchJob`: stores information about and operates on a batch job
- `batch.BatchJob.from_id`: retrieves an existing batch job by ID and builds it
- `batch.BatchJob.stream`: gets the log stream for this batch job
- `batch.BatchJob.submit`: submits the batch job
- `batch.BatchJob.get_status`: gets the current status (ex. SUCCEEDED) for the batch job
- `batch.BatchJob.cancel_job`: cancels a batch job (does nothing if RUNNING)
- `batch.BatchJob.terminate_job`: terminates a batch job (cancels, or terminates if RUNNING)
- `batch.BatchJob.describe_job`: gets the current state of the job and associated job metadata
- `batch.BatchJob.wait_on`: waits on the batch job to enter one or more states, with each state having success or failure conditions
- `batch.BatchJob.wait_on_running`: waits on the batch job to start running
- `batch.BatchJob.wait_on_complete`: waits on the batch job to either succeed or fail

The following APIs are new for AWS CloudWatch logs:
- `logs.Log`: stores information about a log stream to enable easy iteration over logs
- `logs.Log.reset`: resets the `Log` object to iterate over potentially new log events