Load and transform Opsgenie escalation policies

[AlexisJasso]

https://firehydrant.atlassian.net/browse/PDE-1847

Wow it was a road getting here. Some salient decisions made when creating this (there are code comments for these also):

• Our targets for escalation policy steps don't map well to most of the target types for Opsgenie escalation policy steps. For some of those types, we've simply dropped them an logged to the console that we're doing so. For escalating to a schedule specifically, we're targeting all FH schedules (which are actually OG rotations) that were created from the OG schedule, which might be awkward if that separation changed the semantics here enough.
• OG tracks a delay (starting from esc policy start) before an escalation policy step starts, where FH uses a timeout on the previous step before the next step begins. Also, OG allows delay times on the order of hours, days, weeks, or months (!?), where FH only allows minutes. We've locked in a minimum of 1min and a max of 60min for the resulting FH steps, but anyone using an escalation policy step on a different order of magnitude almost certainly has a different semantic meaning there and this should be viewed with suspicion. As above, we log to the console when doing this, and can tag this for future work to insert comments into the tf file or something else as we choose.
• A query was added to get schedules where the schedule ID is LIKE ?. This is to support the fan out mentioned above and can be used for Pagerduty schedule layers also.

[wilsonehusin]

good stuff -- thank you AJ! some questions and should be ok to merge after they're addressed

[wilsonehusin]

is this correct? seems like the test reveals that the query returns none?

[AlexisJasso]

Heh. I missed this is the flurry of generating files. The test runs LoadTeamMembers() without any users or teams so... yeah, null.

Updating this test to run LoadUser() and LoadTeams() first and then I'll get the correct golden file in here.