class MainWindow(QtGui.QWidget):
def __init__(self):
super(MainWindow, self).__init__()
self.initUI()
def initUI(self, params=None):
if params == None:
self.params = {
"subjects": [
"P03",
"P04",
"P05",
"P06",
"P07",
"P08",
"P09",
"eeg_plot0",
"eeg_plot1",
"eeg_plot2",
"eeg_plot3",
"eeg_plot4",
"eeg_plot5",
"eeg_plot6",
"eeg_plot7",
],
"channels": ["Cz", "Pz", "Fz"],
"base_offset": 0,
"chan_offset": 150,
"gain": 1,
"features": ["theta", "alpha", "beta", "means", "kurtosis", "skewness", "svd_entropy"],
#'features' : ['alpha', 'means', 'entropy', 'kurtosis'],
"envir": ["luxmetre", "sonometre"],
"data_repo": "./data/",
"fe": 256.0, # Hz
"featCalc_win_size": 30.0, # s
"featCalc_win_overlap": 1.0 / 2,
"eegView_win_size": 30, # s
"feat_offset": 1, # because feat data are normalized
}
self.selected = collections.OrderedDict(
{
"subject": "P03",
"channels": [2, 2, 2],
"features": [2, 2, 2, 2, 2, 2, 2],
"feat_chan": "Cz",
"envir": [2, 2],
}
)
self.last_selected = self.selected.copy()
# Usefull params
self.nb_win_EEG_values = self.params["eegView_win_size"] * self.params["fe"]
print "nb EEG values : ", self.nb_win_EEG_values
self.EEG_Xpos = np.round(self.nb_win_EEG_values / 2)
print "EEG_Xpos: ", self.EEG_Xpos
self.Feat_Xpos = 0
self.dialogPlotEdit = DialogWindow(self, self.params, self.selected)
# Main Window and Layout
self.main_win = QtGui.QWidget()
self.main_win.setWindowTitle("EEG Super Viewer")
self.main_win.resize(1200, 1000)
self.main_layout = QtGui.QGridLayout()
self.main_win.setLayout(self.main_layout)
# Graph Properties Options
self.prop_layout = QtGui.QHBoxLayout()
# Subject selected
vb_txt_subject = QtGui.QLabel("Subject selected :")
self.prop_layout.addWidget(vb_txt_subject)
subject_combo = QtGui.QComboBox()
subject_combo.addItems(self.params["subjects"])
# TODO set combo initial to self.selected['subject']
self.prop_layout.addWidget(subject_combo)
self.connect(subject_combo, QtCore.SIGNAL("activated(QString)"), self.change_subject)
# EEG Time #TODO
vb_time_EEG = QtGui.QLabel("EEG time :")
self.prop_layout.addWidget(vb_time_EEG)
# Edition windows
self.EditButton = QtGui.QPushButton("Edit Plots properties")
self.EditButton.clicked.connect(self.on_push_Edit_Button)
self.prop_layout.addWidget(self.EditButton)
self.main_layout.addLayout(self.prop_layout, 0, 0)
# ~ self.dialogPlotEdit = DialogWindow(self, self.params, self.selected)
# Grapfics window
self.gWin = pg.GraphicsWindow()
# ~ self.label = pg.LabelItem(justify='right')
# ~ self.gWin.addItem(self.label)
axisEEG = EEGAxis(
sampling_rate=self.params["fe"], eegView_win_size=self.params["eegView_win_size"], orientation="bottom"
)
# eeg_plot and feat_plot are viewbox
self.eeg_plot = self.gWin.addPlot(row=1, col=0, axisItems={"bottom": axisEEG})
self.eeg_plot.invertY()
self.eeg_plot.setMouseEnabled(x=True, y=True)
##axis in date
s_rate = 1 / (self.params["featCalc_win_size"] * self.params["featCalc_win_overlap"])
print "sample_rate : ", 1 / self.params["featCalc_win_size"]
axisFeat = DateAxisFeat(rec_datetime0=dt.datetime.now(), sampling_rate=s_rate, orientation="bottom")
self.feat_plot = self.gWin.addPlot(row=2, col=0, axisItems={"bottom": axisFeat})
self.feat_plot.setMouseEnabled(x=True, y=True)
self.main_layout.addWidget(self.gWin, 1, 0)
# Line EEG : always in the center
self.vLine_eeg = pg.InfiniteLine(pos=np.round(self.nb_win_EEG_values / 2), angle=90, movable=False)
self.eeg_plot.addItem(self.vLine_eeg, ignoreBounds=True)
# Line feat (movable and update EEG view)
self.vLine_feat = pg.InfiniteLine(pos=self.Feat_Xpos, angle=90, movable=True)
self.feat_plot.addItem(self.vLine_feat, ignoreBounds=True)
self.vLine_feat.sigPositionChangeFinished.connect(self.update_cursor_pos)
# Load and init plots
self.nbEEGsamples = 0 # updated in loadSubjectdata
self.nbFeatSamples = 0
self.env_sigs = 0
self.loadSubjectdata()
self.update_eeg_plots()
self.load_feat_plots()
self.main_win.show()
def on_push_Edit_Button(self):
print "Properties Pushed"
self.dialogPlotEdit.exec_()
def on_valid_dialogWindow(self):
print "validation dialogWindow"
self.dialogPlotEdit.close()
def update_cursor_pos(self, evt):
win = self.params["featCalc_win_size"]
featCalc_win_overlap = self.params["featCalc_win_overlap"]
fe = self.params["fe"]
# Or int, depends how you want to proceed betwen feat dots ?
self.Feat_Xpos = np.round(self.vLine_feat.value())
if self.Feat_Xpos < 0:
self.Feat_Xpos = 0
if self.Feat_Xpos > self.nbEEGsamples:
self.Feat_Xpos = self.nbEEGsamples
print "new feat pos is : ", self.Feat_Xpos # TODO put it in time
self.EEG_Xpos = self.Feat_Xpos * (win * featCalc_win_overlap * fe)
print "new eeg pos is : ", self.EEG_Xpos # TODO put it in time
self.update_eeg_plots()
def load_feat_plots(self):
nb_feat = len(self.params["features"])
self.feat_plot.clear()
for j in range(nb_feat):
data = self.feat_df[self.params["features"][j]][self.selected["subject"]][
self.selected["feat_chan"]
].values.reshape((self.nbFeatSamples))
min_feat = min(data)
max_feat = np.percentile(data, 90)
data = (data - max_feat) / (max_feat - min_feat) + (nb_feat - j) * self.params["feat_offset"]
print np.shape(data)[0]
self.feat_plot.plot(x=range(np.shape(data)[0]), y=data, pen=(j, nb_feat * 1.3))
self.vLine_feat.setValue(self.Feat_Xpos)
self.feat_plot.addItem(self.vLine_feat, ignoreBounds=True)
self.feat_plot.setLimits(yMin=0, yMax=nb_feat + 1)
def update_eeg_plots(self):
print "update EEG plots, self.EEG_Xpos is :", self.EEG_Xpos
# EEG params
nb_chan = len(self.params["channels"])
base_offset = self.params["base_offset"]
chan_offset = self.params["chan_offset"]
gain = self.params["gain"]
x0 = np.int(self.EEG_Xpos - self.nb_win_EEG_values / 2)
x1 = np.int(self.EEG_Xpos + self.nb_win_EEG_values / 2)
print "x0 : ", x0
print "x1 : ", x1
eeg_win = np.empty((x1 - x0))
# EEG Lines plots
self.eeg_plot.clear()
for i, name_chan in enumerate(self.params["channels"]):
eeg_win = self.eeg_df[name_chan][x0:x1].values * gain + (nb_chan - i) * chan_offset
self.eeg_plot.plot(eeg_win, pen="g")
# Update EEG central Line
self.eeg_plot.addItem(self.vLine_eeg, ignoreBounds=True)
def change_subject(self, text):
self.selected["subject"] = str(text)
print "Patient selected : ", self.selected["subject"]
self.loadSubjectdata()
self.update_eeg_plots()
self.load_feat_plots()
def loadSubjectdata(self):
# load eeg data
print "subject file : ", join(
self.params["data_repo"], self.selected["subject"], self.selected["subject"] + ".h5"
)
eeg_Store = pd.HDFStore(
join(self.params["data_repo"], self.selected["subject"], self.selected["subject"] + ".h5")
)
self.eeg_df = eeg_Store.df
# load features data
self.feat_df = {}
for feat in self.params["features"]:
print "file : ", join(self.params["data_repo"], feat + ".h5")
feat_Store = pd.HDFStore(join(self.params["data_repo"], feat + ".h5")) # read_hdf ??
self.feat_df[feat] = feat_Store.df.dropna()
# load environment data
env_raw = join(self.params["data_repo"], self.selected["subject"], self.selected["subject"] + ".raw")
env_raw_header = join(self.params["data_repo"], self.selected["subject"], self.selected["subject"] + ".header")
d = self.read_header(env_raw_header)
# (re)set params
self.nbEEGsamples = np.shape(self.eeg_df["Cz"])[0]
print "size EEG : ", self.nbEEGsamples
self.nbFeatSamples = np.shape(self.feat_df[self.params["features"][0]])[0]
print "nb feat samples : ", self.nbFeatSamples
self.env_sigs = np.fromfile(env_raw, dtype=d["dtype"]).reshape(-1, d["nbvoies"])
print "size env sigs : ", np.shape(self.env_sigs)[0]
self.EEG_Xpos = np.round(self.nb_win_EEG_values / 2)
self.Feat_Xpos = 0
def read_header(self, header_filename):
d = {}
for line in open(header_filename):
k, v = line.replace("\n", "").replace("\r", "").split(":")
d[k] = v
if "frequence" in d:
d["frequence"] = float(d["frequence"])
if "dtype" in d:
d["dtype"] = np.dtype(d["dtype"])
if "nbvoies" in d:
d["nbvoies"] = int(d["nbvoies"])
channelnames = []
for i in range(d["nbvoies"]):
channelnames.append(d["nom" + str(i + 1)])
d["channelnames"] = channelnames
return d
