Documents why a brew commonly takes ~250-270 min against a ~160 min
static estimate: the PID cascade needs real time to spin up and settle
within the tight 0.2°C "reached" tolerance, a roughly per-step-constant
lag that accumulates across every ramp/hold transition. Found and
confirmed by a live measured run on the real server.
On DONE/Stop, the forecast used to revert to show_schedule() seeded
with the *current* temperature - which by then is wherever the brew
ended up (e.g. near the last step's target), so walking the whole
original schedule again from there produced a nonsensical total
("152 min" for a brew that actually took ~250). Now it just leaves the
dynamic view's last frame (the real measured history) up instead.
Recomputing from the current temperature is still correct for a fresh
Load, where no run has happened yet - distinguished via whether
sud_start_time was set before this transition.
process_pid() was advancing pid_rate and pid_cool unconditionally on
every tick regardless of which one actually drove y. While COOL had
pid_cool driving, pid_rate kept silently integrating against the same
(very negative) heatrate_err with no way to act on it, building a
large stale windup. The moment the FSM returned to HOLD (which only
reset pid_hold), y read straight from that saturated pid_rate instead
of a fresh value, so the heater stayed off well past the target
before the windup finally unwound - an undershoot of about 1 degree
in scripts/demos/pid/demo_temp_controller.py's last (cooling) step.
Now only the PID currently driving y is process()'d each tick, and
pid_rate is also reset on IDLE->HOLD and COOL->HOLD (it was already
reset on the transitions into HEAT/COOL). Undershoot drops to ~0.06
degrees, in line with normal PID settling.
TempControllerBase.cooling/set_cooling() forced y=0 in advance of a
Sud ramp-down step; redundant now that the FSM's own COOL state
detects and handles a negative diff itself. Removed from the
controller, SudTask's anticipatory detection (and the Step message's
now-unused "Cooling" field), and demo_sud.py's mirror of it.
gui: the "Cool down" status-bar indicator now reflects the
controller's actual State ("States.COOL") via the TempCtrl channel,
rather than Sud's removed anticipatory flag - renamed
sud_cooling/set_sud_cooling to tc_cooling/set_tc_cooling to match.
IDLE conflated two unrelated things: "controller disabled" and "needs
to cool, which this heat-only actuator fakes via zero output." Split
them apart:
- IDLE now means disabled only - the FSM forces it whenever
enabled=False, regardless of the temperature gap, and process_pid()
zeroes y for it same as before.
- New COOL state (mirroring HEAT) takes over the negative-diff case,
with its own pid_cool (separate "Cool" gains, alongside Hold/Heat)
instead of reusing pid_rate. Its output can legitimately be negative
- it's the actuator (tasks/heater.py's actor(), already max(0, ...))
that clamps it to 0 because *this* plant (Pot) can only heat. A
plant with real cooling capability could one day honor it directly.
- Thresholds renamed accordingly (HoldIdle->HoldCool, HeatIdle->
HeatCool, new CoolHold/CoolHeat); config.json/.sim/.templ and the
hand-rolled ctrl_params in scripts/demos/pid/ and demo_sud.py
updated with a "Cool" params section (mirrors "Heat" for now, since
there's no real cooling actuator to tune against yet).
Also fixes a regression in demo_sud.py/the other PID demos: none of
them ever called set_enabled(True), so since enabled defaults to
False they never drove the heater at all - only caught because this
change's demo_sud.py re-run got stuck in RAMPING forever.
New TempControllerBase.enabled (default False) forces y=0 in
process_pid() when off, same mechanism as the existing cooling
override - the controller tries not to drive the heater at all.
tasks/tempctrl.py: new 'Enable' recv command and 'Enabled' broadcast
on the TempCtrl channel. tasks/sud.py: SudTask now enables the
controller on any non-IDLE/DONE Sud state and disables it (forcing
power back to 0) on IDLE/DONE, so automatic mode owns it for the
duration of a run and hands back manual control once it stops/finishes.
gui: new "Enabled" checkbox on the Manual tab's Controller group,
synced from the server; the temp/heatrate setpoint controls are
disabled whenever the controller itself is disabled.
The codebase logs everything via plain print() scattered across
tasks/ws/components - rather than touch every call site, stdout/stderr
are now duplicated (Tee) into logs/brewpi.<timestamp>.log alongside
the existing tracer .mat files, line-buffered so a crash doesn't lose
the tail of the log.
Pure rename (brewpi.py, brewpi.sh, requirements.txt, __init__.py) plus
the README's path references - no code changes. "server/" describes
the directory's role rather than duplicating the project name already
in brewpi.py itself.
Persists whenever set via the entry, and re-applies automatically on
connect - so it survives a server restart (which would otherwise fall
back to config.json's ambient_temperature) without retyping it.
New btn_pot_reset next to the ambient entry, sends {"Pot": {"Reset":
true}}; tasks/pot.py's PotTask.recv() (previously a no-op) now resets
the plant to its current ambient temperature (Pot.initial()) on that
command.
The button only appears when the server reports its Pot is simulated
(new "PlantSim" field on the System channel, derived from
config.json's previously-unused Controller.plant_name) - resetting a
simulated plant's temperature is meaningless for a real one.
New lineEdit_ambient next to the host/connect row - type a value and
press Enter to send {"System": {"AmbientTemp": ...}}. The server now
has a recv handler for the System channel (previously send-only):
updates the real Pot's ambient temperature and, where the controller
has an internal model (Smith), its model Pots too (new
set_ambient_temperature() on TempControllerSmith), then echoes the new
value back so the status bar label and entry itself stay in sync.
New client/user_config.py - a small JSON-backed key/value store at
~/.config/brewpi/gui.json for GUI preferences that should survive
restarts (distinct from the server's config.json and from sude/*.json
schedules). First use: Save/Load Sud's file dialogs now open in
whichever directory was last used, instead of always starting at cwd.
The status bar now appends "(MM:SS remaining)" to a hold step's text,
kept live by recomputing it on every HoldRemaining message (not just
when the step itself changes) - tasks/sud.py already sent this, the
GUI just wasn't tracking/displaying it.
Nonideal ramp rates and user pauses make the schedule's nominal-rate
estimate drift from reality the longer a brew runs. While a run is in
progress, the forecast plot now redraws every tick: the already-
elapsed part is the actual measured trace (solid line), and only the
remaining steps are projected forward (dashed line) from the live
current temperature, step index, and hold_remaining - rather than
recomputing the whole curve once from t=0 at load time.
components/sud.py's Step/HoldRemaining messages already carried
everything needed; the GUI just wasn't tracking step index or
hold_remaining at all before this. Reverts to the static full-schedule
view once the run stops/finishes.
Progress line switches from red to green; a new horizontal red line
tracks the live measured temperature, drawn at a lower zorder than the
forecast/progress lines so it never visually covers them.
Caption becomes "BrewPi - <name> - <description>" once a schedule
with both is loaded, degrading gracefully to just the name, or plain
"BrewPi", when either is empty.
Window caption becomes "BrewPi - <name>" once a schedule is loaded
(back to plain "BrewPi" when empty/disconnected); the Automatic tab's
forecast title is now "<name> - N min total", replacing the generic
"Estimated course..." text while keeping the time estimate.
On (re)connect, the manual-mode controls (heater power/activate, temp
soll, heatrate soll, stirrer speed/activate) sync themselves once to
the server's current value via setValue()/setCheckState(). Without
blockSignals(), that programmatic update also fires valueChanged/
stateChanged, echoing the value straight back as a command.
For the heater power slider this was the cause of the "stale power >
0 after restart" bug: if the slider happened to sync to a nonzero
power (e.g. reconnecting mid-heat), the echoed {'Power': ...} set
tasks/heater.py's power_soll, which power_soll = max(power_soll,
power_actor) then latches as a floor that never drops back down on
its own - the heater would then stay at least that hot regardless of
what the actual controller/Sud wanted afterwards.
Several sude/*.json schedules can exist on disk; only one runs at a
time, chosen by a client Load. Sud no longer takes a path at all - it
starts empty (EMPTY_SUD) and the server config no longer names a
specific schedule file, removing the implicit link to sud_0010.json
in particular. Reading/writing sude/*.json is now entirely the
client's job (e.g. the GUI's Load/Save file dialogs); the server's Sud
only ever holds whatever was last Load()ed, in memory, until replaced
or the server restarts.
Updates demo_sud.py/demo_sud_save_load.py to construct an empty Sud()
and load() a schedule explicitly, matching the new constructor.
load() used to immediately persist whatever was loaded back to
self.path - the file the server was originally configured with. This
meant any Load (including "New Sud"'s empty schedule) silently
clobbered the currently-configured brew file, with no way to get it
back. load()/save() now operate purely on an in-memory copy; a client
that wants to persist a schedule does so explicitly itself (the GUI's
Save already prompts for a destination - this just removes the
hidden, automatic disk write load() was doing underneath it).
File > New Sud loads an empty schedule (zero steps) onto the server,
same mechanism as Load but with a built-in empty document. Tracks
whether the loaded schedule actually has steps (sud_empty) separately
from whether a Sud is configured at all (sud_loaded) - Start/Pause/Stop
now require both, and the Automatic tab's forecast shows "No sud
loaded" instead of an empty plot while sud_empty.
A step's target temperature applies to the whole step (it's what a
ramp ramps to and what a hold then holds at), not just its ramp phase
- promote it from a nested "ramp": {"temp": ...} to a step-level
"temperature" field, defaulted like descr/grain_mass/etc. Updates all
consumers (Sud._build_step, SudTask, the GUI's forecast estimator,
demo_sud.py) and migrates sude/sud_0010.json and the README.
Found while live-testing the previous fix: brewpi.py unconditionally
wired heater.power_set to tc.set_model_power, but "Normal" has no
internal model and thus no such method - same hasattr guard already
used elsewhere (tasks/sud.py's apply_plant_params) for this.
PidFactory.create() calls both controllers with the same (dt, params,
model_params, theta_amb=...) signature, but TempController ("Normal")
only accepted (dt, params) - any config using pid_type "Normal" would
crash with a TypeError on startup. It has no internal model, so just
accepts and ignores the extra arguments.
get_theta_ist() can return numpy.float64 (Pot's transport delay line
is a numpy array internally), so the cooldown comparison produced a
numpy.bool_ instead of a plain bool. json.dumps() can't serialize that,
which silently killed the websocket connection's send loop the moment
a ramp step started - found by actually driving the feature through a
live server+GUI run instead of only unit-style tests with a directly
constructed controller.
A ramp step whose target is below the current actual temperature can
only be reached by ambient cooling - the heater can't actively cool.
SudTask now detects this once per ramp phase and calls the controller's
new set_cooling() override (forces y=0, bypassing the FSM) instead of
waiting for its hysteresis-based IDLE transition. The GUI shows a
blinking blue "Cool down" label in the status bar while this is active.
Progression already only advances once theta_ist matches theta_soll
regardless of direction, so no change was needed there.
A step can now carry both 'ramp' and 'hold': it ramps to the target
temp, then holds there for the given duration, instead of needing two
separate schedule entries. Sud.temp_reached() switches such a step
from its ramp phase into its hold phase in place (re-firing the
step-changed callback so SudTask/demo_sud re-apply the hold's stirrer
settings); user_wait_for_continue still applies once the whole step -
both phases - is done.
Merges sud_0010.json's three ramp/hold pairs into combined steps.
Adds a 'System' channel for static, global info that doesn't belong
to any one task - sent once at startup directly from brewpi.py's
__main__ (a client connecting later still gets it via the
dispatcher's global_state replay on subscribe, same mechanism
SudTask's startup Name/Description already relies on).
Moves warp_factor out of SudTask's startup message and into this new
channel, since it's not actually Sud-specific - it's now available
(and the status bar shows it) even without a Sud configured. The
forecast plot's progress-line scaling switches to the same general
Window.warp_factor field.
Displayed via a permanent status bar widget (not cleared by the
existing transient step/schedule showMessage() calls).
Reverts 19dc7a9 ("Fix Sud hold duration units: seconds, not minutes")
per explicit instruction - duration is minutes again. Re-adds the
*60.0 conversion in Sud._advance() (hold_remaining is ticked in
simulated seconds), the matching conversion in the forecast plot's
_estimate_course(), and flips the README wording back.
SudTask now computes warp_factor = dt/interval (the same simulated-vs-
wall-clock split Pot/TempController/Stirrer already use internally)
and sends it once at startup as 'WarpFactor'. The GUI uses it to scale
real elapsed time into the forecast's simulated-schedule time axis,
so the progress line lines up with the actual sim speed instead of
assuming real time.
Also fixes SudTask.tick() to decrement hold_remaining by dt (simulated
seconds) instead of interval (wall-clock seconds) - it was the only
place in the sim using the wall-clock interval for something meant to
track simulated time, so hold steps were taking warp_factor times
longer in real time than their declared duration intends. Without
this fix the GUI's new warp-scaled progress line would race ahead of
the real server during holds.
Adds a thread-safe path from on_sud_changed() (websocket recv thread)
to a modal QMessageBox on the GUI thread via Qt's cross-thread
signal/slot queuing, fired when the State transition lands on
WAIT_USER - which the schedule is already blocked on server-side, so
the dialog just gives it a UI: closing it (OK) sends Confirm.
Deliberately scoped to WAIT_USER transitions, not every UserMessage -
most steps in sude/sud_0010.json inherit the schedule's default
placeholder text ("Put user message here") rather than overriding it,
so popping a dialog for every step would mostly show that placeholder.
SudTask.on_step_changed() set the controller's target temp/rate on
every step, but never touched the real Pot's or the Smith predictor's
internal model's thermal mass (M/C) - those stayed frozen at
brewpi.py's startup DEFAULT_PLANT_PARAMS for the whole brew, even
though grain_mass/water_mass change per step (malt going in, water
boiling off) and demo_sud.py already recomputes them every step via
derive_plant_params(). Production SudTask just never got the same
treatment.
Adds SudTask.apply_plant_params(), called alongside the existing
theta_soll/heatrate_soll push, mirroring the demo. Needs the real Pot
now, so SudTask takes a `pot` constructor arg; set_model_params() is
only called if the configured controller actually defines one (the
"Normal" pid_type doesn't).
Renames everywhere the concept appears: Sud.download()/upload() ->
save()/load(), the 'Download'/'Upload' websocket messages ->
'Save'/'Load', the brewpi_gui menu actions, handlers and
sud_download_path -> sud_save_path, and the demo script. Behavior is
unchanged - this is naming only, from the client's perspective (save
to / load from a local file) rather than the server's.
components/sud.py: adds SudState.PAUSED. pause() freezes RAMPING/
HOLDING (the hold countdown and ramp-reached checks both no-op while
PAUSED, since they're gated on the exact state they froze); start()
now doubles as resume when paused. stop() aborts back to IDLE from
any in-progress state, reusing the same reset logic as __init__/
upload() (factored into _reset_run_state()).
tasks/sud.py: maps 'Pause'/'Stop' messages to the new methods, and
turns the stirrer off on IDLE (not just DONE) so stop() actually
silences it instead of leaving the last step's stirring running.
client/brewpi_gui.py: wires the Pause/Stop toolbar actions, extends
update_sud_actions() so Start doubles as Resume and Stop stays usable
while paused, and freezes the forecast plot's progress line for the
duration of a pause (tracked via accumulated paused time) instead of
letting it drift ahead of actual progress.
on_sud_changed() now handles 'Step' messages, showing the step's
description (and clearing it once there's no active step), and tracks
IDLE/DONE <-> running transitions to time a red vertical progress line
on the Automatic tab's forecast - elapsed real time since the brew
started, plotted against the forecast's nominal-schedule time axis so
you can see how actual progress compares to the plan.
Sends {'Start': True} on the Sud channel, which SudTask.recv() already
maps to sud.start(). Only reachable while enabled (loaded and idle),
per update_sud_actions().
Start is enabled only once a schedule is loaded and idle; Pause/Stop
only while it's actually running (RAMPING/HOLDING/WAIT_USER). All
three stay disabled until something arrives on the Sud channel (which
only happens if a SudTask is actually configured server-side), and
get disabled again on disconnect since nothing reported over a dead
connection can still be trusted.
Also drops actionStop's leftover wiring to disconnect() - now that
Stop is meant to stop a running brew, it shouldn't tear down the
websocket connection instead.
ws_client.disconnect() blocks until the close handshake completes
(observed ~10s when the server doesn't ack promptly), and was being
called straight from the Connect/Disconnect button's click handler -
joining that wait on the GUI thread, so the click appeared to do
nothing until the timeout passed. Run it off a throwaway thread
instead; on_ws_connect_changed() already flips the button once the
disconnect actually finishes.
MessageDispatcherSync gains an optional on_connect_changed(bool)
callback, fired from the websocket client's background loop once the
connection is actually live/lost, so the button only flips state on a
real connection change rather than optimistically on click.
Start/Pause/Stop on the toolbar were never about the websocket
connection - Start now does nothing until that's wired to its own
purpose, and connecting is its own explicit action next to the host
text entry.
Adds a static forecast plot to the (previously empty) Automatic tab,
showing the planned temperature course derived from a schedule's
declared ramp rates and hold durations alone (no plant/controller
model - just how long the schedule itself says each step takes).
The client now requests the schedule via Download right after
connecting, routing the reply to the forecast instead of a save
dialog since sud_download_path is only set for the user-initiated
download. SudTask also echoes the just-applied document back after a
successful Upload, so editing the schedule refreshes the forecast
without an extra round trip.
Reformats it to json.dump()'s standard 2-space-free, ensure_ascii
output (from exercising the new upload feature against the real
schedule) and adds the matching downloaded copy. No schedule content
changed.
Drop dark_background in favor of a plain white figure, matching
demo_sud's blue/red ist/soll color convention. Add faint major/minor
gridlines, drop the top/right spines, and use label_outer() so only
the bottom subplot repeats x tick labels across the shared time axis.
Embeds a dark-themed matplotlib canvas (3 stacked strip charts: theta
ist/soll, heatrate ist/soll, heater power_set/power_eff) in a new Plot
tab, sampled once per second from the latest websocket-received values.
HeaterTask now also emits a PowerSet message (the pre-discretization
target power) alongside the existing Power (effective) message, since
the GUI had no way to see the set-point side of figure 1's bottom plot
otherwise.
Adds File menu actions to save the running Sud's schedule to a local
sud.json and to replace it from one, via the Sud websocket channel's
new Download/Upload messages. File dialogs are opened up front on the
GUI thread; the actual read/write happens off the websocket recv
callback, which runs on a worker thread and must not touch Qt itself.
main_window.py is regenerated from brewpi.ui via pyuic5 5.15.11, which
also picks up an unrelated pre-existing drift (the default URI had
fallen out of sync with the .ui source).
Sud.download() returns the on-disk document; Sud.upload() validates,
persists, and resets the schedule, refusing mid-brew or malformed input
so the current schedule is never left half-applied.
Both blocks were nearly identical (config.json.sim's differing M and
unused gain key being the only divergence), so there's no need for
per-config tuning here. DEFAULT_PLANT_PARAMS in brewpi.py now feeds
both the real Pot plant and the temperature controller's
Smith-predictor model. No demo scripts read these config keys, so
nothing else needed updating.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CgR9tPaSzFkAwRAUyeaaCD
jens
