kwargs & bugs (#40)

* exclude dynamicripple from pytest * test workflow * add flake8 config file * exclude soundfile from flake8 * exclude soundcard from flake8 * ignore soundcard for pytest * ignore soundfile for pytest * fix bu in test routine * don't check sound level * add testing for filters * change filter documentation * import re * change filters to filter in documentation * change f_lower to low_cutoff * remove brownnoise * ignore build * adding filter doc * ignore build * ignore build * ignore build * removing _build folder * convert tabs to spaces * debug calibrate * remove time_windows from test * remove matlabdoc builder * add matlabdoc builder to requirements * remove matlabdoc * resolve some warnings * syntax fixes * . * test recording * add play method * do test * move test folder one level up * addplotting code in the .rst file * one mire plotting example * add filter docu * add example for equalization filter * fix plotting * fix plotting * remove pyplots folder * debug plotting * ignoring cached scripts * simplify plot_tf * change header * remake intro * add reference for spectral features * move audiogram to examples * add some references * add fft filter test * chnage header * merge * fix reference * add psychoacoustics test * integrate mmn in Trialsequence class * for mmn sequence n_reps = n_trials * condition numbers start at 1 * condtions start at 1 * rename function to oddball, encode deviant as 0 * change oddball to deviant_indices * deviants should work now * make random permutation if n_conds = 1 * give correct trialnr * write response in list of lists * make list of None elementrs fro response * make list of None for response * add deviants to n conditions * fix bug * trial sequence should always be a list * dont insert empty array to avoid warning * add test routines * more tests * fix bug when loading sequence * fix import order * copy _init_ from upstream * all functions except curses now tested * fix syntax error * use uniform syntax for plotting * add hrtf tests * add h5netcdf * add SoundFile * add sudo apt-get install sndfile-tools * fix syntax * separate system and python dependencies * fix syntax error * return condition in __next__ * fix bug in deviant list generation * in self.data change None to -999 to avoid error when writing json files * make triallist python int * syntax error * return 0 in __next__() when trial is deviant * resolve type error * never modify anything in place, remove copy_channel * fix syntax error * remove copychannel, add am method * changing doku * overhaul of the psychoacoustics documentation * fix flake8 error * fix syntax error * fix assertion error * remove redundant function * writing stuff * add bibliography * change paper * new paper version * new unittests for Trialsequence + debugging * debug variable name * convert trials to list * raise error when condition is a string but not a file * add method for saving and loading pickle files and rename to LoadSave_mixin * when writing JSON, catch value AND type errors * when making a Trialsequence instance from a file, try to load JSON, catch exception and then load pickle * catch JSONDecodeError when loading trialsequence from file * JSON decode error is part of json module * fix indentation error * change class names and add docstrings * rename classes according to python conventions * removing attributes this_trial_n and this_rep_n from trialsequence * adding docstrings * add docstrings * adding docstrings * docstrings * move simulate_response to mixin * refactoring docstrings * debug test_psychoacoustics * avoid endless loop when computing deviant indices * enable plotting to axis when plotting an "image" * change plotting routine to avoid timeout when pytesting * add docstrings * change nchannels to n_channels and nsamples to n_samples * change Signal.resize to return a new instance * changed the crossfade function to deal with multiple inputs (still a BUG) * dont transpose array if data array is empty * add unit test for crossfading * docstrings * add test for frames and frametimes * finished docstring overhaul * add tests for signal.py * add test routine for generating and reading/writing sound * add test for power law noise * add tests for binaural cues * pack get and set envelope into one function * change envelope and itd/ild to get/set scheme * add docstrings for binaural generator functions * adding tests for band filters * finish filter class refractor * dont use unittest class * debug test_tone * add type in docstring * negative ild means sound to the left * add test for loading/saving filters * add test routines for externalize and interaural level spectrum * add the interaural level spectrum file to gitignore * add test for making HRTF instances * overhaul hrtf docstrings & tests * resolve some TODOs * remove normalise argument * debuging tests * debug endless loop * debig test_envelope * debug assertion errors * debug assertion errors * change filter n_samples * debug assertion error * renamed the n_ parameters * debug sound docs * implement writing objects to resultsfile * debugging docs * debug save results file * add import statements to docs * debug tests * debug test frames * add default level, changel handling of default sampling rate * hide default level and samplerate variables * remove conditions from setup dependencies * remove level argument from all functions * add SoudCard * remove soundfile * remove soundcard from setup.py * install pulseaudio * debug * remove SounCard from pipeline * debug test * debug * debug get_future_trial * improve documentation * uninstall soundcard after setup.py * add soundcard to requirements * say yes to pip uninstall * fix some warnings from pytest * add error message if libsndfile is not installed * add data_path function * change usage of data_path in docs * Update paper.bib * Update paper.bib * dont print every iteration in Trialsequence.transitions * some small fixes in psychoacoustics docs * catch h5py warning when loading sofa file * change docstrings for proper sphinx rendering * add level argument to generator methods * add kemar * remove all plot kwargs and add kwargs to spectrogram * debug binaural.externalize * debug hrtf plot_tf * make assert condition more liberal Co-authored-by: Marc <[email protected]>
DrMarc · May 28, 2021 · b5b11cf · b5b11cf
1 parent 21a81dc
commit b5b11cf
Show file tree

Hide file tree

Showing 7 changed files with 30 additions and 34 deletions.
diff --git a/slab/binaural.py b/slab/binaural.py
@@ -292,7 +292,7 @@ def externalize(self, hrtf=None):
         idx_frontal = numpy.where((hrtf.sources[:, 1] == 0) & (hrtf.sources[:, 0] == 0))[0][0]
         if not idx_frontal.size: # idx_frontal is empty
             raise ValueError('No frontal direction [0,0] found in HRTF.')
-        _, h = hrtf.data[idx_frontal].tf(channels=0, nbins=12, show=False)  # get low-res version of HRTF spectrum
+        _, h = hrtf.data[idx_frontal].tf(channels=0, n_bins=12, show=False)  # get low-res version of HRTF spectrum
         h[0] = 1  # avoids low-freq attenuation in KEMAR HRTF (unproblematic for other HRTFs)
         resampled_signal = copy.deepcopy(self)
         # if sound and HRTF has different samplerates, resample the sound, apply the HRTF, and resample back:

diff --git a/slab/filter.py b/slab/filter.py
@@ -189,7 +189,7 @@ def apply(self, sig):
                     'Number of filters must equal number of sound channels, or either one of them must be equal to 1.')
         return out
 
-    def tf(self, channels='all', n_bins=None, show=True, axis=None, **kwargs):
+    def tf(self, channels='all', n_bins=None, show=True, axis=None):
         """
         Compute a filter's transfer function (magnitude over frequency) and optionally plot it.
 
@@ -201,7 +201,6 @@ def tf(self, channels='all', n_bins=None, show=True, axis=None, **kwargs):
                 If None, use the maximum number of bins.
             show (bool): whether to show the plot right after drawing.
             axis (matplotlib.axes.Axes | None): axis to plot to. If None create a new plot.
-            ** kwargs: keyword arguments for the plot, see documentation of matplotlib.pyplot.plot for details.
         Returns:
             (numpy.ndarray): the frequency bins in the range from 0 Hz to the Nyquist frequency.
             (numpy.ndarray: the magnitude of each frequency in `w`.
@@ -253,7 +252,7 @@ def tf(self, channels='all', n_bins=None, show=True, axis=None, **kwargs):
                 raise ImportError('Plotting transfer functions requires matplotlib.')
             if axis is None:
                 _, axis = plt.subplots()
-            axis.plot(w, h, **kwargs)
+            axis.plot(w, h)
             axis.set(title='Frequency [Hz]', xlabel='Amplitude [dB]', ylabel='Frequency Response')
             axis.grid(True)
             if show:

diff --git a/slab/hrtf.py b/slab/hrtf.py
@@ -308,7 +308,7 @@ def plot_tf(self, sourceidx, ear='left', xlim=(1000, 18000), n_bins=None, kind='
             vlines = numpy.arange(0, len(sourceidx)) * linesep
             for idx, s in enumerate(sourceidx):
                 filt = self.data[s]
-                freqs, h = filt.tf(channels=chan, nbins=n_bins, show=False)
+                freqs, h = filt.tf(channels=chan, n_bins=n_bins, show=False)
                 axis.plot(freqs, h + vlines[idx],
                           linewidth=0.75, color='0.0', alpha=0.7)
             ticks = vlines[::3]  # plot every third elevation
@@ -329,7 +329,7 @@ def plot_tf(self, sourceidx, ear='left', xlim=(1000, 18000), n_bins=None, kind='
             elevations = self.sources[sourceidx, 1]
             for idx, source in enumerate(sourceidx):
                 filt = self.data[source]
-                freqs, h = filt.tf(channels=chan, nbins=n_bins, show=False)
+                freqs, h = filt.tf(channels=chan, n_bins=n_bins, show=False)
                 img[:, idx] = h.flatten()
             img[img < -25] = -25  # clip at -40 dB transfer
             contour = axis.contourf(freqs, elevations, img.T, cmap='hot', origin='upper', levels=20)

diff --git a/slab/psychoacoustics.py b/slab/psychoacoustics.py
@@ -595,20 +595,19 @@ def response_summary(self):
         else:
             return None
 
-    def plot(self, axis=None, show=True, **kwargs):
+    def plot(self, axis=None, show=True):
         """
         Plot the trial sequence as scatter plot.
 
         Arguments:
             axis (matplotlib.pyplot.Axes): plot axis to draw on, if none a new plot is generated
             show (bool): show the plot immediately, defaults to True
-            **kwargs: pyplot.plot keyword arguments, see matplotlib documentation
         """
         if plt is None:
             raise ImportError('Plotting requires matplotlib!')
         if axis is None:
             axis = plt.subplot()
-        axis.scatter(range(self.n_trials), self.trials, **kwargs)
+        axis.scatter(range(self.n_trials), self.trials)
         axis.set(title='Trial sequence', xlabel='Trials', ylabel='Condition index')
         if show:
             plt.show()
@@ -864,14 +863,13 @@ def save_csv(self, filename):
             f.write(responses)
         return True
 
-    def plot(self, axis=None, show=True, **kwargs):
+    def plot(self, axis=None, show=True):
         """
         Plot the staircase. If called after each trial, one plot is created and updated.
 
         Arguments:
             axis (matplotlib.pyplot.Axes): plot axis to draw on, if none a new plot is generated
             show (bool): whether to show the plot right after drawing.
-            **kwargs: pyplot.plot keyword arguments, see matplotlib documentation.
         """
         if plt is None:
             raise ImportError('Plotting requires matplotlib!')
@@ -883,7 +881,7 @@ def plot(self, axis=None, show=True, **kwargs):
                 fig = plt.figure('stairs')  # figure 'stairs' is created or made current
                 axis = fig.gca()
             axis.clear()
-            axis.plot(x, y, **kwargs)
+            axis.plot(x, y)
             axis.set_xlim(-self.n_pretrials, max(20, (self.this_trial_n + 15)//10*10))
             axis.set_ylim(min(0, min(y)) if self.min_val == -numpy.Inf else self.min_val,
                           max(y) if self.max_val == numpy.Inf else self.max_val)

diff --git a/slab/signal.py b/slab/signal.py
@@ -403,24 +403,23 @@ def delay(self, duration=1, channel=0, filter_length=2048):
                     new.data[i, channel] = numpy.convolve(sig_portion, tap_weight, mode='valid')
         return new
 
-    def plot_samples(self, show=True, axis=None, **kwargs):
+    def plot_samples(self, show=True, axis=None):
         """
         Stem plot of the samples of the signal.
 
         Arguments:
             show (bool): whether to show the plot right after drawing.
             axis (matplotlib.axes.Axes | None): axis to plot to. If None create a new plot.
-            ** kwargs: keyword arguments for the plot, see documentation of matplotlib.pyplot.plot for details.
         """
         if matplotlib is False:
             raise ImportError('Plotting signals requires matplotlib.')
         if axis is None:
             _, axis = plt.subplots()
         if self.n_channels == 1:
-            axis.stem(self.channel(0), **kwargs)
+            axis.stem(self.channel(0))
         else:
             for i in range(self.n_channels):
-                axis.stem(self.channel(i), label=f'channel {i}', **kwargs)
+                axis.stem(self.channel(i), label=f'channel {i}')
             plt.legend()
         axis.set(title='Samples', xlabel='Number', ylabel='Value')
         if show:

diff --git a/slab/sound.py b/slab/sound.py
@@ -925,7 +925,7 @@ def play_file(filename):
                     'Playing from files on Linux without SoundCard module requires SoX. '
                     'Install: sudo apt-get install sox libsox-fmt-all or pip install SoundCard')
 
-    def waveform(self, start=0, end=None, show=True, axis=None, **kwargs):
+    def waveform(self, start=0, end=None, show=True, axis=None):
         """
         Plot the waveform of the sound.
 
@@ -934,7 +934,6 @@ def waveform(self, start=0, end=None, show=True, axis=None, **kwargs):
             end (int | float | None): the end of the plot in seconds (float) or samples (int), defaults to None.
             show (bool): whether to show the plot right after drawing.
             axis (matplotlib.axes.Axes | None): axis to plot to. If None create a new plot.
-            ** kwargs: keyword arguments for the plot, see documentation of matplotlib.pyplot.plot for details.
         """
         if matplotlib is False:
             raise ImportError('Plotting waveforms requires matplotlib.')
@@ -945,14 +944,14 @@ def waveform(self, start=0, end=None, show=True, axis=None, **kwargs):
         if axis is None:
             _, axis = plt.subplots()
         if self.n_channels == 1:
-            axis.plot(self.times[start:end], self.channel(0)[start:end], **kwargs)
+            axis.plot(self.times[start:end], self.channel(0)[start:end])
         elif self.n_channels == 2:
-            axis.plot(self.times[start:end], self.channel(0)[start:end], label='left', **kwargs)
-            axis.plot(self.times[start:end], self.channel(1)[start:end], label='right', **kwargs)
+            axis.plot(self.times[start:end], self.channel(0)[start:end], label='left')
+            axis.plot(self.times[start:end], self.channel(1)[start:end], label='right')
             axis.legend()
         else:
             for i in range(self.n_channels):
-                axis.plot(self.times[start:end], self.channel(i)[start:end], label=f'channel {i}', **kwargs)
+                axis.plot(self.times[start:end], self.channel(i)[start:end], label=f'channel {i}')
             plt.legend()
         axis.set(title='Waveform', xlabel='Time [sec]', ylabel='Amplitude')
         if show:
@@ -971,7 +970,7 @@ def spectrogram(self, window_dur=0.005, dyn_range=120, upper_frequency=None, oth
             show (bool): whether to show the plot right after drawing. Note that if show is False and no `axis` is
                 passed, no plot will be created.
             axis (matplotlib.axes.Axes | None): axis to plot to. If None create a new plot.
-            **kwargs: keyword arguments for the plot. See documentation for matplotlib.pyplot.imshow.
+            **kwargs: keyword arguments for computing the spectrogram. See documentation for scipy.signal.spectrogram.
         Returns:
             (None | tuple): If `show == True` or an axis was passed, a plot is drawn and nothing is returned. Else,
                 a tuple is returned which contains frequencies, time bins and a 2D array of powers.
@@ -993,10 +992,13 @@ def spectrogram(self, window_dur=0.005, dyn_range=120, upper_frequency=None, oth
         window = scipy.signal.windows.gaussian(window_n_samples, window_sigma)
         # convert step size into number of overlapping samples in adjacent analysis frames
         n_overlap = window_n_samples - step_n_samples
-        # compute the power spectral density
-        freqs, times, power = scipy.signal.spectrogram(
-            x, mode='psd', fs=self.samplerate, scaling='density', noverlap=n_overlap, window=window,
-            nperseg=window_n_samples)
+        # compute the power spectral density, use default configuration if no values were specified
+        kwargs.setdefault("mode", "psd")
+        kwargs.setdefault("scaling", "density")
+        kwargs.setdefault("noverlap", n_overlap)
+        kwargs.setdefault("window", window)
+        kwargs.setdefault("nperseg", window_n_samples)
+        freqs, times, power = scipy.signal.spectrogram(x, fs=self.samplerate, **kwargs)
         if show or (axis is not None):
             if matplotlib is False:
                 raise ImportError('Ploting spectrograms requires matplotlib.')
@@ -1011,14 +1013,14 @@ def spectrogram(self, window_dur=0.005, dyn_range=120, upper_frequency=None, oth
             if axis is None:
                 _, axis = plt.subplots()
             axis.imshow(power, origin='lower', aspect='auto',
-                        cmap=cmap, extent=extent, vmin=vmin, vmax=None, **kwargs)
+                        cmap=cmap, extent=extent, vmin=vmin, vmax=None)
             axis.set(title='Spectrogram', xlabel='Time [sec]', ylabel='Frequency [Hz]')
             if show:
                 plt.show()
         else:
             return freqs, times, power
 
-    def cochleagram(self, bandwidth=1 / 5, show=True, axis=None, **kwargs):
+    def cochleagram(self, bandwidth=1 / 5, show=True, axis=None):
         """
         Computes a cochleagram of the sound by filtering with a bank of cosine-shaped filters with given bandwidth
         and applying a cube-root compression to the resulting envelopes.
@@ -1028,7 +1030,6 @@ def cochleagram(self, bandwidth=1 / 5, show=True, axis=None, **kwargs):
             show (bool): whether to show the plot right after drawing. Note that if show is False and no `axis` is
                 passed, no plot will be created
             axis (matplotlib.axes.Axes | None): axis to plot to. If None create a new plot.
-            **kwargs: keyword arguments for the plot. See documentation for matplotlib.pyplot.imshow.
         Returns:
             (None | numpy.ndarray): If `show == True` or an axis was passed, a plot is drawn and nothing is returned.
                 Else, an array with the envelope is returned.
@@ -1057,7 +1058,7 @@ def cochleagram(self, bandwidth=1 / 5, show=True, axis=None, **kwargs):
         else:
             return envs
 
-    def spectrum(self, low_cutoff=16, high_cutoff=None, log_power=True, axis=None, show=True, **kwargs):
+    def spectrum(self, low_cutoff=16, high_cutoff=None, log_power=True, axis=None, show=True):
         """
         Compute the spectrum of the sound.
 
@@ -1069,7 +1070,6 @@ def spectrum(self, low_cutoff=16, high_cutoff=None, log_power=True, axis=None, s
             show (bool): whether to show the plot right after drawing. Note that if show is False and no `axis` is
                 passed, no plot will be created.
             axis (matplotlib.axes.Axes | None): axis to plot to. If None create a new plot.
-            **kwargs: keyword arguments for the plot. see documentation for matplotlib.pyplot.semilogx.
         Returns:
             If show=False, returns `Z, freqs`, where `Z` is a 1D array of powers
                 and `freqs` are the corresponding frequencies.n
@@ -1100,7 +1100,7 @@ def spectrum(self, low_cutoff=16, high_cutoff=None, log_power=True, axis=None, s
                 raise ImportError('Plotting spectra requires matplotlib.')
             if axis is None:
                 _, axis = plt.subplots()
-            axis.semilogx(freqs, Z, **kwargs)
+            axis.semilogx(freqs, Z)
             ticks_freqs = numpy.round(32000 * 2 **
                                       (numpy.arange(12, dtype=float) * -1))
             axis.set_xticks(ticks_freqs)

diff --git a/tests/test_binaural.py b/tests/test_binaural.py
@@ -83,7 +83,7 @@ def test_externalize():
         filtered = hrtf.data[idx_frontal].apply(sound)
         external = sound.externalize()
         assert numpy.abs(filtered.data-external.data).sum() < numpy.abs(filtered.data-sound.data).sum()
-        assert numpy.abs(sound.level - external.level).max() < 0.25
+        assert numpy.abs(sound.level - external.level).max() < 0.6
 
 
 def test_interaural_level_spectrum():