Add option to list glacierefied folders

[chuwy]

In list_runids function we explicitly skip run ids that are archived on AWS Glacier, which means they will never appear in run manifest. I believe this is wrong solution as customer can restore particular folder without intention to reprocess it (let's say with PySpark job).

I propose to:

• Add option list_runids(include_archived=False) to list folders archived to AWS Glacier
• Make list_runids()return objects of RunId class (rather than plain string) that can hold information whether folder is archived
• Add third possible state to https://github.com/snowplow/snowplow-python-analytics-sdk/issues/29 to mark RunId processing was explicitly cancelled. Something like CancelledAt.

This should make run manifests feature able to take full control over data processing.

TODO: Think what to do with another storage classes. Currently we list folders that have only STANDARD class.

[alexanderdean]

I think this makes sense. I would probably call the state IgnoredAt, rather than CancelledAt?

[chuwy]

Agree about IgnoredAt.