TMSS-152: refactored observation task->subtask specification conversion, so it's more generic, and easier to read.

TMSS-152: made sure that create_task_drafts_from_scheduling_unit_draft is always called when processing specification in scheduling_unit_draft. Check and skip already created task_drafts and/or relations

TMSS-189: create UC1 SchedulingUnit with a whole task grapsh specified in json conform the scheduling-unit-schema (based on Roy's example data in the t_tasks test)

TMSS-189: return fields sorted alphabetically, with id and url first so it's easy to identify and 'click' them

TMSS-152: Calculate start time using previous observation subtask info and time_offset specified in the scheduling_relation. Also use duration from task specification to calculate stop time

TMSS-189: made seperate methods for populating cycles and projects. removed example TaskDrafts, and made/called example SchedulingUnit

TMSS-152: Create also SchedulingRelationBlueprint of SchedulingRelationDraft during draft-to-blueprint. Fix in name test data for rest api.

TMSS-152: use same related names of task scheduling relation draft/blueprint as currently delivered with TMSS-50 in master

TMSS-152: Get subtasks specification from task specification and remove hard coded sub-task settings to prepare for UC1 observation run

TMSS-189: fixed all the little details for scheduling a pipeline from TMSS with the ResourceAssigner

TMSS-166 (already merged) implemented the same Data Model  (TaskSchedulingRelation[Draft,Blueprint]) requested in TMSS-50

TMSS-189: stupid parset looses type info. So, return typed dicts for those who want it, return parsets for those who don't care about types