async def coro_calculate(self, calc_all: bool):
"""
Coroutine to calculate all results.
"""
self.is_calculating_all = calc_all
self.mp_handler = MPHandler()
self.mp_handler.stop()
params = self.get_params(all_signals=calc_all)
if params.transform == _wft:
if self.view.get_fmin() is None:
raise Exception("Minimum frequency must be defined for WFT.")
self.is_plotted = False
self.view.main_plot().clear()
self.view.main_plot().set_in_progress(True)
self.invalidate_data()
log: bool = (params.transform == _wt)
self.view.main_plot().set_log_scale(logarithmic=log)
self.view.amplitude_plot().set_log_scale(logarithmic=log)
self.view.on_calculate_started()
all_data = await self.mp_handler.coro_transform(
params, self.on_progress_updated)
for d in all_data:
self.on_transform_completed(*d)
async def coro_preprocess_selected_signal(self) -> List[ndarray]:
sig = self.get_selected_signal()
if not self.preproc_mp_handler:
self.preproc_mp_handler = MPHandler()
return await self.preproc_mp_handler.coro_preprocess(sig, None, None)
async def coro_calculate(self) -> None:
self.view.enable_save_data(False)
self.view.groupbox_stats.setEnabled(
self.should_stats_be_enabled(calculating=True))
if self.mp_handler:
self.mp_handler.stop()
self.is_plotted = False
self.invalidate_data()
self.view.main_plot().clear()
self.view.main_plot().set_in_progress(True)
self.view.amplitude_plot().clear()
self.view.amplitude_plot().set_in_progress(True)
params = self.get_params()
self.params = params
self.mp_handler = MPHandler()
self.view.on_calculate_started()
if "PYMODALIB_CACHE" not in os.environ:
cache = Settings().get_pymodalib_cache()
if cache and cache is not "None":
os.environ["PYMODALIB_CACHE"] = cache
sig1a, sig1b, sig2a, sig2b = self.signals.get_all()
if sig2a is None:
self.results = (await self.mp_handler.coro_group_coherence(
sig1a,
sig1b,
fs=self.signals.frequency,
on_progress=self.on_progress_updated,
**params,
))[0]
else:
self.results = (await self.mp_handler.coro_dual_group_coherence(
sig1a,
sig1b,
sig2a,
sig2b,
fs=self.signals.frequency,
on_progress=self.on_progress_updated,
**params,
))[0]
self.view.groupbox_stats.setEnabled(
self.should_stats_be_enabled(calculating=False))
self.enable_save_data(True)
self.update_plots()
self.view.on_calculate_stopped()
self.on_all_tasks_completed()
async def coro_calculate(self):
"""Coroutine which calculates the bispectra."""
if self.mp_handler:
self.mp_handler.stop()
self.is_plotted = False
self.invalidate_data()
# We need to use the same params for biphase later; save now because the UI could change.
self.params = self.get_params()
self.mp_handler = MPHandler()
data = await self.mp_handler.coro_bispectrum_analysis(
self.signals, self.params, self.on_progress_updated
)
for d in data:
self.on_bispectrum_completed(*d)
self.view.on_calculate_stopped()
self.update_plots()
async def coro_preprocess_selected_signal(self) -> List[ndarray]:
"""
Coroutine to preprocess the currently selected signal.
:return: the preprocessed signal as a 1D array
"""
sig = self.get_selected_signal()
fmin = self.view.get_fmin()
fmax = self.view.get_fmax()
if not self.preproc_mp_handler:
self.preproc_mp_handler = MPHandler()
return await self.preproc_mp_handler.coro_preprocess(sig, fmin, fmax)
async def coro_plot_preprocessed_signal(self) -> None:
"""
Coroutine to preprocess the signal and plot the result.
"""
sig = self.get_selected_signal()
fmin = self.view.get_fmin()
fmax = self.view.get_fmax()
if not self.preproc_mp_handler:
self.preproc_mp_handler = MPHandler()
result = await self.preproc_mp_handler.coro_preprocess(sig, fmin, fmax)
if result and result[0] is not None:
self.view.plot_preprocessed_signal(sig.times, sig.signal, result[0])
async def coro_preprocess_all_signals(self) -> List[ndarray]:
"""
Coroutine to preprocess all signals.
:return: a list containing the preprocessed signals as a 1D array each
"""
signals = self.signals
try:
fmin = self.view.get_fmin()
fmax = self.view.get_fmax()
except AttributeError:
fmin, fmax = None, None
if not self.preproc_mp_handler:
self.preproc_mp_handler = MPHandler()
return await self.preproc_mp_handler.coro_preprocess(signals, fmin, fmax)
async def coro_calculate(self, calculate_all: bool) -> None:
self.view.enable_save_data(False)
self.is_calculating_all = calculate_all
if self.mp_handler:
self.mp_handler.stop()
self.is_plotted = False
self.invalidate_data()
self.view.main_plot().clear()
self.view.main_plot().set_in_progress(True)
self.view.amplitude_plot().clear()
self.view.amplitude_plot().set_in_progress(True)
params = self.get_params(all_signals=calculate_all)
self.params = params
self.surrogate_count = self.view.get_surr_count()
self.surrogates_enabled = self.view.get_surr_enabled()
self.mp_handler = MPHandler()
self.view.main_plot().set_log_scale(logarithmic=True)
self.view.amplitude_plot().set_log_scale(logarithmic=True)
self.view.on_calculate_started()
data = await self.mp_handler.coro_transform(
params=params, on_progress=self.on_progress_updated)
for d in data:
self.on_transform_completed(*d)
all_data = await self.coro_phase_coherence(self.signals_calc, params,
self.on_progress_updated)
for d in all_data:
self.on_phase_coherence_completed(*d)
self.plot_phase_coherence()
self.view.on_calculate_stopped()
self.on_all_tasks_completed()
print("Finished calculating phase coherence.")
async def coro_calculate(self):
if self.mp_handler:
self.mp_handler.stop()
self.mp_handler = MPHandler()
data = await self.mp_handler.coro_bayesian(self.signals,
self.get_paramsets(),
self.on_progress_updated)
for d in data:
self.on_bayesian_inference_completed(*d)
if data:
self.plot_bayesian()
else:
print("No data returned; are any parameter sets added?")
print("Dynamical Bayesian inference completed.")
self.view.on_calculate_stopped()
async def coro_calculate(self, calculate_all: bool) -> None:
"""
Coroutine to perform the calculation.
:param calculate_all: whether to calculate for all signals, or just the current signal
"""
self.is_calculating_all = calculate_all
self.view.enable_save_data(False)
for s in self.signals:
s.data = None
if self.mp_handler:
self.mp_handler.stop()
self.mp_handler = MPHandler()
params = self.get_params(all_signals=calculate_all)
self.params = params
self.params.preprocess = self.view.get_preprocess()
self.is_plotted = False
self.view.main_plot().clear()
self.invalidate_data()
self.view.main_plot().set_log_scale(logarithmic=True)
self.view.on_calculate_started()
all_data = await self.mp_handler.coro_harmonics(
self.signals_calc, params, self.params.preprocess,
self.on_progress_updated)
if not isinstance(all_data, List):
all_data = [all_data]
for signal, data in zip(self.signals_calc, all_data):
signal.data = data
self.view.enable_save_data(bool(all_data))
self.view.on_calculate_stopped()
self.update_plots()
class BAPresenter(BaseTFPresenter):
def __init__(self, view):
super().__init__(view)
self.params: BAParams = None
from gui.windows.bispectrum.BAWindow import BAWindow
self.view: BAWindow = view
def init(self):
super().init()
self.view.switch_to_dual_plot()
def calculate(self, calculate_all: bool):
"""Starts the coroutine which calculates the data."""
asyncio.ensure_future(self.coro_calculate())
self.view.on_calculate_started()
async def coro_calculate(self):
"""Coroutine which calculates the bispectra."""
if self.mp_handler:
self.mp_handler.stop()
self.is_plotted = False
self.invalidate_data()
# We need to use the same params for biphase later; save now because the UI could change.
self.params = self.get_params()
self.mp_handler = MPHandler()
data = await self.mp_handler.coro_bispectrum_analysis(
self.signals, self.params, self.on_progress_updated
)
for d in data:
self.on_bispectrum_completed(*d)
self.view.on_calculate_stopped()
self.update_plots()
def add_point(self, x: float, y: float):
asyncio.ensure_future(self.coro_biphase(x, y))
self.view.on_calculate_started()
async def coro_biphase(self, x: float, y: float):
self.mp_handler.stop()
fs = self.params.fs
f0 = self.params.f0
fr = self.view.get_selected_freq_pair()
x, y = fr
if x is not None and y is not None:
data = await self.mp_handler.coro_biphase(
self.signals, fs, f0, fr, self.on_progress_updated
)
for d in data:
self.on_biphase_completed(*d)
self.view.on_calculate_stopped()
self.update_side_plots(self.get_selected_signal_pair()[0].output_data)
def update_plots(self):
"""
Updates all plots according to the currently selected plot type and current data.
"""
data = self.get_selected_signal_pair()[0].output_data
self.update_main_plot(data)
self.update_side_plots(data)
def update_main_plot(self, data):
"""
Updates the main plot, plotting the wavelet transform or bispectrum depending
on the selected plot type.
:param data: the data object
"""
x, y, c, log = self.get_main_plot_data(self.view.get_plot_type(), data)
freq_x, freq_y = self.view.get_selected_freq_pair()
if self.view.is_wt_selected():
self.view.switch_to_dual_plot()
else:
self.view.switch_to_triple_plot()
if freq_x is not None and freq_y is not None:
self.view.plot_main.draw_crosshair(freq_x, freq_y)
if c is not None:
self.view.plot_main.set_log_scale(log, axis="x")
self.view.plot_main.set_log_scale(True, axis="y")
self.view.plot_main.update_xlabel("Frequency (Hz)")
self.view.plot_main.update_ylabel("Frequency (Hz)")
self.view.plot_main.plot(x=x, y=y, c=c)
def update_side_plots(self, data: BAOutputData):
"""
Updates the side plot(s). For wavelet transforms this will be
the average amplitude/power, but for bispectra this will be the
biphase and biamplitude.
:param data: the data object
"""
if self.view.is_wt_selected(): # Plot average amplitude or power.
amp_not_power = self.view.is_amplitude_selected()
x, y = self.get_side_plot_data_wt(
self.view.get_plot_type(), data, amp_not_power
)
if x is not None and y is not None and len(x) > 0:
plot = self.view.plot_right_top
plot.set_xlabel("Average amplitude")
plot.set_ylabel("Frequency (Hz)")
plot.set_log_scale(True, "y")
plot.plot(x, y)
else: # Plot biphase and biamplitude.
biamp, biphase = self.get_side_plot_data_bispec(
self.view.get_plot_type(), self.view.get_selected_freq_pair(), data
)
times = self.get_selected_signal().times
if biamp is not None:
self.view.plot_right_middle.update_ylabel("Biamplitude")
self.view.plot_right_middle.update_xlabel("Time (s)")
self.view.plot_right_middle.plot(times, biamp)
if biphase is not None:
self.view.plot_right_bottom.update_ylabel("Biphase")
self.view.plot_right_bottom.update_xlabel("Time (s)")
self.view.plot_right_bottom.plot(times, biphase)
@staticmethod
def get_side_plot_data_wt(
plot_type: str, data: BAOutputData, amp_not_power: bool
) -> Tuple[ndarray, ndarray]:
"""
Gets the data required to plot the average amplitude/power on the side plot. Used
when a wavelet transform is selected.
:param plot_type: the plot type shown in the QComboBox, e.g. "Wavelet transform 1"
:param data: the data object
:param amp_not_power: whether the data should be amplitude, or power
:return: the x-values, the y-values
"""
_dict = {
"Wavelet transform 1": (
data.avg_amp_wt1 if amp_not_power else data.avg_pow_wt1,
data.freq,
),
"Wavelet transform 2": (
data.avg_amp_wt2 if amp_not_power else data.avg_pow_wt2,
data.freq,
),
}
return _dict.get(plot_type)
@staticmethod
def get_side_plot_data_bispec(
plot_type: str, freq: Tuple[float, float], data: BAOutputData
):
"""
Gets the data required to plot the biphase and biamplitude on the side plots.
Used when bispectrum is selected.
:param freq: the selected frequencies (x and y)
:param plot_type the plot type shown in the QComboBox, e.g. "b111"
:param data the data object
:return: biamplitude and biphase
"""
key = ", ".join([str(f) for f in freq])
try:
biamp = data.biamp.get(key)
biphase = data.biphase.get(key)
except AttributeError:
biamp = None
biphase = None
if "None" in key or biamp is None:
return None, None
_dict = {
"b111": (biamp[0], biphase[0]),
"b222": (biamp[1], biphase[1]),
"b122": (biamp[2], biphase[2]),
"b211": (biamp[3], biphase[3]),
}
biamp, biphase = _dict.get(plot_type)
return biamp, biphase
@staticmethod
def get_main_plot_data(
plot_type: str, data: BAOutputData
) -> Tuple[ndarray, ndarray, ndarray, bool]:
"""
Gets the relevant arrays to plot in the main plot (WT or bispectrum).
:param plot_type: the type of plot, as shown in the QComboBox; e.g. "Wavelet transform 1"
:param data: the data object with all data
:return: the x-values, the y-values, the c-values, and whether to use a log scale
"""
if not isinstance(data, BAOutputData):
return [None for _ in range(4)]
_dict = {
"Wavelet transform 1": (data.times, data.freq, data.amp_wt1, False),
"Wavelet transform 2": (data.times, data.freq, data.amp_wt2, False),
"b111": (data.freq, data.freq, data.bispxxx, True),
"b222": (data.freq, data.freq, data.bispppp, True),
"b122": (data.freq, data.freq, data.bispxpp, True),
"b211": (data.freq, data.freq, data.bisppxx, True),
}
data = _dict.get(plot_type)
if data is None: # All plots.
pass # TODO
return data
def on_bispectrum_completed(
self,
name: str,
freq: ndarray,
amp_wt1: ndarray,
pow_wt1: ndarray,
avg_amp_wt1: ndarray,
avg_pow_wt1: ndarray,
amp_wt2: ndarray,
pow_wt2: ndarray,
avg_amp_wt2: ndarray,
avg_pow_wt2: ndarray,
bispxxx: ndarray,
bispppp: ndarray,
bispxpp: ndarray,
bisppxx: ndarray,
surrxxx: ndarray,
surrppp: ndarray,
surrxpp: ndarray,
surrpxx: ndarray,
opt: dict,
):
# Attach the data to the first signal in the current pair.
sig = self.signals.get(name)
sig.output_data = BAOutputData(
amp_wt1,
pow_wt1,
avg_amp_wt1,
avg_pow_wt1,
amp_wt2,
pow_wt2,
avg_amp_wt2,
avg_pow_wt2,
sig.times,
freq,
bispxxx,
bispppp,
bispxpp,
bisppxx,
surrxxx,
surrppp,
surrxpp,
surrpxx,
opt,
{},
{},
)
def on_biphase_completed(
self,
name: str,
freq_x: float,
freq_y: float,
biamp1: ndarray,
biphase1: ndarray,
biamp2: ndarray,
biphase2: ndarray,
biamp3: ndarray,
biphase3: ndarray,
biamp4: ndarray,
biphase4: ndarray,
):
sig = self.signals.get(name)
key = f"{freq_x}, {freq_y}"
data = sig.output_data
data.biamp[key] = [[] for _ in range(4)]
data.biphase[key] = [[] for _ in range(4)]
data.biamp[key][0] = biamp1
data.biphase[key][0] = biphase1
data.biamp[key][1] = biamp2
data.biphase[key][1] = biphase2
data.biamp[key][2] = biamp3
data.biphase[key][2] = biphase3
data.biamp[key][3] = biamp4
data.biphase[key][3] = biphase4
def load_data(self):
"""
Loads the data from a file, showing a dialog to set the frequency of
the signal.
"""
self.signals = SignalPairs.from_file(self.open_file)
if not self.signals.has_frequency():
freq = FrequencyDialog().run_and_get()
if freq:
self.set_frequency(freq)
self.on_data_loaded()
else:
raise Exception("Frequency was None. Perhaps it was mistyped?")
def on_data_loaded(self):
self.view.update_signal_listview(self.signals.get_pair_names())
self.plot_signal_pair()
def plot_signal_pair(self):
self.view.plot_signal_pair(self.get_selected_signal_pair())
def on_signal_selected(self, item: Union[QListWidgetItem, str]):
if isinstance(item, QListWidgetItem):
name = item.text()
else:
name = item
self.signals.reset()
if name != self.selected_signal_name:
print(f"Selected '{name}'")
self.selected_signal_name = name
self.plot_signal_pair()
self.view.on_xlim_edited()
self.plot_preprocessed_signal()
def get_selected_signal_pair(self) -> Tuple[TimeSeries, TimeSeries]:
"""
Gets the currently selected signal pair as a tuple containing 2 signals.
"""
return self.signals.get_pair_by_name(self.selected_signal_name)
def get_params(self) -> BAParams:
"""Gets data from the GUI to create the params used by the bispectrum calculation."""
return BAParams(
signals=self.signals,
preprocess=self.view.get_preprocess(),
fmin=self.view.get_fmin(),
fmax=self.view.get_fmax(),
f0=self.view.get_f0(),
nv=self.view.get_nv(),
surr_count=self.view.get_surr_count(),
opt={},
)
class BAPresenter(BaseTFPresenter):
def __init__(self, view):
super().__init__(view)
self.params: BAParams = None
from gui.windows.bispectrum.BAWindow import BAWindow
self.view: BAWindow = view
def init(self):
super().init()
self.view.switch_to_dual_plot()
def calculate(self, calculate_all: bool):
"""Starts the coroutine which calculates the data."""
asyncio.ensure_future(self.coro_calculate())
self.view.on_calculate_started()
async def coro_calculate(self):
"""Coroutine which calculates the bispectra."""
self.enable_save_data(False)
if self.mp_handler:
self.mp_handler.stop()
self.is_plotted = False
self.invalidate_data()
# We need to use the same params for biphase later; save now because the UI could change.
self.params = self.get_params()
self.mp_handler = MPHandler()
data = await self.mp_handler.coro_bispectrum_analysis(
self.signals, self.params, self.on_progress_updated)
for d in data:
self.on_bispectrum_completed(*d)
self.enable_save_data(True)
self.view.on_calculate_stopped()
self.update_plots()
def add_point(self, x: float, y: float):
asyncio.ensure_future(self.coro_biphase(x, y))
self.view.on_calculate_started()
async def coro_biphase(self, x: float, y: float):
self.enable_save_data(False)
self.mp_handler.stop()
fs = self.params.fs
f0 = self.params.f0
fr = self.view.get_selected_freq_pair()
x, y = fr
if x is not None and y is not None:
data = await self.mp_handler.coro_biphase(self.signals, fs, f0, fr,
self.on_progress_updated)
for d in data:
self.on_biphase_completed(*d)
self.enable_save_data(True)
self.view.on_calculate_stopped()
self.update_side_plots(
self.get_selected_signal_pair()[0].output_data)
def update_plots(self):
"""
Updates all plots according to the currently selected plot type and current data.
"""
data = self.get_selected_signal_pair()[0].output_data
try:
self.update_main_plot(data)
self.update_side_plots(data)
except AttributeError:
pass
except ValueError as e:
msg = "zero-size array to reduction operation minimum which has no identity"
if msg in str(e):
print(
f"'{self.view.combo_plot_type.currentText()}' is not available to plot."
)
else:
raise e
def set_plot_type(self, amplitude_selected=True) -> None:
"""
Set the type of plot to display (power or amplitude). This affects
the main plot and the amplitude plot.
:param amplitude_selected: whether to set the plot type as amplitude (not power)
"""
self.plot_ampl = amplitude_selected
self.update_plots()
def update_main_plot(self, data):
"""
Updates the main plot, plotting the wavelet transform or bispectrum depending
on the selected plot type.
:param data: the data object
"""
x, y, c, log = self.get_main_plot_data(self.view.get_plot_type(), data)
freq_x, freq_y = self.view.get_selected_freq_pair()
if self.view.is_wt_selected():
self.view.switch_to_dual_plot()
else:
self.view.switch_to_triple_plot()
if freq_x is not None and freq_y is not None:
self.view.plot_main.draw_crosshair(freq_x, freq_y)
if c is not None:
self.view.plot_main.set_log_scale(log, axis="x")
self.view.plot_main.set_log_scale(True, axis="y")
self.view.plot_main.update_xlabel("Frequency (Hz)")
self.view.plot_main.update_ylabel("Frequency (Hz)")
self.view.plot_main.plot(x=x, y=y, c=c)
def update_side_plots(self, data: BAOutputData):
"""
Updates the side plot(s). For wavelet transforms this will be
the average amplitude/power, but for bispectra this will be the
biphase and biamplitude.
:param data: the data object
"""
if self.view.is_wt_selected(): # Plot average amplitude or power.
x, y = self.get_side_plot_data_wt(self.view.get_plot_type(), data,
self.plot_ampl)
if x is not None and y is not None and len(x) > 0:
plot = self.view.plot_right_top
plot.set_xlabel("Average amplitude")
plot.set_ylabel("Frequency (Hz)")
plot.set_log_scale(True, "y")
plot.plot(x, y)
else: # Plot biphase and biamplitude.
biamp, biphase = self.get_side_plot_data_bispec(
self.view.get_plot_type(), self.view.get_selected_freq_pair(),
data)
times = self.get_selected_signal().times
if biamp is not None:
self.view.plot_right_middle.update_ylabel("Biamplitude")
self.view.plot_right_middle.update_xlabel("Time (s)")
self.view.plot_right_middle.plot(times, biamp)
if biphase is not None:
self.view.plot_right_bottom.update_ylabel("Biphase")
self.view.plot_right_bottom.update_xlabel("Time (s)")
self.view.plot_right_bottom.plot(times, biphase)
@staticmethod
def get_side_plot_data_wt(plot_type: str, data: BAOutputData,
amp_not_power: bool) -> Tuple[ndarray, ndarray]:
"""
Gets the data required to plot the average amplitude/power on the side plot. Used
when a wavelet transform is selected.
:param plot_type: the plot type shown in the QComboBox, e.g. "Wavelet transform 1"
:param data: the data object
:param amp_not_power: whether the data should be amplitude, or power
:return: the x-values, the y-values
"""
_dict = {
"Wavelet transform 1": (
data.avg_amp_wt1 if amp_not_power else data.avg_pow_wt1,
data.freq,
),
"Wavelet transform 2": (
data.avg_amp_wt2 if amp_not_power else data.avg_pow_wt2,
data.freq,
),
}
return _dict.get(plot_type)
@staticmethod
def get_side_plot_data_bispec(plot_type: str, freq: Tuple[float, float],
data: BAOutputData):
"""
Gets the data required to plot the biphase and biamplitude on the side plots.
Used when bispectrum is selected.
:param freq: the selected frequencies (x and y)
:param plot_type the plot type shown in the QComboBox, e.g. "b111"
:param data the data object
:return: biamplitude and biphase
"""
key = ", ".join([str(f) for f in freq])
try:
biamp = data.biamp.get(key)
biphase = data.biphase.get(key)
except AttributeError:
biamp = None
biphase = None
if "None" in key or biamp is None:
return None, None
_dict = {
"b111": (biamp[0], biphase[0]),
"b222": (biamp[1], biphase[1]),
"b122": (biamp[2], biphase[2]),
"b211": (biamp[3], biphase[3]),
}
biamp, biphase = _dict.get(plot_type)
return biamp, biphase
def get_main_plot_data(
self, plot_type: str,
data: BAOutputData) -> Tuple[ndarray, ndarray, ndarray, bool]:
"""
Gets the relevant arrays to plot in the main plot (WT or bispectrum).
:param plot_type: the type of plot, as shown in the QComboBox; e.g. "Wavelet transform 1"
:param data: the data object with all data
:return: the x-values, the y-values, the c-values, and whether to use a log scale
"""
if not isinstance(data, BAOutputData):
return [None for _ in range(4)]
plot_surr = self.view.get_plot_surrogates_selected()
if self.plot_ampl:
wt1 = data.amp_wt1
wt2 = data.amp_wt2
else:
wt1 = data.pow_wt1
wt2 = data.pow_wt2
_dict = {
"Wavelet transform 1": (data.times, data.freq, wt1, False),
"Wavelet transform 2": (data.times, data.freq, wt2, False),
"b111": (data.freq, data.freq, data.bispxxx, True),
"b222": (data.freq, data.freq, data.bispppp, True),
"b122": (data.freq, data.freq, data.bispxpp, True),
"b211": (data.freq, data.freq, data.bisppxx, True),
}
tup = _dict.get(plot_type)
if tup is None: # All plots.
pass # TODO
if plot_surr and self.params.surr_count > 0 and "transform" not in plot_type:
tup = self.apply_surrogates(plot_type, data, tup)
return tup
def apply_surrogates(
self,
plot_type: str,
data: BAOutputData,
tup: Tuple[ndarray, ndarray, ndarray, bool],
) -> Tuple[ndarray, ndarray, ndarray, bool]:
fx, fy, bisp, b = tup
K = np.int(np.floor((self.params.surr_count + 1) * self.params.alpha))
surr = {
"b111": data.surrxxx,
"b222": data.surrppp,
"b122": data.surrxpp,
"b211": data.surrpxx,
}.get(plot_type)[:, :, K]
bisp = bisp.copy()
bisp -= surr
bisp[bisp < 0] = np.NAN
return fx, fy, bisp, b
def on_bispectrum_completed(
self,
name: str,
freq: ndarray,
amp_wt1: ndarray,
pow_wt1: ndarray,
avg_amp_wt1: ndarray,
avg_pow_wt1: ndarray,
amp_wt2: ndarray,
pow_wt2: ndarray,
avg_amp_wt2: ndarray,
avg_pow_wt2: ndarray,
bispxxx: ndarray,
bispppp: ndarray,
bispxpp: ndarray,
bisppxx: ndarray,
surrxxx: ndarray,
surrppp: ndarray,
surrxpp: ndarray,
surrpxx: ndarray,
opt: Dict,
) -> None:
self.opt: Dict = opt
# Attach the data to the first signal in the current pair.
sig = self.signals.get(name)
sig.output_data = BAOutputData(
amp_wt1,
pow_wt1,
avg_amp_wt1,
avg_pow_wt1,
amp_wt2,
pow_wt2,
avg_amp_wt2,
avg_pow_wt2,
sig.times,
freq,
bispxxx,
bispppp,
bispxpp,
bisppxx,
surrxxx,
surrppp,
surrxpp,
surrpxx,
opt,
{},
{},
)
def on_biphase_completed(
self,
name: str,
freq_x: float,
freq_y: float,
biamp1: ndarray,
biphase1: ndarray,
biamp2: ndarray,
biphase2: ndarray,
biamp3: ndarray,
biphase3: ndarray,
biamp4: ndarray,
biphase4: ndarray,
) -> None:
sig = self.signals.get(name)
key = f"{freq_x}, {freq_y}"
data = sig.output_data
data.biamp[key] = [[] for _ in range(4)]
data.biphase[key] = [[] for _ in range(4)]
data.biamp[key][0] = biamp1
data.biphase[key][0] = biphase1
data.biamp[key][1] = biamp2
data.biphase[key][1] = biphase2
data.biamp[key][2] = biamp3
data.biphase[key][2] = biphase3
data.biamp[key][3] = biamp4
data.biphase[key][3] = biphase4
@override
async def coro_get_data_to_save(self) -> Dict:
if not self.opt or not self.params:
return
output_data: List[BAOutputData] = [
s.output_data for s, _ in self.signals.get_pairs()
]
cols = len(output_data)
first = output_data[0]
amp = np.empty((*first.amp_wt1.shape, cols * 2))
avg_amp = np.empty((first.avg_amp_wt1.shape[0], cols * 2))
b111 = np.empty((*first.bispxxx.shape, cols))
b222 = np.empty(b111.shape)
b122 = np.empty(b111.shape)
b211 = np.empty(b111.shape)
freq = first.freq
time = first.times
preproc = [] # TODO: Save this and other params
for index in range(0, len(output_data) * 2, 2):
d = output_data[index]
amp[:, :, index] = d.amp_wt1[:]
avg_amp[:, index] = d.avg_amp_wt1[:]
amp[:, :, index + 1] = d.amp_wt2[:]
avg_amp[:, index + 1] = d.avg_amp_wt2[:]
b111[:, :, index] = d.bispxxx
b222[:, :, index] = d.bispppp
b122[:, :, index] = d.bispxpp
b211[:, :, index] = d.bisppxx
ba_data = {
"amplitude": amp,
"avg_amplitude": avg_amp,
"frequency": freq,
"time": time,
"preprocessed_signals": preproc,
"b111": b111,
"b222": b222,
"b122": b122,
"b211": b211,
"fr": self.view.get_f0() or 1,
"fmin": self.opt["fmin"],
"fmax": self.opt["fmax"],
"preprocessing": "on" if self.params.preprocess else "off",
"sampling_frequency": self.params.fs,
}
return {"BAData": sanitise(ba_data)}
def load_data(self):
"""
Loads the data from a file, showing a dialog to set the frequency of
the signal.
"""
self.signals = SignalPairs.from_file(self.open_file)
if not self.signals.has_frequency():
freq = FrequencyDialog().run_and_get()
if freq:
self.set_frequency(freq)
self.on_data_loaded()
else:
raise Exception("Frequency was None. Perhaps it was mistyped?")
def on_data_loaded(self):
self.view.update_signal_listview(self.signals.get_pair_names())
self.plot_signal_pair()
def plot_signal_pair(self):
self.view.plot_signal_pair(self.get_selected_signal_pair())
def on_signal_selected(self, item: Union[QListWidgetItem, str]):
if isinstance(item, QListWidgetItem):
name = item.text()
else:
name = item
self.signals.reset()
if name != self.selected_signal_name:
print(f"Selected '{name}'")
self.selected_signal_name = name
self.plot_signal_pair()
self.view.on_xlim_edited()
self.plot_preprocessed_signal()
def get_selected_signal_pair(self) -> Tuple[TimeSeries, TimeSeries]:
"""
Gets the currently selected signal pair as a tuple containing 2 signals.
"""
return self.signals.get_pair_by_name(self.selected_signal_name)
def get_params(self) -> BAParams:
"""Gets data from the GUI to create the params used by the bispectrum calculation."""
return BAParams(
signals=self.signals,
preprocess=self.view.get_preprocess(),
fmin=self.view.get_fmin(),
fmax=self.view.get_fmax(),
f0=self.view.get_f0(),
nv=self.view.get_nv(),
surr_count=self.view.get_surr_count(),
alpha=self.view.get_alpha(),
opt={},
)
class TFPresenter(BaseTFPresenter):
"""
The presenter in control of the time-frequency window.
"""
def __init__(self, view):
super(TFPresenter, self).__init__(view)
from gui.windows.timefrequency import TFWindow
self.view: TFWindow = view
self.is_calculating_all = True
def calculate(self, calculate_all: bool):
"""
Calculates the desired transform(s), and plots the result.
"""
# If WFT parameters are incorrect, show error.
if not self.view.get_fmin(
) and not self.view.is_wavelet_transform_selected():
return self.view.show_wft_error()
asyncio.ensure_future(self.coro_calculate(calculate_all))
print("Started calculation...")
async def coro_calculate(self, calc_all: bool):
"""
Coroutine to calculate all results.
"""
self.is_calculating_all = calc_all
self.mp_handler = MPHandler()
self.mp_handler.stop()
params = self.get_params(all_signals=calc_all)
if params.transform == _wft:
if self.view.get_fmin() is None:
raise Exception("Minimum frequency must be defined for WFT.")
self.is_plotted = False
self.view.main_plot().clear()
self.view.main_plot().set_in_progress(True)
self.invalidate_data()
log: bool = (params.transform == _wt)
self.view.main_plot().set_log_scale(logarithmic=log)
self.view.amplitude_plot().set_log_scale(logarithmic=log)
self.view.on_calculate_started()
all_data = await self.mp_handler.coro_transform(
params, self.on_progress_updated)
for d in all_data:
self.on_transform_completed(*d)
def on_transform_completed(self, name, times, freq, values, ampl, powers,
avg_ampl, avg_pow):
"""Called when the calculation of the desired transform(s) is completed."""
self.view.on_calculate_stopped()
t = self.signals.get(name)
t.output_data = TFOutputData(times, values, ampl, freq, powers,
avg_ampl, avg_pow)
print(f"Finished calculation for '{name}'.")
# Plot result if all signals finished.
if self.all_transforms_completed():
self.on_all_transforms_completed()
def all_transforms_completed(self):
"""Returns whether all transforms have been completed."""
return all([s.output_data.is_valid() for s in self.signals_calc])
def on_all_transforms_completed(self):
"""Called when all transforms have been completed."""
self.plot_output()
self.on_all_tasks_completed()
def get_values_to_plot(self, amplitude=None) -> tuple:
"""
Returns the data needed to plotting the transform.
:param amplitude: overrides the normal value of whether to plotting amplitude instead of power
:return: the times, frequencies, amplitudes/powers, and average amplitudes/powers
"""
amp: bool = self.plot_ampl
if amplitude is not None:
amp = amplitude
tf_data = self.get_selected_signal().output_data
if not tf_data.is_valid():
return None, None, None, None
if amp: # Plot amplitudes.
values = tf_data.ampl
avg_values = tf_data.avg_ampl
else: # Plot powers.
values = tf_data.powers
avg_values = tf_data.avg_pow
return tf_data.times, tf_data.freq, values, avg_values
def set_plot_type(self, amplitude_selected=True):
"""
Set the type of plotting to display (power or amplitude). This affects
the main plotting and the amplitude plotting.
:param amplitude_selected: whether to set the plotting type as amplitude (not power)
"""
self.plot_ampl = amplitude_selected
if self.is_plotted:
t, f, values, avg_values = self.get_values_to_plot()
self.plot(t, f, values, avg_values)
def plot(self, times, freq, values, avg_values):
self.view.main_plot().plot(times, values, freq)
self.view.amplitude_plot().plot(avg_values, freq)
def plot_output(self):
"""
Plots the output of the WT/WFT calculations for the currently selected signal.
"""
t, f, values, avg_values = self.get_values_to_plot()
if t is not None and f is not None:
self.plot(t, f, values, avg_values)
self.is_plotted = True
else:
self.view.main_plot().clear()
self.view.amplitude_plot().clear()
def get_params(self, all_signals=True) -> TFParams:
"""
Creates the parameters to use when performing the calculations.
"""
if all_signals:
self.signals_calc = self.signals
else:
self.signals_calc = self.signals.only(self.selected_signal_name)
return create(
params_type=TFParams,
signals=self.signals_calc,
fmin=self.view.get_fmin(),
fmax=self.view.get_fmax(),
f0=self.view.get_f0(),
# Only one of these two will be used, depending on the selected transform.
window=self.view.get_wt_wft_type(),
wavelet=self.view.get_wt_wft_type(),
cut_edges=self.view.get_cut_edges(),
preprocess=self.view.get_preprocess(),
transform=self.view.get_transform_type(),
)
def load_data(self):
self.signals = Signals.from_file(self.open_file)
if not self.signals.has_frequency():
freq = FrequencyDialog().run_and_get()
if freq:
self.set_frequency(freq)
self.on_data_loaded()
def plot_signal(self):
"""Plots the signal on the SignalPlot."""
self.view.plot_signal(self.get_selected_signal())
def on_data_loaded(self):
"""Called when the time-series data has been loaded."""
self.view.update_signal_listview(self.signals.names())
self.plot_signal()
def on_signal_zoomed(self, rect: Rect) -> None:
"""
Override callback to also plot the preprocessed version of
the x-limited signal instead of the whole signal.
:param rect: the rectangle which has been zoomed to
"""
super().on_signal_zoomed(rect)
self.plot_preprocessed_signal()
def on_signal_selected(self, item: Union[QListWidgetItem, str]):
"""
Called when a signal is selected in the QListWidget.
Plots the new signal in the top-left plotting and, if
transform data is available, plots the transform and
amplitude/power in the main plots.
"""
if isinstance(item, QListWidgetItem):
name = item.text()
else:
name = item
self.signals.reset()
if name != self.selected_signal_name:
print(f"Selected signal: '{name}'")
self.selected_signal_name = name
self.plot_signal()
self.view.on_xlim_edited()
self.plot_output()
self.plot_preprocessed_signal()
