address comments

Author: agramfort

doc/whats_new.rst

@@ -142,7 +142,7 @@ API
 
     - Added :class:`mne.decoding.UnsupervisedSpatialFilter` providing interface for scikit-learn decomposition algorithms to be used with MNE data, by `Jean-Remi King`_ and `Asish Panda`_
 
-    - Deprecated :func:`mne.time_frequency.cwt_morlet` and :func:`mne.time_frequency.single_trial_power` in favour of :func:`mne.time_frequency.tfr_morlet` with parameter average=False, by `Jean-Remi King`_ and 'Alex Gramfort'_
+    - Deprecated :func:`mne.time_frequency.cwt_morlet` and :func:`mne.time_frequency.single_trial_power` in favour of :func:`mne.time_frequency.tfr_morlet` with parameter average=False, by `Jean-Remi King`_ and `Alex Gramfort`_
 
 
 .. _changes_0_12:

mne/epochs.py

@@ -30,7 +30,7 @@
                       pick_channels, pick_info, _pick_data_channels,
                       _pick_aux_channels, _DATA_CH_TYPES_SPLIT)
 from .io.proj import setup_proj, ProjMixin, _proj_equal
-from .io.base import _BaseRaw, ToDataFrameMixin, TimeMixin, SizeMixin
+from .io.base import _BaseRaw, ToDataFrameMixin, TimeMixin
 from .bem import _check_origin
 from .evoked import EvokedArray, _check_decim
 from .baseline import rescale, _log_rescale
@@ -43,7 +43,7 @@
                   plot_epochs_image, plot_topo_image_epochs)
 from .utils import (check_fname, logger, verbose, _check_type_picks,
                     _time_mask, check_random_state, warn, _check_copy_dep,
-                    sizeof_fmt)
+                    sizeof_fmt, SizeMixin)
 from .externals.six import iteritems, string_types
 from .externals.six.moves import zip
 

mne/evoked.py

@@ -16,6 +16,7 @@
 from .filter import resample, detrend, FilterMixin
 from .fixes import in1d
 from .utils import check_fname, logger, verbose, _time_mask, warn, sizeof_fmt
+from .utils import SizeMixin
 from .viz import (plot_evoked, plot_evoked_topomap, plot_evoked_field,
                   plot_evoked_image, plot_evoked_topo)
 from .viz.evoked import (_plot_evoked_white, plot_evoked_joint,
@@ -33,7 +34,7 @@
 from .io.write import (start_file, start_block, end_file, end_block,
                        write_int, write_string, write_float_matrix,
                        write_id)
-from .io.base import ToDataFrameMixin, TimeMixin, SizeMixin
+from .io.base import ToDataFrameMixin, TimeMixin
 
 _aspect_dict = {'average': FIFF.FIFFV_ASPECT_AVERAGE,
                 'standard_error': FIFF.FIFFV_ASPECT_STD_ERR}

mne/io/base.py

@@ -34,45 +34,15 @@
 from ..parallel import parallel_func
 from ..utils import (_check_fname, _check_pandas_installed, sizeof_fmt,
                      _check_pandas_index_arguments, _check_copy_dep,
-                     check_fname, _get_stim_channel, object_hash,
-                     logger, verbose, _time_mask, warn, object_size)
+                     check_fname, _get_stim_channel,
+                     logger, verbose, _time_mask, warn, SizeMixin)
 from ..viz import plot_raw, plot_raw_psd, plot_raw_psd_topo
 from ..defaults import _handle_default
 from ..externals.six import string_types
 from ..event import find_events, concatenate_events
 from ..annotations import _combine_annotations, _onset_to_seconds
 
 
-class SizeMixin(object):
-    """Class to estimate MNE object sizes"""
-    @property
-    def _size(self):
-        """Estimate of the object size"""
-        try:
-            size = object_size(self.info)
-        except Exception:
-            warn('Could not get size for self.info')
-            return -1
-        if hasattr(self, 'data'):
-            size += object_size(self.data)
-        elif hasattr(self, '_data'):
-            size += object_size(self._data)
-        return size
-
-    def __hash__(self):
-        from ..evoked import Evoked
-        from ..epochs import _BaseEpochs
-        if isinstance(self, Evoked):
-            return object_hash(dict(info=self.info, data=self.data))
-        elif isinstance(self, (_BaseEpochs, _BaseRaw)):
-            if not self.preload:
-                raise RuntimeError('Cannot hash %s unless data are loaded'
-                                   % self.__class__.__name__)
-            return object_hash(dict(info=self.info, data=self._data))
-        else:
-            raise RuntimeError('Hashing unknown object type: %s' % type(self))
-
-
 class ToDataFrameMixin(object):
     """Class to add to_data_frame capabilities to certain classes."""
     def _get_check_picks(self, picks, picks_check):

mne/time_frequency/tests/test_tfr.py

@@ -33,8 +33,7 @@ def test_morlet():
 
 
 def test_time_frequency():
-    """Test the to-be-deprecated time frequency transform (PSD and ITC).
-    """
+    """Test the to-be-deprecated time frequency transform (PSD and ITC)"""
     # Set parameters
     event_id = 1
     tmin = -0.2
@@ -278,6 +277,8 @@ def test_tfr_multitaper():
                                   n_cycles=freqs / 2., time_bandwidth=4.0,
                                   return_itc=False, average=False).average()
 
+    print(power_evoked)  # test repr for EpochsTFR
+
     assert_raises(ValueError, tfr_multitaper, epochs,
                   freqs=freqs, n_cycles=freqs / 2.,
                   return_itc=True, average=False)
@@ -451,6 +452,7 @@ def test_add_channels():
 
 
 def test_compute_tfr():
+    """Test _compute_tfr function"""
     # Set parameters
     event_id = 1
     tmin = -0.2

mne/time_frequency/tfr.py

@@ -19,11 +19,13 @@
 from ..fixes import partial
 from ..baseline import rescale
 from ..parallel import parallel_func
-from ..utils import logger, verbose, _time_mask, check_fname, deprecated
+from ..utils import (logger, verbose, _time_mask, check_fname, deprecated,
+                     sizeof_fmt)
 from ..channels.channels import ContainsMixin, UpdateChannelsMixin
 from ..channels.layout import _pair_grad_sensors
 from ..io.pick import pick_info, pick_types
 from ..io.meas_info import Info
+from ..utils import SizeMixin
 from .multitaper import dpss_windows
 from ..viz.utils import figure_nobar, plt_show
 from ..externals.h5io import write_hdf5, read_hdf5
@@ -249,7 +251,7 @@ def _cwt(X, Ws, mode="same", decim=1, use_fft=True):
 
 
 def _compute_tfr(epoch_data, frequencies, sfreq=1.0, method='morlet',
-                 n_cycles=7.0, zero_mean=None, time_bandwidth=4.0,
+                 n_cycles=7.0, zero_mean=None, time_bandwidth=None,
                  use_fft=True, decim=1, output='complex', n_jobs=1,
                  verbose=None):
     """Computes time-frequency transforms.
@@ -271,7 +273,8 @@ def _compute_tfr(epoch_data, frequencies, sfreq=1.0, method='morlet',
     zero_mean : bool | None, defaults to None
         None means True for method='multitaper' and False for method='morlet'.
         If True, make sure the wavelets have a mean of zero.
-    time_bandwidth : float, defaults to 4.0 (3 tapers)
+    time_bandwidth : float, defaults to None
+        If method=multitaper, will be set to 4.0 (3 tapers).
         Time x (Full) Bandwidth product. Only applies if
         method == 'multitaper'. The number of good tapers (low-bias) is
         chosen automatically based on this to equal floor(time_bandwidth - 1).
@@ -296,8 +299,8 @@ def _compute_tfr(epoch_data, frequencies, sfreq=1.0, method='morlet',
           coherence across trials.
 
     n_jobs : int, defaults to 1
-        The number of epochs to process at the same time. The parallization is
-        implemented across channels.
+        The number of epochs to process at the same time. The parallelization
+        is implemented across channels.
     verbose : bool, str, int, or None, defaults to None
         If not None, override default verbose level (see mne.verbose).
 
@@ -341,7 +344,8 @@ def _compute_tfr(epoch_data, frequencies, sfreq=1.0, method='morlet',
     # Default zero_mean = True if multitaper else False
     zero_mean = method == 'multitaper' if zero_mean is None else zero_mean
     if not isinstance(zero_mean, bool):
-        raise ValueError('')
+        raise ValueError('zero_mean should be of type bool. Got %s.'
+                         % type(zero_mean))
     frequencies = np.asarray(frequencies)
 
     # XXX Can we compute single-trial phases with multitaper?
@@ -354,10 +358,12 @@ def _compute_tfr(epoch_data, frequencies, sfreq=1.0, method='morlet',
     if method == 'morlet':
         W = morlet(sfreq, frequencies, n_cycles=n_cycles, zero_mean=zero_mean)
         Ws = [W]  # to have same dimensionality as the 'multitaper' case
-        if time_bandwidth != 4.0:
+        if time_bandwidth is not None:
             raise ValueError('time_bandwidth only applies to "multitaper"'
                              ' method.')
     elif method == 'multitaper':
+        if time_bandwidth is None:
+            time_bandwidth = 4.0
         Ws = _make_dpss(sfreq, frequencies, n_cycles=n_cycles,
                         time_bandwidth=time_bandwidth, zero_mean=zero_mean)
 
@@ -395,6 +401,7 @@ def _compute_tfr(epoch_data, frequencies, sfreq=1.0, method='morlet',
         out[channel_idx] = tfr
 
     if ('avg_' not in output) and ('itc' not in output):
+        # This is to enforce that the first dimension is for epochs
         out = out.transpose(1, 0, 2, 3)
     return out
 
@@ -759,11 +766,11 @@ def tfr_morlet(inst, freqs, n_cycles, use_fft=False, return_itc=True, decim=1,
     zero_mean : bool, defaults to True
         Make sure the wavelet has a mean of zero.
 
-        .. versionadded:: 0.12.0
+        .. versionadded:: 0.13.0
     average : bool, defaults to True
         If True average accross Epochs.
 
-        .. versionadded:: 0.12.0
+        .. versionadded:: 0.13.0
     verbose : bool, str, int, or None, defaults to None
         If not None, override default verbose level (see mne.verbose).
 
@@ -825,7 +832,7 @@ def tfr_multitaper(inst, freqs, n_cycles, time_bandwidth=4.0,
     average : bool, defaults to True
         If True average accross Epochs.
 
-        .. versionadded:: 0.12.0
+        .. versionadded:: 0.13.0
     verbose : bool, str, int, or None, defaults to None
         If not None, override default verbose level (see mne.verbose).
 
@@ -853,7 +860,7 @@ def tfr_multitaper(inst, freqs, n_cycles, time_bandwidth=4.0,
 
 # TFR(s) class
 
-class _BaseTFR(ContainsMixin, UpdateChannelsMixin):
+class _BaseTFR(ContainsMixin, UpdateChannelsMixin, SizeMixin):
     @property
     def ch_names(self):
         return self.info['ch_names']
@@ -1419,6 +1426,7 @@ def __repr__(self):
         s += ", freq : [%f, %f]" % (self.freqs[0], self.freqs[-1])
         s += ", nave : %d" % self.nave
         s += ', channels : %d' % self.data.shape[0]
+        s += ', ~%s' % (sizeof_fmt(self._size),)
         return "<AverageTFR  |  %s>" % s
 
     def save(self, fname, overwrite=False):
@@ -1464,7 +1472,7 @@ class EpochsTFR(_BaseTFR):
 
     Notes
     -----
-    .. versionadded:: 0.12.0
+    .. versionadded:: 0.13.0
     """
     @verbose
     def __init__(self, info, data, times, freqs, comment=None,
@@ -1493,6 +1501,7 @@ def __repr__(self):
         s += ", freq : [%f, %f]" % (self.freqs[0], self.freqs[-1])
         s += ", epochs : %d" % self.data.shape[0]
         s += ', channels : %d' % self.data.shape[1]
+        s += ', ~%s' % (sizeof_fmt(self._size),)
         return "<EpochsTFR  |  %s>" % s
 
     def average(self):

mne/utils.py

@@ -1727,6 +1727,37 @@ def sizeof_fmt(num):
         return '1 byte'
 
 
+class SizeMixin(object):
+    """Class to estimate MNE object sizes"""
+    @property
+    def _size(self):
+        """Estimate of the object size"""
+        try:
+            size = object_size(self.info)
+        except Exception:
+            warn('Could not get size for self.info')
+            return -1
+        if hasattr(self, 'data'):
+            size += object_size(self.data)
+        elif hasattr(self, '_data'):
+            size += object_size(self._data)
+        return size
+
+    def __hash__(self):
+        from ..evoked import Evoked
+        from ..epochs import _BaseEpochs
+        from ..io.base import _BaseRaw
+        if isinstance(self, Evoked):
+            return object_hash(dict(info=self.info, data=self.data))
+        elif isinstance(self, (_BaseEpochs, _BaseRaw)):
+            if not self.preload:
+                raise RuntimeError('Cannot hash %s unless data are loaded'
+                                   % self.__class__.__name__)
+            return object_hash(dict(info=self.info, data=self._data))
+        else:
+            raise RuntimeError('Hashing unknown object type: %s' % type(self))
+
+
 def _url_to_local_path(url, path):
     """Mirror a url path in a local destination (keeping folder structure)"""
     destination = urllib.parse.urlparse(url).path