import common_functions as cf

import scipy.stats as stats

import matplotlib.pyplot as plt

modulation = 1

cell_type = 'dspn'

mod_type = 'ACh'

if not modulation:

# load data

data = cf.load_data('Data/{}_HFI[0]+0_validation.json'.format(cell_type))

clus_info = data['meta']['clustered']

clus_labels = clus_info['label']

# normality checking =====

# gets stats

data['stats'] = {}

for clus_lab in clus_labels:

data['stats'][clus_lab] = {'dur':{}, 'amp':{}}

# duration data

data['stats'][clus_lab]['dur'] = cf.norm_dist(data['all'][clus_lab]['dur'])

# amplitude data

data['stats'][clus_lab]['amp'] = cf.norm_dist(data['all'][clus_lab]['amp'])

# plot histograms

fig, axs = plt.subplots(2,2)

fig.suptitle(cell_type)

for i, clus_lab in enumerate(clus_labels):

# duration data

axs[i,0].hist(data['all'][clus_lab]['dur'])

axs[i,0].set_title('dur, ' + clus_lab + ', ' + str(data['stats'][clus_lab]['dur']))

# amplitude data

axs[i,1].hist(data['all'][clus_lab]['amp'])

axs[i,1].set_title('amp, ' + clus_lab + ', ' + str(data['stats'][clus_lab]['amp']))

plt.tight_layout()

# tests on data =====

# for dspn, dur and amp data was normally distributed; for ispn, a bit iffy but still within reason for normal distribution

data['stats']['diff'] = {'dur':{}, 'amp':{}}

# duration

data['stats']['diff']['dur']['result'] = stats.anderson_ksamp([data['all'][clus_labels[1]]['dur'],

data['all'][clus_labels[0]]['dur']])[-1]

data['stats']['diff']['dur']['label'] = 'Anderson-Darling; proximal:distal duration'

# ttest_rel() doesn't have one-sided implemented, so to see if duration of one greater than other,

# t-statistic must be > 0 and p-value/2 < .05

# amplitude

data['stats']['diff']['amp']['result'] = stats.anderson_ksamp([data['all'][clus_labels[1]]['amp'],

data['all'][clus_labels[0]]['amp']])[-1]

data['stats']['diff']['amp']['label'] = 'Anderson-Darling; proximal:distal amplitude'

else:

# load data

data = cf.load_data('Data/{}_HFI[0]+0_{}-modulation.json'.format(cell_type, mod_type))

ctrl_data = cf.load_data('Data/{}_HFI[0]+0_validation.json'.format(cell_type))

clus_info = data['meta']['clustered']

clus_labels = clus_info['label']

mod_info = data['meta'][mod_type + ' info']

mod_labels = mod_info['label']

cell_type = data['meta']['cell type']

# normality checking =====

# gets stats

data['stats'] = {}

for clus_lab in clus_labels:

data['stats'][clus_lab] = {}

data['stats'][clus_lab][clus_lab] = {'dur':{}, 'amp':{}}

# duration data

data['stats'][clus_lab][clus_lab]['dur'] = cf.norm_dist(data['all'][clus_lab][clus_lab]['dur'])

# amplitude data

data['stats'][clus_lab][clus_lab]['amp'] = cf.norm_dist(data['all'][clus_lab][clus_lab]['amp'])

for mod_lab in mod_labels:

data['stats'][clus_lab][mod_lab] = {'dur':{}, 'amp':{}}

# duration data

data['stats'][clus_lab][mod_lab]['dur'] = cf.norm_dist(data['all'][clus_lab][mod_lab]['dur'])

# amplitude data

data['stats'][clus_lab][mod_lab]['amp'] = cf.norm_dist(data['all'][clus_lab][mod_lab]['amp'])

# plot histograms

for i, clus_lab in enumerate(clus_labels):

fig, axs = plt.subplots(2,len(mod_labels)+1)

fig.suptitle(cell_type + ', ' + clus_lab)

# duration data

axs[0,0].hist(data['all'][clus_lab][clus_lab]['dur'])

axs[0,0].set_title('dur, on-site, ' + str(data['stats'][clus_lab][clus_lab]['dur']))

# amplitude data

axs[1,0].hist(data['all'][clus_lab][clus_lab]['amp'])

axs[1,0].set_title('amp, on-site, ' + str(data['stats'][clus_lab][clus_lab]['amp']))

for j, mod_lab in enumerate(mod_labels):

# duration data

axs[0,j+1].hist(data['all'][clus_lab][mod_lab]['dur'])

axs[0,j+1].set_title('dur, ' + mod_lab + ', ' + str(data['stats'][clus_lab][mod_lab]['dur']))

# amplitude data

axs[1,j+1].hist(data['all'][clus_lab][mod_lab]['amp'])

axs[1,j+1].set_title('amp, ' + mod_lab + ', ' + str(data['stats'][clus_lab][mod_lab]['amp']))

plt.tight_layout()

# tests on data =====

# for dspn and ispn, dur and amp data was pretty normally distributed

data['stats']['diff'] = {}

# tests whether duration and amplitude values for each modulation target are signficantly different to control values

for i, clus_lab in enumerate(clus_labels):

data['stats']['diff'][clus_lab] = {}

data['stats']['diff'][clus_lab][clus_lab] = {'dur':{}, 'amp':{}}

# duration data

data['stats']['diff'][clus_lab][clus_lab]['dur']['result'] = stats.anderson_ksamp([ctrl_data['all'][clus_lab]['dur'],

data['all'][clus_lab][clus_lab]['dur']])[-1]

data['stats']['diff'][clus_lab][clus_lab]['dur']['label'] = 'Anderson-Darling; control:modulated duration'

# amplitude data

data['stats']['diff'][clus_lab][clus_lab]['amp']['result'] = stats.anderson_ksamp([ctrl_data['all'][clus_lab]['amp'],

data['all'][clus_lab][clus_lab]['amp']])[-1]

data['stats']['diff'][clus_lab][clus_lab]['amp']['label'] = 'Anderson-Darling; control:modulated amplitude'

for j, mod_lab in enumerate(mod_labels):

data['stats']['diff'][clus_lab][mod_lab] = {'dur':{}, 'amp':{}}

# duration data

data['stats']['diff'][clus_lab][mod_lab]['dur']['result'] = stats.anderson_ksamp([ctrl_data['all'][clus_lab]['dur'],

data['all'][clus_lab][mod_lab]['dur']])[-1]

data['stats']['diff'][clus_lab][mod_lab]['dur']['label'] = 'Anderson-Darling; control:modulated duration'

# amplitude data

data['stats']['diff'][clus_lab][mod_lab]['amp']['result'] = stats.anderson_ksamp([ctrl_data['all'][clus_lab]['amp'],

data['all'][clus_lab][mod_lab]['amp']])[-1]

data['stats']['diff'][clus_lab][mod_lab]['amp']['label'] = 'Anderson-Darling; control:modulated amplitude'

# tests whether duration and amplitude values for each modulation target are signficantly different to each other for

# proximal and distal clustered stimulation

mod_targets = ['on-site']

mod_targets.extend(mod_labels)

variables = ['dur', 'amp']

for i, tar in enumerate(mod_targets):

data['stats']['diff'][tar] = {'dur':{}, 'amp':{}}

if tar == 'on-site':

for var in variables:

data['stats']['diff'][tar][var]['result'] = stats.anderson_ksamp([data['all'][clus_labels[0]][clus_labels[0]][var],

data['all'][clus_labels[1]][clus_labels[1]][var]])[-1]

data['stats']['diff'][tar][var]['label'] = 'Anderson-Darling; proximal:distal duration'

else:

for var in variables:

data['stats']['diff'][tar][var]['result'] = stats.anderson_ksamp([data['all'][clus_labels[0]][tar][var],

data['all'][clus_labels[1]][tar][var]])[-1]

data['stats']['diff'][tar][var]['label'] = 'Anderson-Darling; proximal:distal duration'