Replace legacy standings and scheduling dataset queries with Sprocket-native sources

[gankoji]

Summary

Move standings and scheduling datasets off legacy team/division/scheduled-time metadata and onto Sprocket-native sources.

Why

Public standings and match schedule datasets still depend on legacy MLEDB metadata for division/conference labels and approved match times.

Candidate queries

• queries/public/standings.sql
• queries/public/s18/standings_s18.sql
• queries/public/s19/standings_s19.sql
• queries/public/schedules/matches.sql

Scope

• Replace legacy team/division metadata dependencies with Sprocket-owned metadata.
• Replace legacy scheduled-time joins with the Sprocket-native scheduling source once available.
• Preserve ordering and public data shape where practical.

Acceptance criteria

• Standings and scheduling queries no longer require mledb.team, mledb.division, or mledb.series.
• Dataset generation succeeds from Sprocket-native metadata.
• Any schema changes are documented for downstream consumers.