Welcome to pyseries documentation!¶

import sys
sys.path.insert(0, '/Users/user/Desktop/repo_for_pyseries/pyseries')

import pyseries.LoadingData as loading
import pyseries.Preprocessing as prep
import pyseries.Analysis as analysis
import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np
from scipy import stats

# Notebook settings (i.e. this thing settings)
%matplotlib inline
#%matplotlib notebook
#Change to %matplotlib notebook to be able to zoom, pan, etc the figures,
#inline is only used so the notebook can be exported to .rst format and hosted on readthedocs
%load_ext autoreload
%autoreload 2

def calc_corr(path):
    #Read single subject data - eeg, info about eeg and info from experimental app (unity)
    recording = loading.Read_edf.Combine_EDF_XML(path, 3, 70)
    #Display markers over the whole signal
    prep.Epochs.mark_events(recording,['EEG O1'], subject_name = path)

    #Define epochs for analysis
    epochs_info= {"Please Count": [0, 500*10], "Only Look": [0, 500 *10]}
    #Create them by slicing the signal
    epochs = prep.Epochs.Make_Epochs_for_Channels(recording, ['EEG O1','EEG O2','EEG P3', 'EEG P4'],epochs_info)
    #Re-reference, because only by subtracting P from O-electrodes ssvep response becomes visible
    new_ref = {}
    new_ref['Only Look'] = epochs ['EEG O2']['Only Look'] - epochs['EEG P4']['Only Look']
    new_ref['Please Count'] = epochs ['EEG O2']['Please Count'] - epochs ['EEG P4']['Please Count']
    new_epochs  = {"O-P":new_ref}

    #Get the accuracy in counting condition
    responses = recording['events'][recording['events']["code"] == "responded"]
    accuracy = responses['response'] / responses['expected']

    #Get the power spectra in two conditions
    power_density= analysis.Explore.PlotPowerSpectrum(new_epochs['O-P'], 498, mode = 'period', name = path, freq_min = 0, freq_max = 20)

    ssvep = analysis.Normalize.Z_score( power_density['Please Count'][1][:,49] )
    accuracy = analysis.Normalize.Z_score( accuracy )

    return ssvep, accuracy, power_density

def plot_slow_ssvep():
    #Slow is 5Hz flicker
    paths = ['/Users/user/Desktop/nagrania_eeg/ssvep/Blazej_13_06_16/',
            '/Users/user/Desktop/nagrania_eeg/ssvep/Ania_14_06_16/',
            '/Users/user/Desktop/nagrania_eeg/ssvep/Karen_14_06_16/',
            '/Users/user/Desktop/nagrania_eeg/ssvep/Agnieszka_03_06/',
            '/Users/user/Desktop/nagrania_eeg/ssvep/Kuba_14_06_16/',
            '/Users/user/Desktop/nagrania_eeg/ssvep/Rysiek_03_06/'
            ]

    all_ssvep = []
    all_acc = []

    saving = {}
    for p in paths:
        ssvep, acc, pxx = calc_corr(p)
        all_ssvep.extend(ssvep)
        all_acc.extend(acc)

        saving[p] = ssvep

    sns.jointplot(x = np.array(all_ssvep),y =  np.array(all_acc), kind="reg")
    return saving

def plot_fast_ssvep():
    #fast is 20 Hz flicker
    paths = ['/Users/user/Desktop/nagrania_eeg/ssvep_20hz/Agnieszka_03_06/',
             '/Users/user/Desktop/nagrania_eeg/ssvep_20hz/Rysiek_03_06/']

    for path in paths:

        #Read single subject data - eeg, info about eeg and info from experimental app (unity)
        recording = loading.Read_edf.Combine_EDF_XML(path, 0, 70)
        #Define epochs for analysis
        epochs_info= {"Only Look": [0, 500 *10]}
        #Create them by slicing the signal
        epochs = prep.Epochs.Make_Epochs_for_Channels(recording, ['EEG O1','EEG O2','EEG P3','EEG P4'],epochs_info)
        #Re-reference, because oonly then ssvep response becomes visible
        new_ref = {}
        new_ref['Only Look'] = epochs ['EEG O2']['Only Look'] - epochs['EEG P4']['Only Look']
        new_epochs  = {"O-P":new_ref}

        #Get the power spectra in two conditions
        power_density= analysis.Explore.PlotPowerSpectrum(new_epochs['O-P'], 498, mode = 'period', name = path, freq_min = 0, freq_max = 30)

plot_slow_ssvep()

plot_fast_ssvep()

import LoadingData.ReadData as rd
import Preprocessing.PrepareData as pd

in_path = '/Users/ryszardcetnarski/Desktop/Nencki/Badanie_NFB/Dane/fft_treningi/'
filter_list = ['P3_trening_','P4_trening_', 'F3_trening_', 'F4_trening_']
#load averaged fft from trainings
signals_dict = rd.LoadAvg(in_path, filter_list, freq_lim = [15,40])
#label the outliers
labels, signals_colum_major = pd.FilterIndividualAndGroup(signals_dict, plot_on = True, log = True)
#Plot'em
pd.MarkOutliers(labels[labels['mask'] ==1], rd.LoadAvg(in_path, filter_list, freq_lim = [0,50]))

import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np
import pandas as pd
import Analysis.Anova as anova


%matplotlib inline
%load_ext autoreload
%autoreload 2

groups = []
for mean in [1, 1.5, 1.5, 1]:
    groups.append(np.random.normal(mean, 0.5, 100))

groups = pd.DataFrame(groups).T
ax = sns.boxplot(groups)

anova.one_way(groups)

#Conduct anova on one population which was exposed to different levels of two treatments

factorByfactor = np.array([[np.random.normal(1, 0.5, 100), np.random.normal(1.5, 0.5, 100), np.random.normal(1, 0.5, 100)],
                          [np.random.normal(1, 0.5, 100), np.random.normal(1.5, 0.5, 100), np.random.normal(1, 0.5, 100)]]
                         ).swapaxes(0,2).swapaxes(1,2)

anova.two_way(factorByfactor, 'seed', 'fertilizer')