refactor(planner): remove the inert battery-schedule feature

[woopstar]

Summary

• Removes the fixed battery discharge/charge-schedule feature (batteries_schedule_1/2/3), which issue Triage: battery-schedule feature appears inert since MILP became the sole active planner #860 found to be functionally inert whenever MILP is active (the default planner path).
• Also removes apply_arbitrage_grid_charge, which was gated entirely on at least one schedule being enabled.
• current_required_battery / calculate_required_battery_until_solar are not touched — they're a separate mechanism feeding excess-export and the EV discharge-cap reserve, unrelated to fixed schedules.

Branch

refactor/860-remove-battery-schedules

Why

In MILP-only mode, candidate_generator.py only produces no_action, passive, and milp candidates — the schedule-consuming heuristic candidates are commented out (deleted entirely by #897, a separate PR). Both surviving fallback candidates (no_action, passive) explicitly clear all charge/discharge recommendations before scoring, and the MILP itself never reads battery_schedules at all. So whichever candidate wins, schedule-driven recommendations never reach the final plan.

This was already partially documented as intentional (docs/candidate-generation.md, docs/planner-spec.md), but the config, entities, and planner passes remained fully wired with no indication to users that they had no effect. See issue #860 for the full triage — the user chose full removal over re-wiring it as a MILP-unavailable fallback or leaving it documented as inert, explicitly overriding an earlier "keep fixed battery schedules" decision recorded in .github/memories.md (updated in this PR to reflect the reversal, and again to point at #897's follow-up cleanup).

This PR does not touch candidate_generator.py/candidate_selector.py/docs/candidate-generation.md/docs/planner-spec.md — those files are exclusively owned by #897, so the two PRs have zero file overlap and can merge in either order.

What changed

Config / UI

• Removed the batteries_schedules config-flow step (flows/batteries_schedules.py, flows/schedule_helpers.py) and repointed ocpp → batteries_wait_mode directly in both config_flow.py and options_flow.py.
• Removed the six hsem_batteries_enable_batteries_schedule_* config keys from const.py.
• Removed the 3 schedule switches and 6 start/end time entities (switch.py, time.py, utils/sensornames/controls.py, custom_switches/description.py, custom_times/description.py).
• Removed the now-dead validate_time_window validator (its only caller was the schedule step).

Models

• Deleted models/battery_schedule.py, models/battery_schedule_input.py, and the BatteryScheduleConfig dataclass in models/sensor_config.py.
• Removed the battery_schedules field from PlannerInput and CoordinatorData.

Planner

• Deleted planner/charging/pre_charge.py (apply_charge_schedules) and planner/charging/arbitrage_charge.py (apply_arbitrage_grid_charge).
• Removed apply_discharge_schedules from discharge_scheduler.py.
• Removed all three call sites from engine_core.py::_schedule_slots.
• charge_scheduler.py / planner/charging/__init__.py are now thin re-export shims for the surviving apply_opportunistic_charge / apply_window_hysteresis.

Coordinator

• Removed _batteries_schedules / _batteries_schedules_remaining_capacity_needed state, the build_battery_schedules() call, and the batteries_schedules kwarg threaded through build_planner_input().
• Removed the resolver rule in recommendation_resolver.py that forced BatteriesDischargeMode based on remaining schedule capacity.

Docs / translations

• Updated docs/architecture-overview.md, docs/config-flow-reference.md, docs/planner-guide.md, docs/sensors-reference.md, and docs/dashboard.yaml (removed the Battery Schedules dashboard section).
• Removed the corresponding keys from translations/en.json and translations/da.json.

Tests

• Deleted whole files that were 100% schedule-specific: test_schedule_validation.py, test_arbitrage_grid_charge.py, test_charge_scheduler_capacity.py, test_48h_second_day.py.
• Deleted schedule-only classes/methods embedded in otherwise-unrelated files (e.g. TestP004Schedule3Default, TestBatteryAboveScheduleNeed, TestEvLoadDoesNotInflateChargeNeeded, TestScheduleValidator).
• Updated shared fixtures (tests/planner/fixtures.py and others) that threaded a schedules/battery_schedules kwarg through PlannerInput construction.
• tests/planner/test_cycle_cost_guard.py::_make_two_slot_input and tests/planner/test_invariants.py::test_grid_import_exceeds_stored_with_conversion_loss were redesigned to use natural price-arbitrage incentives instead of a forced discharge schedule — the MILP doesn't need a schedule to decide when to charge, so this preserves the invariant coverage without the removed feature.
• Repointed tests that used a schedule key purely as generic example data (test_entity_platform_classes.py, test_platform_entity_refactor.py) to a surviving key (hsem_ev_deadline_time).

Test and lint results

All four quality gates pass:

./scripts/quality.sh lint     ✅ ruff format + ruff check + prettier — clean
./scripts/quality.sh typing   ✅ mypy — 0 errors
./scripts/quality.sh quality  ✅ pyright + vulture — 0 errors (1 pre-existing unrelated warning)
./scripts/quality.sh test     ✅ 2926 passed

Known limitations / follow-ups

None expected. flows/migrations.py's _V1_DEPRECATED_KEYS (a different, already-deprecated ..._min_price_difference sub-key list) is left untouched — it's a pre-existing migration cleanup unrelated to the current batteries_schedule_N toggle/start/end keys removed here.

Fixes #860