"""
.. _exa-vol:

Volume source space
===================

.. currentmodule:: eelbrain

Basic analysis of volume source space vector data.

.. contents:: Contents
   :local:

Dataset
^^^^^^^
Load the :mod:`mne` sample data with responses to tones.
Data contains multiple trials for a single subject, with
tones presented to the left or right ear:
"""
# sphinx_gallery_thumbnail_number = 2
# dataset: mne_sample

from pathlib import Path

from eelbrain import *
import mne


# Load the dataset
data = datasets.get_mne_sample(src='vol', ori='vector')
data.head()

###############################################################################
# One-sample test
# ^^^^^^^^^^^^^^^
# A one-sample test can be used to detect significant activations.
# Perform a one-sample test on the right-ear trials only:

result = testnd.Vector('src', sub="side == 'R'", data=data, samples=250, tfce=True, tstart=0.05, tstop=0.200)

###############################################################################
# Morph the data to the template brain for anatomical visualization.

# Store the test result as NDVar
y = result.masked_difference()

# Make sure the source space for fsaverage exists
fname_src_fsaverage = Path(y.source.subjects_dir) / "fsaverage" / "bem" / "fsaverage-vol-7-src.fif"
if fname_src_fsaverage.exists():
    src_fs = mne.read_source_spaces(fname_src_fsaverage)
else:
    src_fs = mne.setup_volume_source_space('fsaverage', 7, subjects_dir=y.source.subjects_dir)
    src_fs.save(Path(y.source.subjects_dir) / "fsaverage" / "bem" / "fsaverage-vol-7-src.fif")

# Compute the transformation from the sample subject to fsaverage
morph = mne.compute_source_morph(
    y.source.get_source_space(),
    subject_from=y.source.subject,
    subjects_dir=y.source.subjects_dir,
    niter_affine=[10, 10, 5],
    niter_sdr=[10, 10, 5],  # just for speed
    src_to=src_fs,
)

# Morph the test result
y = morph_source_space(y, 'fsaverage', morph=morph)

###############################################################################
# For static visualization, we can use a combination of :class:`plot.Butterfly`
# and :class:`plot.GlassBrain` plots.
# A :class:`plot.Butterfly` plot can give a quick overview of amplitudes over
# time:

# Extract vector norm (amplitude)
y_norm = y.norm('space')
# Split data by hemisphere
butterfly_data = [y_norm.sub(source=hemi, name=hemi.capitalize()) for hemi in ['lh', 'rh']]
# Buterfly plot
p = plot.Butterfly(butterfly_data, axh=2, axw=3)
# Mark time points for anatomical visualization
times = [0.090, 0.160]
for t in times:
    p.add_vline(t)

###############################################################################
# :class:`plot.GlassBrain` plots show anatomical distribution of activity at
# relevant time points. Glassbrains project activity through the head, i.e.,
# each voxel in a projection shows the largest value along the line of sight.
# By comparing different projections, the 3D distribution of activity can be
# inferred. Here, the lateral view shows activity in temporal regions over
# auditory cortex, while the coronal view shows how lateral/medial the peaks
# are:

for t in times:
    p = plot.GlassBrain(y.sub(time=t), title=f"{t * 1000:.0f} ms", vmax=4)

###############################################################################
# In a notebook, `LiveNeuron <https://github.com/liang-bo96/LiveNeuron>`_
# can provide interactive visualization. Start the visualization with
# the code below (after uncommenting):

# from eelbrain_plotly_viz import EelbrainPlotly2DViz

# viz = EelbrainPlotly2DViz(result.difference, layout_mode='horizontal', realtime=True, arrow_scale=0.2)
# viz.show_in_jupyter()

###############################################################################
# In an interactive iPython session, butterfly and glassbrain plots can be
# linked for interactive visualization using:

# butterfly, brain = plot.GlassBrain.butterfly(result)