Implement performance improvements across backend and frontend

- models.py: add composite (user_id, date) indexes to flock_history,
  feed_purchases, and other_purchases for faster date-filtered queries
  (egg_collections already had one via its unique constraint)
- main.py: add v2.2 migration to create the three composite indexes on
  existing installs at startup
- stats.py: fix N+1 query in monthly_stats — flock history is now fetched
  once and looked up per month using bisect_right instead of one DB query
  per month row; also remove unnecessary Decimal(str(...)) round-trips
  since SQLAlchemy already returns Numeric columns as Decimal
- eggs.py: add limit parameter (default 500, max 1000) to list_eggs to
  cap unbounded fetches on large datasets
- dashboard.js: pass start= (30 days ago) when fetching eggs so the
  dashboard only loads the data it actually needs for the chart and
  recent collections list

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

backend/main.py

@@ -79,6 +79,18 @@ def _run_migrations():
         except Exception:
             db.rollback()  # column already exists — safe to ignore
 
+        # v2.2 — composite (user_id, date) indexes for faster filtered queries
+        for sql in [
+            "CREATE INDEX ix_flock_history_user_date  ON flock_history  (user_id, date)",
+            "CREATE INDEX ix_feed_purchases_user_date  ON feed_purchases  (user_id, date)",
+            "CREATE INDEX ix_other_purchases_user_date ON other_purchases (user_id, date)",
+        ]:
+            try:
+                db.execute(text(sql))
+                db.commit()
+            except Exception:
+                db.rollback()  # index already exists — safe to ignore
+
 
 @asynccontextmanager
 async def lifespan(app: FastAPI):