refactor(givtcp): finish moving GivTCP REST handling into the component

[chalfontchubby]

Follow-on to #4649, per your ask there. Stacked on that branch rather than main - review/merge that one first.

Takes the three things #4649 explicitly left behind off the direct REST client, so inverter.py no longer parses GivTCP's REST blob.

Commits

1. 6c83993c - inverter mode and pause controls as entities. inverter_mode (Control.Mode) plus, on v3 only, pause_mode/pause_start_time/pause_end_time. All four already existed as writable sensor_list keys, so the entity paths in Inverter were already there and the REST branches simply go. Pause is gated on every configured inverter reporting v3, mirroring the existing givtcp_rest_power_ignore "any opts out, leave them all" rule - v2 has no /setBatteryPauseMode, which is why the old REST path was v3-gated too.

2. d614c316 - the GivTCP discharge target written every cycle even when unchanged (regression from #4492) #4517 discharge-target model check. Rather than plumbing the raw model string out to a new apps.yaml key so Inverter could keep making the decision, the component just doesn't publish a discharge_target_soc entity for a model whose register doesn't work. adjust_force_export already leaves a target it can't read alone, so an absent entity is the whole mechanism - and inverter.py no longer knows about inverter models at all. A mixed fleet still works per-inverter.

3. 97ad0a0e - battery/capacity discovery. GivTCPRest gains normalised accessors that absorb the version differences previously open-coded in Inverter.__init__: v2's Invertor_Details vs v3 renaming it to the serial number, nominal capacity in raw register units vs kWh, per-pack temperature averaging across three field shapes, and the two calibration representations. The 19.53125 divisor is carried over verbatim - it was back-calculated rather than derived, and this doesn't pretend to explain it.

Two bugs found on the way

Both mutation-tested (confirmed failing without the fix):

• _handle_write had no branch for a select that isn't a time slot. A plain-select write was silently dropped and the entity then republished unchanged, which reads as a successful no-op. Would have bitten the moment inverter_mode/pause_mode landed.
• The rate entities advertised a generic max of 20000W. Inverter derives battery_rate_max_raw for a GE inverter from exactly that attribute. Harmless while REST discovery still overrode it - but the moment discovery moved, every GivTCP user's battery rate limit would silently have become 20kW. They now carry Invertor_Max_Bat_Rate.

Judgement calls worth a look

• adjust_inverter_mode's 30s changed_start_end sleep stays gated on rest_api. It works around GivTCP's own HA integration lagging behind a window write; a GivTCP-REST inverter has no such lag, because the component applies the write and republishes inline before returning. Dropping the guard would have quietly added 30s to every window change for GivTCP users.
• One new optional apps.yaml key, battery_calibration (documented). A calibration cycle drives the battery outside its normal SoC range, so Predbat disables itself for that inverter while one runs. Absent means never calibrating, so nothing changes for inverters that don't report it.
• The pause entity carries GivTCP's native vocabulary (Disabled/PauseCharge/...), not the GE Cloud spelling - adjust_pause_mode picks between the two from the value it reads back, so this keeps it on the same side of that branch as the REST path was.

Deliberately still on REST in inverter.py

The version/serial/firmware metadata read, update_status's refresh, and the reserve read. That last one differs between the branches - REST doesn't apply the battery_min_soc floor the entity path does, and reserve_percent_current is a reading of what the inverter is actually set to, so collapsing them is a behaviour change rather than a refactor. Left for a separate decision rather than changed silently.

Testing

31 component tests (up from 23). The two real captures (rest_v2.json/rest_v3.json) move from test_inverter to the component tests and now assert the accessors rather than Inverter's internals, plus per-version calibration coverage. test_inverter's REST-shaped mode/target assertions updated - the REST phase now asserts the entity write and that no REST command is issued. Full run_pre_commit green.

Also fixed a test-isolation bug: a queue_rest_data entry fed setBatteryMode's read-back verification, and with that call gone it leaked into the next test's runAll and made an unrelated REST test retry 5x.

Same caveat as #4649: no validation against real hardware. Draft until someone dogfoods it. The write/republish timing is the part only a live inverter really proves.

🤖 Generated with Claude Code