On KeyboardInterrupt, brewpi.py cancels all asyncio tasks (letting
HeaterTask's `with device.open()` finally block run), then calls
heater.close() as belt-and-suspenders before closing the event loop.
On shutdown HeaterHendi.activate(False) now sets the switch to 0 instead
of disabling remote control, so the Hendi stays in remote mode at 0 W
rather than dropping back to its local panel state. HendiCtrl.close()
does the same (setSwitch(0)) then closes the serial port.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds a configurable IIR pre-filter (beta, default 0.05, τ≈19 s at dt=1 s)
applied to theta_ist before the derivative, so stirrer-induced temperature
noise is attenuated before being amplified by differentiation rather than
after. The shared filter/derivative/post-filter logic is factored into
TempControllerBase._compute_heatrate(); both Normal and Smith subclasses
call it, replacing their duplicated inline rate blocks.
beta is read from TempCtrl.beta in the config (default 0.05 if absent) and
added to config-sim.json.tpl and config-real.json.tpl.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replaces the open-loop (feed real temp_ist) approach with a proper
closed-loop simulation: a Pot plant model is driven by the controller's
power output each tick, generating sim_temp_ist that diverges from the
real log's real_temp_ist whenever the PID params produce different
behaviour. Temperature panel now shows real_temp_ist vs sim_temp_ist
vs temp_soll so the effect of candidate gains is immediately visible.
Plant params (M, C, L, Td) default to 20 kg water / L=0.2 / Td=30 s
and can be overridden via --plant-* flags. Ambient defaults from config.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EY449HByk6UpKhDN2HTnfK
Feeds real temp_ist/temp_soll readings from a run log back into a fresh
TempController (Normal mode) tick by tick, then plots a 3-panel comparison
of heat rate and normalized power against the original logged values. Any
PID gain can be overridden via --<section>-<gain> flags so candidate params
can be evaluated against a past run without touching hardware.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EY449HByk6UpKhDN2HTnfK
Documents the stirring-induced rate noise problem, its root cause
(post-differentiation filtering), and three mitigation options: pre-filter
before differentiating, sliding-window regression, and model-based rate
from Pot.p_pot (Smith only).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NSo8R6GjoBQdStB3j67zSs
Renders the same two plots the PyQt GUI shows (realtime 3-panel strip
chart and forecast-vs-actual comparison) from a log/forecast JSON pair,
specified by date_time and sud_name on the command line.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NSo8R6GjoBQdStB3j67zSs
pyserial and spidev are Linux/Pi-only (spidev requires the kernel SPI
driver and won't install on macOS/Windows). Dev machines use
requirements.txt; the Pi uses requirements-hw.txt which pulls in the
base file via -r.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
HoldRemaining now calls updateHeaderCountdown() directly so the header
hold row refreshes every second from its own message, not only via the
Elapsed → updateStepPlates → updateHeaderCountdown chain.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
- Power soll label centered below flame area (amber, clears when off)
- Stirrer RPM label next to motor housing (clears when stopped)
- Flame base moved 20 units closer to pot bottom
- Drop Ist/Soll LCD header row to save vertical space
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
Catch ConnectionClosed in User.send() and handler_recv() so that an
unclean client disconnect no longer produces "Task exception was never
retrieved" / "connection handler failed" log noise.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
- Add viewport meta tag so Safari uses device CSS pixels
- Move LCD units to a dedicated right-hand column; strip them from labels
- Tighten LCD grid: 4-column layout, narrower gap, font-size 1.9em
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
- HeaterTask.shutdown() on DONE/IDLE: document the _on_end callback,
power-zero, TC disable, and ClosedLoop:false broadcast to clients.
- Resume with inherited temperature: document _paused_temp_soll fallback
for steps that carry no temperature of their own.
- Browser client additional features: update Header countdown (now three
rows: total/ramp/hold, always h:mm:ss) and describe the step-plate
Ramp/Hold green countdown and configured hold-duration field.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
Each step plate now shows the scheduled hold time ("Hold h:mm:ss") as a
static field instead of the live actual temperature. The value is set once
at plate-creation time from the step's hold.duration; steps without a hold
phase show "Hold —".
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
Step plates now show "Ramp h:mm:ss" (ramp time only) during ramp phase
and "Hold h:mm:ss" (hold remaining) during hold phase, both in the same
digit area in green (#4ade80). Inactive steps show planned ramp duration.
Header countdown gains a "hold" row alongside the existing "ramp" (renamed
from "step") and "total" rows. All three use the same h:mm:ss format.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
Adds HeaterTask.shutdown() which switches to open-loop at power 0,
disables TC, and broadcasts ClosedLoop=false to clients so the browser
checkbox unchecks immediately. SudTask fires it via a new set_on_end()
callback whenever state reaches DONE or IDLE, giving both a completed
schedule and a manual Stop the same safe behaviour: stirrer and heater
both off, no heater left running unattended.
Re-enabling after shutdown: check the closed-loop checkbox (sends
ClosedLoop=true → TC re-enabled) or adjust the power slider directly;
the next brew's Play press already sends ClosedLoop=true automatically.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
Total remaining now appears above step remaining in the header. Both
countdowns and all other duration displays always use h:mm:ss format
regardless of whether hours is zero, so the layout never shifts width.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
Steps with no 'temperature' field inherit their target from a prior step
(temperature=None in the resolved schedule). The resume-from-pause code
only called set_theta_soll when the current step had an explicit temperature,
so any manual setpoint change made during pause was never reverted for
inherited steps (e.g. sud_0010 steps 2-3 both hold at step 1's 57 °C).
Fix: save theta_soll_set at pause time; use it as fallback on resume when
the current step's temperature is None. Also set initialTempSollSync=true
in the browser on play so the slider accepts the next server Soll push
regardless of message arrival order relative to the State update.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
- Rename "State replay on connect" to "State is server-only; GUIs only
reflect it" and expand it with the core design principle: neither GUI
client ever maintains its own model of server state — controls update
only on server push, never speculatively on user input. Multi-client
sync, stale-overwrite safety, and reconnect correctness all follow from
this single rule.
- Update TC-enable description: HeaterTask now owns the TC's enabled
switch via the closed_loop flag; SudTask no longer touches it.
- Add "Heater control modes" subsection describing Closed-loop/Open-loop,
the enable/disable rules per mode and sud state, the auto-reset on play,
and the schedule parameter restore on resume.
- Refresh the Browser client section: document the Heater control mode
checkbox, header countdown timers, and the always-reflect-server-values
slider policy; call out the server-only-state principle explicitly.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
On resume from PAUSED, on_step_changed does not re-fire (same step
continues), so any manual TC/stirrer adjustments made while paused
would persist into the run. Explicitly re-apply the current step's
temperature, heatrate, and stirrer config on the resume path in
SudTask.recv().
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
Closed-loop (default, checkbox checked): TC has exclusive control over
the heater; temp-soll and heatrate-soll sliders are active; direct
heater power entry is disabled. TC now stays enabled at server startup
and whenever closed-loop mode is active — SudTask no longer disables it
on IDLE/DONE, so adjusting temp-soll manually with no sud running
immediately drives the heater via the TC.
Open-loop (checkbox unchecked, only selectable when paused/stopped):
TC is disconnected from the heater; only the heater-power slider is
active. Switching back to closed-loop resets power_soll to 0 and
re-enables the TC.
On play, the mode is automatically reset to closed-loop.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
Shows remaining time for the current step and the entire sud in large
accent-coloured monospace text to the right of the transport buttons,
matching the per-step plate colour. Hidden until a step is active.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
Remove one-time sync guards (initialPowerSync, initialSpeedSync) so heater
power and stirrer speed sliders track the server's ist readbacks on every
push, not just on first connect. Also remove the zero-out on sud
stop/pause: with continuous ist tracking the animation and slider always
agree, and the zero-out caused them to diverge when the server kept sending
non-zero values during a pause.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SH2D4LRuCnhLSzoYerAfJX
- Remove tab system (nav#tabs, tab-btn, tab-content) from HTML/CSS/JS
- Permanent 3-column grid: manual controls (left, 16em) | LCD+pot (center) | progress (right, fills remaining)
- LCD panel stacks above pot SVG in center column, both hidden until connected
- Step plates no longer max-width capped, fill the right column
- Slider labels moved above the slider+readout row via flex-direction: column
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DF16nV2ispNw1C9SbvqbRQ
Mirrors PyQt client which seeds doubleSpinBox_pot_temp from the
remembered ambient on startup. In the browser there is no persistence,
so the first AmbientTemp message from the server is used instead.
Flag resets on reconnect so it re-syncs if the server restarts.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DF16nV2ispNw1C9SbvqbRQ
HeaterTask.on_changed_active() was never registered as a callback so
never fired; TcTask.recv() no longer needs to handle {Enable: ...}
since no client sends it after the override checkbox removal.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DF16nV2ispNw1C9SbvqbRQ
During a running sud, all four manual sliders (temp soll, heat rate,
heater power, stirrer speed) are disabled for user input but update
their positions continuously from server pushes. They re-enable when
the sud stops or is paused, or on disconnect. Browser client also
resets initial-sync flags on reconnect so sliders re-sync correctly.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DF16nV2ispNw1C9SbvqbRQ
Removing the Enabled else-if branch in a previous commit left the Soll
if-block unclosed, crashing app.js at parse time so no JS ran at all.
Also adds: onerror handler surfaces connection failures in the status
label; try/catch for invalid URL; 'Connecting…' feedback; localhost
fallback when location.hostname is empty (file:// open); wider host
input field.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DF16nV2ispNw1C9SbvqbRQ
Drops the tc-enable, heater-activate and stirrer-activate checkboxes
and all associated state, signal connections, initial-sync flags, and
channel-handler logic. Regenerates main_window.py from the updated .ui.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DF16nV2ispNw1C9SbvqbRQ
Replace clip-rect height manipulation with translate-scale-translate SVG
transform on fire-group, pivoting at the fixed base coordinate y=248.
Eliminates base-hopping caused by fill-box bounding-box drift from
individual flame flicker animations. Clip rect is now static (guards
against flicker overshoot above the pot bottom). Minimum visible flame
size floored at the scale corresponding to 1200W.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JLR2133M2PyjiYq584i2G7
sudStepIndex was reset to null in updateSudForecast whenever
isNewSchedule=true (i.e. on every fresh page load). If the Step
message arrived before Json in the global-state replay, or in a
separate message, the schedule would be rebuilt but the index
would be null, leaving grain mass as 0 with no subsequent Step
to restore it.
Step messages are the authoritative source for the current index,
so the reset is removed. The Step handler always overwrites it.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JLR2133M2PyjiYq584i2G7
Water fill height now reflects water_mass + grain_mass*0.7 (grain
displacement). A small blue tick and label on the left pot wall shows
the utilized volume in litres, hidden when the pot is unconfigured.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JLR2133M2PyjiYq584i2G7
- Stirrer starts at speed 0 (was 50) so blades don't spin on connect
- Grain particles only levitate when stirrer is running
- Fire size driven by PowerSet (soll) instead of Power (ist)
- Split animation-play-state out of stirrer shorthand for reliable pausing
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JLR2133M2PyjiYq584i2G7
Replace scaleY group transform with a dynamic clipPath that reveals
flame height proportional to heater power. Base at y=256 stays fixed;
only the tips grow with intensity. Includes flicker overshoot buffer.
Fire paths shifted +8px down, viewBox extended to 265.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DGKe8JzTM2Gr1fdx9wYXTi
35 particles distributed across full pot height, clipped to dynamic
water area via water-clip-rect updated alongside water fill level.
Particle opacity transitions on grain_mass changes per step.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DGKe8JzTM2Gr1fdx9wYXTi
- SVG fire group with 3 layered flame paths (red/orange/yellow gradients)
- CSS keyframe flicker animations with different timings per flame layer
- JS controls fire opacity and scaleY proportional to heater power
- Fix buildStep() reading grain_mass/water_mass from nested step.pot
instead of flat step keys, matching the JSON structure
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DGKe8JzTM2Gr1fdx9wYXTi
- Add volumen field to all sud JSON pot entries (30 L)
- Fix parseSudDoc to read from nested doc.pot (was using stale flat keys)
- SVG widget shows transparent rectangular pot with stirrer, thermometer
with live temperature, and water fill level (water_mass/volumen) when
a sud is loaded; no water shown when no sud is loaded
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DGKe8JzTM2Gr1fdx9wYXTi
Update set_step(), status bar mass display, and on_action_sud_new() to
read grain_mass/water_mass from step['pot'] and use the new pot key
structure when sending an empty Load document.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01T52JH848ojhXXHn1bAzdC3
In estimate()'s on_step_changed closure, the new 'pot = step.get("pot", {})'
local variable shadowed the outer 'pot' (Pot plant instance), causing
pot.set_plant_params() to be called on a plain dict and crash on load.
Renamed to 'step_pot' to avoid the collision.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01T52JH848ojhXXHn1bAzdC3
Top-level pot_mass/pot_material/L/Td/grain_mass/water_mass and their
per-step overrides are now grouped under a 'pot' sub-object, matching
the structure already applied to sud_0010.json's parent level. Updated
all sude/*.json files and the parsing/consuming code in components/sud.py,
tasks/sud.py, components/sud_forecast.py, and scripts/demos/sud/demo_sud.py.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01T52JH848ojhXXHn1bAzdC3
- Browser: OK button sends Confirm without closing dialog; Escape blocked;
dialog closes when state leaves WAIT_USER (server-driven).
- PyQt: non-blocking QDialog replaces blocking QMessageBox; OK sends Confirm
without closing; dialog closed by server state change or on disconnect.
- ws_client.py: remove deprecated loop= params from websockets.connect() and
asyncio.ensure_future() (same fix already applied to ws_server.py) so the
PyQt client can actually connect with websockets >= 10.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Tqxrk8uj4M3w3d3eXm3xK8
With TempControllerBase(APid, TempControllerFsm), APid.is_holding()
(abstract stub returning None) shadowed TempControllerFsm.is_holding()
in the MRO, so tc.is_holding() always returned None. The forecast
simulation never called sud.temp_reached() and always hit MAX_TICKS.
Swapping to TempControllerBase(TempControllerFsm, APid) puts the
concrete FSM methods before APid's abstract stubs in the MRO.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Tqxrk8uj4M3w3d3eXm3xK8
States, DEFAULT_THRESHOLDS, the three PIDs, FSM state vars, process_fsm(),
on_state_entered(), set_enabled(), and is_holding() move to a new
TempControllerFsm class in temp_controller_fsm.py. TempControllerBase
inherits from both APid and TempControllerFsm; States is re-exported so
existing imports in temp_controller_smith.py are unaffected.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Tqxrk8uj4M3w3d3eXm3xK8
All four name-dispatch factories repeated the same if/elif lazy-import
pattern. Replace with a ComponentFactory base class holding a callable
registry and a _lazy() helper for the common import-and-construct case.
PlantFactory is an orchestrator, not a dispatcher, and is left as-is.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Tqxrk8uj4M3w3d3eXm3xK8
Callbacks were registered inside the async on_process() coroutine,
meaning they weren't active until the event loop started ticking.
Moving them to __init__ ensures they're wired up at construction time.
Also removes remaining debug prints from the affected task files.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Tqxrk8uj4M3w3d3eXm3xK8
Both handler loops caught all exceptions, printed e, and broke — hiding
errors entirely. Let exceptions propagate naturally; the parent handler's
FIRST_COMPLETED logic already handles task cleanup on disconnect.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Tqxrk8uj4M3w3d3eXm3xK8
Same pattern as set_model_power: these methods only existed on the Smith
controller, forcing five hasattr guards at call sites. No-op stubs on the
base class give all controllers a stable interface; Smith overrides them.
Remove the now-redundant hasattr guards.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Tqxrk8uj4M3w3d3eXm3xK8
The Normal controller had no set_model_power(), causing callers to
guard with hasattr() instead of relying on a stable interface. Adding a
no-op stub to the base class means all controllers have the method;
TempControllerSmith overrides it to actually update its internal model.
Remove the now-redundant hasattr guards at both call sites.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Tqxrk8uj4M3w3d3eXm3xK8
Notes Mainsail (Vue 3 + TypeScript + Vuetify, uPlot for live charts) as
the reference point for a future rewrite of app.js's hand-rolled DOM
wiring and the still-missing Automatic/Plot tab charts.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0189FkmyJkY77HqsNomr1DwV
Mirrors client/brewpi_gui.py's actionPause: btn-sud-pause sends
{"Sud": {"Pause": true}} and is only enabled while actually running
(not idle, not already paused), matching update_sud_actions()'s
enabled/disabled logic - Start already doubles as Resume while paused,
so no separate Resume control is needed.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0189FkmyJkY77HqsNomr1DwV
New/Load/Save/Start/Stop now live in #sud-controls inside
#connection-bar, visible regardless of which tab is active, instead of
a standalone panel above the tabs. Drops the now-unused .button-row
rule along with it.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0189FkmyJkY77HqsNomr1DwV
Ports show_user_message() from client/brewpi_gui.py: a plain <dialog>
pops with the step's user_message the instant WAIT_USER is newly
entered (tracked via a prevState comparison in onSudChanged()'s State
branch), and sends {"Sud": {"Confirm": true}} on the dialog's close
event - covers both the OK button and Escape dismissal.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0189FkmyJkY77HqsNomr1DwV
Ports on_action_sud_new/_save/_load() and on_action_sud_start/_stop()
from client/brewpi_gui.py: Load/Save stand in for the native file
dialogs with <input type="file">/FileReader and a Blob/<a download>,
New sends the same hardcoded empty-doc Load, and Start/Stop are gated
by updateSudActions() mirroring update_sud_actions()'s enabled/disabled
logic. Also fixes the status line's step number being one below what
the Progress tab's plates show for the same step (same bug just fixed
in the desktop client).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0189FkmyJkY77HqsNomr1DwV
Re-enabling the temp controller landed unconditionally in HOLD and
relied on the next tick to correct it, but SudTask.on_process() reads
is_holding() synchronously in the same call chain that pushes a fresh
step's setpoint - so a restart could finish a hold-less step instantly
instead of actually ramping. Resolve the FSM against the real gap right
away instead. Also fix the status bar showing one step number below
what the Progress tab's plates show for the same step.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0189FkmyJkY77HqsNomr1DwV
A new HTML/CSS/JS client, added next to client/brewpi_gui.py rather
than replacing it, speaking the exact same WebSocket pub/sub protocol
- no server-side protocol changes needed. server/brewpi.py gains a
--http-port (default 8080) stdlib http.server.ThreadingHTTPServer,
in its own daemon thread and deliberately decoupled from the existing
asyncio WebSocket server, serving web/ statically.
v1 covers the Manual tab (direct heater/stirrer/controller control)
and the new Progress tab - the two most actionable surfaces, neither
needing a charting library. web/app.js ports the relevant logic from
brewpi_gui.py directly: components/sud.py's _build_step()/
_merge_defaults() for resolving the raw schedule doc, and the
StepPlate/_update_step_plates()/update_status_step_label() logic
behind the Progress tab and status line.
Deliberately deferred rather than stubbed (see web/TODO.md for the
full parity backlog against the Qt GUI): Sud control actions (Start/
Pause/Stop/Confirm), the Automatic tab's forecast plot and the Plot
tab's live strip charts, Sud file management, and auth (matching the
WebSocket server's own existing lack of it).
No new Python dependencies - functools/threading/http.server are all
stdlib.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019qvu5giu7gvRCyEWzf2Vpx
sud_elapsed_seconds is intentionally reset to None once a run leaves
a running state (it gates the forecast plot's dynamic-vs-static view
switch) - but the status bar/Progress tab reused it too, so they lost
the final total exactly when it mattered most: right as the run
finished. Added sud_elapsed_last, which only ever gets overwritten by
a real 'Elapsed' push and is reset on a fresh Start/Load, for those
two displays to use instead.
Separately, a fresh Start (restarting an already-loaded, now IDLE/
DONE schedule) never sent a new Load, so SudTask's energy bookkeeping
- previously only reset in the Load handler - kept showing the
previous run's banked totals until each step was revisited. Now also
reset on a fresh Start (not a Pause->resume, which should keep it).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019qvu5giu7gvRCyEWzf2Vpx
The "Forecast vs. actual duration" section still described the old
confirm_points/_continue_forecast_after_confirm() mechanism and
claimed the forecast was deliberately not re-anchored mid-run - both
no longer true since _reanchor_forecast() now fires on every step
boundary. Replaced with the actual current mechanism (the reanchor
itself, the _forecast_generation race guard, forecast_step_starts/
StepStarts, and the Save-while-running fix on both server and
client), added a new "Progress tab" section (LED semantics, per-step
fields, energy integration/banking, status-bar sums), and documented
--dt/--sim-warp-factor replacing config.json's Controller.dt/
sim_warp_factor.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019qvu5giu7gvRCyEWzf2Vpx
Elapsed is just Sud's own tick-counted total (every step's time,
WAIT_USER dwell included, already adds up sequentially with no gaps).
Energy has no equivalent single running counter server-side (it's
banked per step, not accumulated as one grand total - see tasks/
sud.py's SudTask) - summed here instead from every finished step's
own Wh total plus the active step's running one, then converted to
kWh.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019qvu5giu7gvRCyEWzf2Vpx
Both are run-mode knobs (how fast to drive this particular process),
not config.json material (a plant/hardware description that stays the
same regardless of how a run happens to be invoked) - --dt and
--sim-warp-factor, defaulting to 1.0 each (real time, 1-simulated-
second ticks) unless a dev/test run asks for a faster warp explicitly.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019qvu5giu7gvRCyEWzf2Vpx
Each StepPlate on the Progress tab now shows a step's energy use in
Wh - live and growing while it's the active step, frozen at its final
total once finished, blank before it's ever been reached.
Integrated server-side from the heater's own live effective power
(pot.get_power(), already fed from heater.power_eff - see server/
brewpi.py's wiring), since actual energy used can't be predicted from
the forecast like the timing/temperature fields are, only measured as
it happens. Banked into energy_by_step on every genuine step
transition (on_step_changed(), keyed off the same index-change check
used for the ramp->hold phase switch), including the final transition
to DONE; reset on every fresh Load.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019qvu5giu7gvRCyEWzf2Vpx
Reconnecting to (or any client-side Save fetch of) an already-running
Sud showed the wrong step active and bogus remaining times, traced to
two independent bugs:
- recv()'s 'Save' handler always called send_forecast(doc), which
resimulates the *entire* schedule cold from step 0 - fine for a
not-yet-started Sud, but for one already running it silently
overwrote the forecast/forecast_step_starts that _reanchor_forecast()
had been accurately, continuously maintaining with a context-free
"starting now" guess. Now skipped whenever a run is already in
progress, re-sending what's already there instead.
- _reanchor_forecast() recomputes each step's real start time
asynchronously, so that recomputation can itself be (and routinely
is) superseded and discarded by a later transition's reanchor before
ever committing - and the next *successful* commit's "everything
before my index is real" filter would then preserve whatever stale
prediction was sitting there before, sometimes for a step that
hadn't even happened yet by the schedule's real position.
on_step_changed() now also records each step's start synchronously,
immune to that race.
- Client-side, update_sud_forecast() unconditionally reset sud_step_
index/forecast_step_starts/etc. on every 'Json' push - correct for
an actual Load, but connect()'s unconditional Save request produces
a standalone 'Json' push (no Step/State alongside it, unlike the
subscribe-triggered replay) for the *same* already-running schedule,
wiping the correct state with nothing left to restore it. Now only
resets when the incoming doc actually differs from the last one
processed.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019qvu5giu7gvRCyEWzf2Vpx
A glance at the Automatic tab's forecast plot couldn't show where the
brew actually stands within its own schedule - just a curve. The new
Progress tab stacks one StepPlate per step instead: an LED for its
status (off/done/active-ramping/active-holding), its resolved target
temp, masses and ramp rate, plus, for whichever step is currently
active, live actual temp and stirrer state (frozen at their last
value once that step finishes) and a forecast-based remaining-time
countdown.
The countdown needs to know where each step actually begins in the
forecast's timeline, which the server didn't track before -
SudForecastEstimator.estimate() now returns per-step start times
alongside t/theta, and SudTask threads them through send_forecast()/
_reanchor_forecast() the same way (including the same generation-guard
against the concurrent-call race) as a new 'StepStarts' field on the
'Forecast' message.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019qvu5giu7gvRCyEWzf2Vpx
jens
