The callbacks API is experimental. The interface may change in future releases.
logs dictionary containing current simulation state.
pybamm.callbacks.Callback
Base class. Override any of the on_<event> methods in your subclass. Each method receives logs and should return None.
class Callback:
def on_experiment_start(self, logs): ...
def on_cycle_start(self, logs): ...
def on_step_start(self, logs): ...
def on_step_end(self, logs): ...
def on_cycle_end(self, logs): ...
def on_experiment_end(self, logs): ...
def on_experiment_error(self, logs): ...
def on_experiment_infeasible_time(self, logs): ...
def on_experiment_infeasible_event(self, logs): ...
Hook methods
| Method | When called | Typical logs keys |
|---|
on_experiment_start | Before the first cycle | — |
on_cycle_start | Before each cycle | "cycle number", "elapsed time" |
on_step_start | Before each step within a cycle | "cycle number", "step number", "step operating conditions" |
on_step_end | After each step | "stopping conditions", "experiment time" |
on_cycle_end | After each cycle | "stopping conditions", "summary variables", "Minimum voltage [V]", "start capacity" |
on_experiment_end | After the final cycle | "elapsed time" |
on_experiment_error | When a SolverError occurs | "error" |
on_experiment_infeasible_time | Default duration reached without termination event | "step duration", "cycle number", "step number", "step operating conditions" |
on_experiment_infeasible_event | Termination event triggered before experiment completes | "termination", "cycle number", "step number", "step operating conditions" |
pybamm.callbacks.CallbackList
A container that holds multiple Callback objects and calls each one in sequence. Construct it with a list of callbacks:
callbacks = pybamm.callbacks.CallbackList([
pybamm.callbacks.LoggingCallback(),
my_custom_callback,
])
CallbackList mirror those of Callback and automatically dispatch to every registered callback.
pybamm.callbacks.LoggingCallback
The default callback. Writes progress information to PyBaMM’s logger (or an optional log file).
pybamm.callbacks.LoggingCallback(logfile=None)
Path to a log file. If None, output is sent to pybamm.logger (stdout by default).
LoggingCallback is automatically added to every simulation that does not already have one.
Custom Callback Example
import pybamm
class StopOnCapacityFade(pybamm.callbacks.Callback):
"""
Raise an error if capacity drops below a threshold.
"""
def __init__(self, capacity_threshold):
self.capacity_threshold = capacity_threshold
def on_cycle_end(self, logs):
cap = logs["summary variables"]["Capacity [A.h]"]
if cap < self.capacity_threshold:
raise pybamm.SolverError(
f"Capacity {cap:.3f} Ah fell below threshold "
f"{self.capacity_threshold:.3f} Ah"
)
experiment = pybamm.Experiment(
["Discharge at 1C until 2.5 V", "Charge at 1C until 4.2 V"] * 100
)
sim = pybamm.Simulation(
pybamm.lithium_ion.SPM(),
experiment=experiment,
parameter_values=pybamm.ParameterValues("Chen2020"),
)
sim.solve(
callbacks=StopOnCapacityFade(capacity_threshold=4.5)
)
Using Multiple Callbacks
import pybamm
class SaveProgress(pybamm.callbacks.Callback):
def __init__(self):
self.cycles = []
def on_cycle_end(self, logs):
cap = logs["summary variables"]["Capacity [A.h]"]
self.cycles.append(cap)
save_cb = SaveProgress()
log_cb = pybamm.callbacks.LoggingCallback(logfile="run.log")
sim.solve(callbacks=[save_cb, log_cb])
print(save_cb.cycles) # list of capacity at each cycle end
