fix(inverter): skip the discharge target write on AC Coupled GivTCP inverters (#4517)

AC Coupled inverters don't have a working Discharge_Target_SOC_1 register -
GivTCP still reports each write as successful, but it never persists
between cycles, so adjust_force_export() saw a permanent mismatch and
rewrote it every single cycle indefinitely.

Confirmed against GivEnergy's own firmware archive
(github.com/DJBenson/giv-firmware): "AC Coupled" is a single product line
with exactly two firmware releases ever published, not an early/late
generational split the way the Hybrid line has Gen1/2/3 - so there's no
newer AC-coupled variant this would need to keep working for. Skip the
write outright when raw.invertor.model reports "Ac", rather than attempting
one already known to be doomed.

Supersedes the earlier enable-before-write approach on this same issue -
that theory relied on a correlation with a second reporter's data that
turned out to be from before #4492/#4518 existed and was confounded by the
exact race condition those fixes address, so it was never solid evidence
that AC Coupled's register works anywhere. Dropped rather than kept
alongside this fix, since it has no remaining supporting evidence.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

Author: chalfontchubby

apps/predbat/inverter.py

@@ -2523,13 +2523,24 @@ def adjust_force_export(self, force_export, new_start_time=None, new_end_time=No
         if force_export:
             target_soc = int(self.reserve_percent)
             if self.rest_data and self.rest_v3:
-                current = self.rest_readDischargeTarget()
-                if current is None:
-                    self.log("Inverter {} No current discharge target to read, export target not written".format(self.id))
-                elif current != target_soc:
-                    self.rest_setDischargeTarget(target_soc)
+                # AC Coupled inverters don't have a working Discharge_Target_SOC_1 register - GivTCP
+                # still reports a write as successful, but it never persists between cycles, so the
+                # caller sees a permanent mismatch and rewrites indefinitely (#4517). Confirmed
+                # against GivEnergy's own firmware archive (github.com/DJBenson/giv-firmware): "AC
+                # Coupled" is a single product line with exactly two firmware releases ever published
+                # (D212-A212, A214-D214) - not an early/late generational split the way Hybrid has
+                # Gen1/2/3 - so there's no newer AC-coupled variant this would need to keep working for.
+                inverter_model = self.rest_data.get("raw", {}).get("invertor", {}).get("model", "")
+                if inverter_model == "Ac":
+                    self.log("Inverter {} is AC Coupled, discharge target register not supported, export target not written".format(self.id))
                 else:
-                    self.log("Inverter {} Current discharge target is already set to {}".format(self.id, current))
+                    current = self.rest_readDischargeTarget()
+                    if current is None:
+                        self.log("Inverter {} No current discharge target to read, export target not written".format(self.id))
+                    elif current != target_soc:
+                        self.rest_setDischargeTarget(target_soc)
+                    else:
+                        self.log("Inverter {} Current discharge target is already set to {}".format(self.id, current))
             elif "discharge_target_soc" in self.base.args:
                 current = self.base.get_arg("discharge_target_soc", index=self.id, required_unit="%")
                 try:

apps/predbat/tests/test_inverter.py

@@ -1927,6 +1927,66 @@ def test_discharge_target_control_signal(test_name, ha, inv, dummy_rest):
     return failed
 
 
+def test_discharge_target_skipped_for_ac_coupled(test_name, ha, inv, dummy_rest):
+    """
+    Regression test for issue #4517: AC Coupled inverters (raw.invertor.model == "Ac") don't have a
+    working Discharge_Target_SOC_1 register - GivTCP reports a write as successful, but it never
+    persists between cycles, so the caller sees a permanent mismatch and rewrites indefinitely.
+    Confirmed against GivEnergy's own firmware archive: "AC Coupled" is a single product line with
+    only two firmware releases ever published, not an early/late generational split - so there's no
+    newer AC-coupled variant that would need this write to still happen. Skip it outright rather than
+    attempting a write already known to be doomed.
+    """
+    failed = False
+    print("Test: {}".format(test_name))
+
+    saved_rest_data = inv.rest_data
+    saved_rest_api = inv.rest_api
+    saved_rest_v3 = inv.rest_v3
+
+    try:
+        inv.rest_api = "dummy"
+        inv.rest_v3 = True
+        inv.reserve_percent = 20
+
+        start_time = "03:33:00"
+        end_time = "04:44:00"
+        ts = datetime.strptime(start_time, "%H:%M:%S")
+        te = datetime.strptime(end_time, "%H:%M:%S")
+
+        inv.rest_data = {
+            "Control": {"Mode": "Timed Export", "Enable_Discharge_Schedule": "on"},
+            "Timeslots": {"Discharge_start_time_slot_1": start_time, "Discharge_end_time_slot_1": end_time},
+            "raw": {"invertor": {"discharge_target_soc_1": "4", "model": "Ac"}},
+        }
+        dummy_rest.clear_queue()
+        dummy_rest.rest_data = copy.deepcopy(inv.rest_data)
+
+        inv.adjust_force_export(True, ts, te)
+
+        commands = dummy_rest.get_commands()
+        if commands:
+            print("ERROR: {}: AC Coupled should attempt no discharge-target REST commands at all, got {}".format(test_name, commands))
+            failed = True
+
+        # A non-AC-Coupled model (or no model reported at all) must still attempt the write as before.
+        for model in ["Hybrid", ""]:
+            inv.rest_data["raw"]["invertor"]["model"] = model
+            dummy_rest.clear_queue()
+            dummy_rest.rest_data = copy.deepcopy(inv.rest_data)
+            inv.adjust_force_export(True, ts, te)
+            commands = dummy_rest.get_commands()
+            if not any(c[0] == "dummy/setDischargeTarget" for c in commands):
+                print("ERROR: {}: model={!r} should still attempt setDischargeTarget, got {}".format(test_name, model, commands))
+                failed = True
+    finally:
+        inv.rest_data = saved_rest_data
+        inv.rest_api = saved_rest_api
+        inv.rest_v3 = saved_rest_v3
+
+    return failed
+
+
 def test_discharge_target_read_prefers_control(test_name, ha, inv):
     """
     Regression test for issue #4517: a discharge target write kept firing every cycle even when
@@ -3173,6 +3233,12 @@ def run_inverter_tests(my_predbat_dummy):
     if failed:
         return failed
 
+    # Regression test for issue #4517 (follow-up): AC Coupled inverters don't have a working
+    # discharge target register, skip the write entirely rather than retrying it forever
+    failed |= test_discharge_target_skipped_for_ac_coupled("discharge_target_skipped_for_ac_coupled", ha, inv, dummy_rest)
+    if failed:
+        return failed
+
     # Regression test for issue #4517: the caller deciding whether to write at all must also prefer
     # Control.Discharge_Target_SOC_1 over the slow raw.invertor fallback, or it re-writes every cycle
     failed |= test_discharge_target_read_prefers_control("discharge_target_read_prefers_control", ha, inv)