Merge branch 'fix/ev-teltocharge-reg15-and-session-visibility' into dev

# Conflicts:
#	docs/planning-changelog.md

backend/services/control/outputs.py

@@ -18,33 +18,52 @@ logger = logging.getLogger(__name__)
 TELTO_REG_AMPS_TO_USE = 15        # 0 = stop, 6–32 A
 TELTO_REG_COMM_TIMEOUT_S = 19     # watchdog: bez komunikace → failsafe
 TELTO_REG_FAILSAFE_CURRENT_A = 20
-#: Výpadek EMS: po 5 min bez zápisu wallbox přejde na failsafe proud —
-#: auto se přes noc nabije i bez EMS (pomalu), místo aby stálo na 0 A.
+#: Výpadek EMS: po watchdog_comm_timeout_s bez komunikace wallbox přejde na
+#: failsafe proud — auto se přes noc nabije i bez EMS (pomalu), místo aby
+#: stálo na 0 A. Defaulty (fallback, když řádek chargeru nemá vlastní hodnoty).
 TELTO_WATCHDOG_TIMEOUT_S = 300
 TELTO_WATCHDOG_FAILSAFE_A = 8
 
 
-def _telto_setpoint_registers(current_a: int) -> list[tuple[int, str, int]]:
+def _telto_setpoint_registers(
+    current_a: int,
+    *,
+    comm_timeout_s: int = TELTO_WATCHDOG_TIMEOUT_S,
+    failsafe_a: int = TELTO_WATCHDOG_FAILSAFE_A,
+) -> list[tuple[int, str, int]]:
     """Registry pro jeden export tick: limit proudu + watchdog konfigurace.
 
-    Write-on-change: volající VŽDY filtruje přes drop-unchanged proti
-    fn_modbus_device_state_map (poslední written/verified per registr) —
-    watchdog 19/20 se reálně zapíše jen po startu / po výpadku zařízení,
-    amps (15) jen při změně plánu. Watchdog timer TeltoCharge sytí jakákoli
-    validní Modbus komunikace (i FC3 čtení telemetrie každých 60 s), takže
-    periodické zápisy k udržení spojení NEJSOU potřeba (oficiální protokol,
-    docs/04-modules/modbus-registers-teltocharge.md).
+    **Reg 15 (amps to use) NENÍ write-on-change** — viz `_assert_amps_register`.
+    Je to volatilní řídicí registr: TeltoCharge ho po výpadku komunikace sám
+    přepíše na failsafe (reg 20), aniž by o tom vznikl journal řádek. Kdyby se
+    reg 15 dropoval proti journalu (poslední „0 verified"), EMS by tichý drift
+    0 → 8 A NIKDY nezahlédlo (verify čte zpět jen `written` řádky) a nikdy ho
+    neopravilo. Proto se reg 15 re-asertuje KAŽDÝ export tick (≤ 8×/hod) —
+    EEPROM wear se týká jen konfiguračních 19/20, které write-on-change zůstávají.
+
+    Watchdog timer TeltoCharge sytí jakákoli validní Modbus komunikace (i FC3
+    čtení telemetrie každých 60 s), takže periodické zápisy k udržení spojení
+    NEJSOU potřeba; failsafe/timeout (19/20) per charger z DB.
     """
     a = int(current_a)
     if a < 6:
         a = 0
     return [
         (TELTO_REG_AMPS_TO_USE, "telto_amps_to_use", min(a, 32)),
-        (TELTO_REG_COMM_TIMEOUT_S, "telto_comm_timeout_s", TELTO_WATCHDOG_TIMEOUT_S),
-        (TELTO_REG_FAILSAFE_CURRENT_A, "telto_failsafe_a", TELTO_WATCHDOG_FAILSAFE_A),
+        (TELTO_REG_COMM_TIMEOUT_S, "telto_comm_timeout_s", int(comm_timeout_s)),
+        (TELTO_REG_FAILSAFE_CURRENT_A, "telto_failsafe_a", max(0, min(int(failsafe_a), 32))),
     ]
 
 
+def _split_amps_and_watchdog(
+    registers: list[tuple[int, str, int]],
+) -> tuple[list[tuple[int, str, int]], list[tuple[int, str, int]]]:
+    """Rozdělí registry na (reg 15 = vždy zapsat) a (19/20 = write-on-change)."""
+    amps = [r for r in registers if r[0] == TELTO_REG_AMPS_TO_USE]
+    watchdog = [r for r in registers if r[0] != TELTO_REG_AMPS_TO_USE]
+    return amps, watchdog
+
+
 def _current_limit_for_charger(charger_code: str, sp: ControlSetpoints) -> int:
     c = (charger_code or "").strip().lower()
     if c == "ev-charger-1":
@@ -74,7 +93,8 @@ async def write_ev_setpoints(
 
     rows = await db.fetch(
         """
-        SELECT ec.id AS asset_id, ec.code, se.host, se.port, se.unit_id
+        SELECT ec.id AS asset_id, ec.code, se.host, se.port, se.unit_id,
+               ec.watchdog_failsafe_a, ec.watchdog_comm_timeout_s
         FROM ems.asset_ev_charger ec
         JOIN ems.site_endpoint se ON se.id = ec.endpoint_id
         WHERE ec.site_id = $1
@@ -97,16 +117,31 @@ async def write_ev_setpoints(
         unit_id = int(row["unit_id"] if row["unit_id"] is not None else 1)
         current_a = _current_limit_for_charger(code, setpoints)
 
-        registers = _telto_setpoint_registers(current_a)
-        # Write-on-change: poslední written/verified stav (ne jen verified) —
-        # zápis se nesmí opakovat každý tick, když verify čtení zaostává.
+        registers = _telto_setpoint_registers(
+            current_a,
+            comm_timeout_s=int(
+                row["watchdog_comm_timeout_s"]
+                if row["watchdog_comm_timeout_s"] is not None
+                else TELTO_WATCHDOG_TIMEOUT_S
+            ),
+            failsafe_a=int(
+                row["watchdog_failsafe_a"]
+                if row["watchdog_failsafe_a"] is not None
+                else TELTO_WATCHDOG_FAILSAFE_A
+            ),
+        )
+        amps_regs, watchdog_regs = _split_amps_and_watchdog(registers)
+        # Reg 15 = vždy (re-asert proti tichému watchdog failsafe driftu na
+        # zařízení, který nemá journal řádek). Reg 19/20 = write-on-change
+        # proti fn_modbus_device_state_map (poslední written/verified stav).
         device_state = await _fetch_device_state_registers(
             site_id, asset_id, db, asset_type="ev_charger"
         )
-        registers, skipped = _drop_registers_matching_last_verified(
-            registers, device_state
+        watchdog_regs, skipped = _drop_registers_matching_last_verified(
+            watchdog_regs, device_state
         )
-        if not registers:
+        to_write = amps_regs + watchdog_regs
+        if not to_write:
             logger.debug("EV setpoint [%s]: beze změny (%s A)", code, current_a)
             continue
 
@@ -119,7 +154,7 @@ async def write_ev_setpoints(
             host,
             port,
             unit_id,
-            registers,
+            to_write,
             db,
         )
         ok = await execute_modbus_commands(cmd_ids, db)
@@ -128,7 +163,7 @@ async def write_ev_setpoints(
             "EV setpoint [%s]: %s A (regs %s%s) -> %s",
             code,
             current_a,
-            [r for r, _, _ in registers],
+            [r for r, _, _ in to_write],
             f", skip {skipped}" if skipped else "",
             "written" if ok else "FAILED",
         )
@@ -155,7 +190,8 @@ async def write_ev_arrival_hold(
 
     row = await db.fetchrow(
         """
-        SELECT ec.id AS asset_id, ec.code, se.host, se.port, se.unit_id
+        SELECT ec.id AS asset_id, ec.code, se.host, se.port, se.unit_id,
+               ec.watchdog_failsafe_a, ec.watchdog_comm_timeout_s
         FROM ems.asset_ev_charger ec
         JOIN ems.site_endpoint se ON se.id = ec.endpoint_id
         WHERE ec.site_id = $1
@@ -170,20 +206,29 @@ async def write_ev_arrival_hold(
     if row is None:
         return False
     asset_id = int(row["asset_id"])
-    registers = _telto_setpoint_registers(0)
+    registers = _telto_setpoint_registers(
+        0,
+        comm_timeout_s=int(
+            row["watchdog_comm_timeout_s"]
+            if row["watchdog_comm_timeout_s"] is not None
+            else TELTO_WATCHDOG_TIMEOUT_S
+        ),
+        failsafe_a=int(
+            row["watchdog_failsafe_a"]
+            if row["watchdog_failsafe_a"] is not None
+            else TELTO_WATCHDOG_FAILSAFE_A
+        ),
+    )
+    amps_regs, watchdog_regs = _split_amps_and_watchdog(registers)
+    # Reg 15 = 0 A se zapíše VŽDY (tvrdé zastavení po píchnutí kabelu; wallbox
+    # po připojení sám rozjíždí nabíjení defaultem). Reg 19/20 write-on-change.
     device_state = await _fetch_device_state_registers(
         site_id, asset_id, db, asset_type="ev_charger"
     )
-    registers, skipped = _drop_registers_matching_last_verified(
-        registers, device_state
+    watchdog_regs, skipped = _drop_registers_matching_last_verified(
+        watchdog_regs, device_state
     )
-    if not registers:
-        logger.info(
-            "EV arrival hold [%s]: 0 A už na zařízení (skip %s)",
-            charger_code,
-            skipped,
-        )
-        return True
+    to_write = amps_regs + watchdog_regs
     cmd_ids = await create_modbus_commands(
         site_id,
         None,
@@ -193,14 +238,14 @@ async def write_ev_arrival_hold(
         str(row["host"]),
         int(row["port"] or 502),
         int(row["unit_id"] if row["unit_id"] is not None else 1),
-        registers,
+        to_write,
         db,
     )
     ok = await execute_modbus_commands(cmd_ids, db)
     logger.info(
         "EV arrival hold [%s]: 0 A (regs %s%s) %s",
         charger_code,
-        [r for r, _, _ in registers],
+        [r for r, _, _ in to_write],
         f", skip {skipped}" if skipped else "",
         "written" if ok else "FAILED",
     )

backend/services/planning/db_io.py

@@ -31,12 +31,15 @@ def _ev_session_from_json(obj: object) -> Optional[SimpleNamespace]:
         obj = json.loads(obj)
     if not isinstance(obj, dict):
         return None
+    # target_deadline SMÍ být None: oportunistická session (auto nad targetem,
+    # nebo bez nastaveného cíle) zůstává v plánu kvůli headroomu i jako známá
+    # zátěž. Tvrdý deadline constraint se aplikuje jen při energy_needed_wh > 0
+    # (a needed > 0 nastane jen s deadlinem). Dřív se taková session zahazovala
+    # (None) a plánovač pak neviděl zátěž auta — bug 2026-06-13.
     td = _parse_json_dt(obj.get("target_deadline"))
-    if td is None:
-        return None
     return SimpleNamespace(
         target_deadline=td,
-        energy_needed_wh=float(obj["energy_needed_wh"]),
+        energy_needed_wh=float(obj.get("energy_needed_wh") or 0.0),
         headroom_wh=float(obj.get("headroom_wh") or 0.0),
         opportunistic_value_czk_kwh=float(obj.get("opportunistic_value_czk_kwh") or 0.0),
     )

backend/tests/test_ev_session_parse.py

@@ -0,0 +1,66 @@
+"""Parser EV session z fn_planning_site_context (_ev_session_from_json).
+
+Bug 2026-06-13: session BEZ deadline (auto nad targetem / bez cíle) se v
+parseru zahazovala (None), takže plánovač neviděl zátěž auta ani oportunismus.
+Oprava: session bez deadline zůstává objektem s energy_needed_wh=0 a headroom.
+"""
+
+import unittest
+
+from services.planning.db_io import _ev_session_from_json
+
+
+class EvSessionParseTests(unittest.TestCase):
+    def test_none_and_empty_return_none(self) -> None:
+        self.assertIsNone(_ev_session_from_json(None))
+        self.assertIsNone(_ev_session_from_json([]))
+        self.assertIsNone(_ev_session_from_json(123))
+
+    def test_session_without_deadline_kept_for_opportunism(self) -> None:
+        sess = _ev_session_from_json(
+            {
+                "target_deadline": None,
+                "energy_needed_wh": 0,
+                "headroom_wh": 18000.0,
+                "opportunistic_value_czk_kwh": 1.0,
+            }
+        )
+        self.assertIsNotNone(sess)
+        assert sess is not None
+        self.assertIsNone(sess.target_deadline)
+        self.assertEqual(sess.energy_needed_wh, 0.0)
+        self.assertEqual(sess.headroom_wh, 18000.0)
+        self.assertEqual(sess.opportunistic_value_czk_kwh, 1.0)
+
+    def test_session_with_deadline_and_need(self) -> None:
+        sess = _ev_session_from_json(
+            {
+                "target_deadline": "2026-06-14T05:00:00+00:00",
+                "energy_needed_wh": 12000.0,
+                "headroom_wh": 6000.0,
+                "opportunistic_value_czk_kwh": 1.0,
+            }
+        )
+        assert sess is not None
+        self.assertIsNotNone(sess.target_deadline)
+        self.assertEqual(sess.energy_needed_wh, 12000.0)
+
+    def test_missing_needed_defaults_zero(self) -> None:
+        sess = _ev_session_from_json(
+            {"target_deadline": None, "headroom_wh": 1000.0}
+        )
+        assert sess is not None
+        self.assertEqual(sess.energy_needed_wh, 0.0)
+        self.assertEqual(sess.opportunistic_value_czk_kwh, 0.0)
+
+    def test_json_string_payload(self) -> None:
+        sess = _ev_session_from_json(
+            '{"target_deadline": null, "energy_needed_wh": 0, '
+            '"headroom_wh": 5000, "opportunistic_value_czk_kwh": 1.0}'
+        )
+        assert sess is not None
+        self.assertEqual(sess.headroom_wh, 5000.0)
+
+
+if __name__ == "__main__":
+    unittest.main()

backend/tests/test_ev_write_on_change.py

@@ -1,11 +1,13 @@
-"""Write-on-change pro TeltoCharge: zápis JEN při skutečné změně hodnoty.
+"""TeltoCharge zápis: reg 15 (amps) VŽDY, watchdog 19/20 write-on-change.
 
-Regrese na opakované zápisy do wallboxu (EEPROM wear): export tick běží
-~8x/hod (control_export :14,:29,:44,:59 + rolling replan */15 s exportem),
-ale reg 15/19/20 se smí zapsat jen při změně proti poslednímu known stavu
-zařízení (fn_modbus_device_state_map: nejnovější written/verified řádek
-journalu per registr). Watchdog (19/20) se zapíše jednou po startu / po
-výpadku; sytí ho i FC3 čtení telemetrie (60 s), periodické zápisy netřeba.
+Export tick běží ~8x/hod (control_export :14,:29,:44,:59 + rolling replan
+*/15 s exportem). **Reg 15 (amps to use) se zapisuje VŽDY** — TeltoCharge ho
+po výpadku komunikace sám přepíše na failsafe (reg 20) bez journal řádku, a
+kdyby byl write-on-change, EMS by tichý drift 0 → 8 A nikdy nezahlédlo
+(verify čte zpět jen `written`). **Reg 19/20 (watchdog config, EEPROM wear)
+zůstávají write-on-change** proti fn_modbus_device_state_map (nejnovější
+written/verified řádek per registr): zapíší se jednou po startu / po výpadku;
+sytí je i FC3 čtení telemetrie (60 s), periodické zápisy netřeba.
 """
 
 import unittest
@@ -20,6 +22,7 @@ from services.control.outputs import (
     TELTO_REG_FAILSAFE_CURRENT_A,
     TELTO_WATCHDOG_FAILSAFE_A,
     TELTO_WATCHDOG_TIMEOUT_S,
+    _split_amps_and_watchdog,
     _telto_setpoint_registers,
     write_ev_arrival_hold,
     write_ev_setpoints,
@@ -59,40 +62,48 @@ class TeltoSetpointRegistersTests(unittest.TestCase):
         self.assertEqual(_telto_setpoint_registers(6)[0][2], 6)
         self.assertEqual(_telto_setpoint_registers(40)[0][2], 32)
 
+    def test_per_charger_failsafe_and_timeout(self) -> None:
+        regs = _telto_setpoint_registers(0, comm_timeout_s=120, failsafe_a=6)
+        self.assertEqual([(r, v) for r, _, v in regs], [(15, 0), (19, 120), (20, 6)])
+
+    def test_failsafe_clamped_to_0_32(self) -> None:
+        self.assertEqual(_telto_setpoint_registers(0, failsafe_a=99)[2][2], 32)
+        self.assertEqual(_telto_setpoint_registers(0, failsafe_a=-5)[2][2], 0)
+
+    def test_split_separates_amps_from_watchdog(self) -> None:
+        amps, watchdog = _split_amps_and_watchdog(_telto_setpoint_registers(0))
+        self.assertEqual([r for r, _, _ in amps], [15])
+        self.assertEqual([r for r, _, _ in watchdog], [19, 20])
+
 
 class DropAgainstDeviceStateTests(unittest.TestCase):
-    def test_steady_state_drops_everything(self) -> None:
+    def test_watchdog_steady_state_drops_19_20(self) -> None:
+        _, watchdog = _split_amps_and_watchdog(_telto_setpoint_registers(0))
         out, skipped = _drop_registers_matching_last_verified(
-            _telto_setpoint_registers(0), _STEADY_STATE_0A
+            watchdog, _STEADY_STATE_0A
         )
         self.assertEqual(out, [])
-        self.assertEqual(skipped, [15, 19, 20])
-
-    def test_amps_change_writes_only_reg_15(self) -> None:
-        out, skipped = _drop_registers_matching_last_verified(
-            _telto_setpoint_registers(16), _STEADY_STATE_0A
-        )
-        self.assertEqual([(r, v) for r, _, v in out], [(15, 16)])
         self.assertEqual(skipped, [19, 20])
 
-    def test_empty_state_after_outage_writes_full_triple(self) -> None:
-        out, skipped = _drop_registers_matching_last_verified(
-            _telto_setpoint_registers(0), {}
-        )
-        self.assertEqual([r for r, _, _ in out], [15, 19, 20])
+    def test_empty_state_after_outage_keeps_19_20(self) -> None:
+        _, watchdog = _split_amps_and_watchdog(_telto_setpoint_registers(0))
+        out, skipped = _drop_registers_matching_last_verified(watchdog, {})
+        self.assertEqual([r for r, _, _ in out], [19, 20])
         self.assertEqual(skipped, [])
 
 
 class _FakeDB:
     """Jen řádky chargeru; journal funkce se patchují v modbus_journal."""
 
-    def __init__(self) -> None:
+    def __init__(self, failsafe_a: int = 8, comm_timeout_s: int = 300) -> None:
         self.row = {
             "asset_id": 7,
             "code": "ev-charger-1",
             "host": "172.16.1.16",
             "port": 502,
             "unit_id": 1,
+            "watchdog_failsafe_a": failsafe_a,
+            "watchdog_comm_timeout_s": comm_timeout_s,
         }
 
     async def fetch(self, query: str, *args: object) -> list[dict]:
@@ -105,9 +116,9 @@ class _FakeDB:
         raise AssertionError(f"unexpected fetchval: {query}")
 
 
-class WriteEvSetpointsOnChangeTests(unittest.IsolatedAsyncioTestCase):
+class WriteEvSetpointsTests(unittest.IsolatedAsyncioTestCase):
     async def _run(
-        self, device_state: dict[int, int], ev1_a: int
+        self, device_state: dict[int, int], ev1_a: int, db: _FakeDB | None = None
     ) -> tuple[AsyncMock, AsyncMock]:
         create = AsyncMock(return_value=[1, 2, 3])
         execute = AsyncMock(return_value=True)
@@ -120,13 +131,17 @@ class WriteEvSetpointsOnChangeTests(unittest.IsolatedAsyncioTestCase):
             patch.object(journal, "create_modbus_commands", create),
             patch.object(journal, "execute_modbus_commands", execute),
         ):
-            await write_ev_setpoints(1, _setpoints(ev1_a), _FakeDB())  # type: ignore[arg-type]
+            await write_ev_setpoints(1, _setpoints(ev1_a), db or _FakeDB())  # type: ignore[arg-type]
         return create, execute
 
-    async def test_steady_state_no_write_at_all(self) -> None:
+    async def test_steady_state_still_reasserts_reg_15(self) -> None:
+        # Reg 15 se zapisuje VŽDY (re-asert proti tichému failsafe driftu),
+        # i když je device-state mapa shodná. Watchdog 19/20 se přeskočí.
         create, execute = await self._run(_STEADY_STATE_0A, ev1_a=0)
-        create.assert_not_awaited()
-        execute.assert_not_awaited()
+        create.assert_awaited_once()
+        registers = create.await_args.args[8]
+        self.assertEqual([(r, v) for r, _, v in registers], [(15, 0)])
+        execute.assert_awaited_once()
 
     async def test_plan_change_writes_only_amps(self) -> None:
         create, execute = await self._run(_STEADY_STATE_0A, ev1_a=16)
@@ -135,14 +150,24 @@ class WriteEvSetpointsOnChangeTests(unittest.IsolatedAsyncioTestCase):
         self.assertEqual([(r, v) for r, _, v in registers], [(15, 16)])
         execute.assert_awaited_once()
 
-    async def test_after_outage_writes_full_triple(self) -> None:
+    async def test_after_outage_writes_amps_then_watchdog(self) -> None:
         create, execute = await self._run({}, ev1_a=0)
         registers = create.await_args.args[8]
         self.assertEqual([r for r, _, _ in registers], [15, 19, 20])
         execute.assert_awaited_once()
 
+    async def test_per_charger_failsafe_from_db(self) -> None:
+        # Failsafe 6 A z DB → po výpadku se zapíše reg 20 = 6 (prázdná mapa).
+        create, _ = await self._run(
+            {}, ev1_a=0, db=_FakeDB(failsafe_a=6, comm_timeout_s=120)
+        )
+        registers = create.await_args.args[8]
+        self.assertEqual(
+            [(r, v) for r, _, v in registers], [(15, 0), (19, 120), (20, 6)]
+        )
 
-class WriteEvArrivalHoldOnChangeTests(unittest.IsolatedAsyncioTestCase):
+
+class WriteEvArrivalHoldTests(unittest.IsolatedAsyncioTestCase):
     async def _run(
         self, device_state: dict[int, int]
     ) -> tuple[bool, AsyncMock, AsyncMock]:
@@ -160,13 +185,15 @@ class WriteEvArrivalHoldOnChangeTests(unittest.IsolatedAsyncioTestCase):
             ok = await write_ev_arrival_hold(1, "ev-charger-1", _FakeDB())  # type: ignore[arg-type]
         return ok, create, execute
 
-    async def test_hold_skips_when_device_already_at_zero(self) -> None:
+    async def test_hold_always_writes_reg_15_even_if_device_at_zero(self) -> None:
+        # Tvrdé zastavení po píchnutí kabelu — reg 15 = 0 se zapíše VŽDY.
         ok, create, execute = await self._run(_STEADY_STATE_0A)
         self.assertTrue(ok)
-        create.assert_not_awaited()
-        execute.assert_not_awaited()
+        registers = create.await_args.args[8]
+        self.assertEqual([(r, v) for r, _, v in registers], [(15, 0)])
+        execute.assert_awaited_once()
 
-    async def test_hold_writes_only_amps_when_watchdog_already_set(self) -> None:
+    async def test_hold_writes_amps_and_watchdog_when_device_drifted(self) -> None:
         ok, create, execute = await self._run(
             {
                 TELTO_REG_AMPS_TO_USE: 16,
@@ -176,6 +203,7 @@ class WriteEvArrivalHoldOnChangeTests(unittest.IsolatedAsyncioTestCase):
         )
         self.assertTrue(ok)
         registers = create.await_args.args[8]
+        # Reg 15 = 0 zapsán (i když device hlásí 16); 19/20 shodné → skip.
         self.assertEqual([(r, v) for r, _, v in registers], [(15, 0)])
         execute.assert_awaited_once()