Move schedule approval and match scheduling state into Sprocket

[gankoji]

Summary

Make Sprocket the source of truth for captain scheduling approvals and scheduled match times instead of storing active scheduling state on legacy mledb.series rows.

Why

Captain scheduling still writes approved match times into the legacy series model. That keeps scheduling metadata coupled to legacy bridge rows even though Sprocket should own active scheduling workflows.

Relevant code

• core/src/scheduling/schedule-group/schedule-group.service.ts
• core/src/scheduling/schedule-fixture/schedule-fixture.service.ts
• MLEDB/src/bot/commands/Management/ScheduleMatch.command.ts
• dataset consumer: datasets/queries/public/schedules/matches.sql

Scope

• Add a Sprocket-native place to store approved match scheduling time and approval state.
• Update scheduling write paths to persist this state in Sprocket first.
• Limit legacy series.scheduled_time writes to temporary compatibility if still needed.
• Identify downstream consumers that must be cut over to the new Sprocket field.

Acceptance criteria

• A captain scheduling action can be fully written and queried from Sprocket.
• Approved match times no longer require mledb.series as the source of truth.
• Any compatibility bridge into legacy scheduling state is documented and temporary.
• Downstream consumers have a clear migration path to the new Sprocket field(s).