perf(subgraph): reduce store operations to speed up indexing

[silent-cipher]

Goal

Increase subgraph indexing speed (backfill and steady-state) by cutting store operations per event.

Sub-issues

Phase 1 — remove dead work (small diffs, no product impact):

• perf(subgraph): remove dead SumTree maintenance from piece handlers #135 Remove dead SumTree maintenance from piece handlers
• perf(subgraph): stop materializing ProvingWindow entities #136 Stop materializing ProvingWindow entities
• perf(subgraph): remove dead tx-input ABI parsing in handlePiecesAdded #137 Remove dead tx-input ABI parsing in handlePiecesAdded
• perf(subgraph): cut per-event store round-trips #138 Cut per-event store round-trips (loadInBlock, redundant existence check)

Phase 2 — structural (schema changes, client updates):

• perf(subgraph): migrate weekly/monthly activity metrics to native timeseries + @aggregation #139 Migrate weekly/monthly activity metrics to native timeseries + @aggregation
• perf(subgraph): slim EventLog/Transaction entities and drop unused counters #140 Slim EventLog / Transaction entities and drop unused counters