def combine_files(files, *dates, output='df'):
dn = {}
filtered_files = [file for file in files for date in dates if date in file]
for i, f in enumerate(filtered_files):
d = h5py.File(f)
if output == 'df':
dn['df_{}'.format(i)] = create_df(d)
else:
dn['df_{}'.format(i)] = d
return dn
def plot_opto_masking(d, ignoreNoRespAfter=None):
matplotlib.rcParams['pdf.fonttype'] = 42
# create dataframe of session
df = create_df(d)
mouse = df.mouse
framerate = df.framerate
date = get_dates(df.date)
if ignoreNoRespAfter is not None:
end = ignore_after(d, ignoreNoRespAfter)[0]
df = df[:end]
df['respDir'] = d['trialResponseDir'][:len(df)]
# separate out the target only, mask only, catch only, and masking trials
# (we are using the averages of these, not left/right)
# remove the ignore trials
# NO OPTO TRIALS
maskOnly = df[(df['trialType'] == 'maskOnly')
& (df['ignoreTrial'] == False)]
targetOnly = df[(df['trialType'] == 'targetOnly')
& (df['ignoreTrial'] == False)]
masking = df[(df['trialType'] == 'mask') & (df['ignoreTrial'] == False)]
catch = df[(df['trialType'] == 'catch') & (df['ignoreTrial'] == False)]
hitsNoOpto = [[] for i in range(4)] # list for each df above
respsNoOpto = [[] for i in range(4)]
totalsNoOpto = [[] for i in range(4)]
for i, frame in enumerate([maskOnly, targetOnly, masking, catch]):
hitsNoOpto[i].append(np.sum((frame['resp'] == 1)))
respsNoOpto[i].append(np.sum((np.isfinite(frame['respDir']))))
totalsNoOpto[i].append(len(frame))
totalTrialsNoOpto = np.array(
[len(frame) for frame in [maskOnly, targetOnly, masking, catch]])
hitsNoOpto = np.squeeze(np.array(hitsNoOpto))
respsNoOpto = np.squeeze(np.array(respsNoOpto))
totalsNoOpto = np.squeeze(np.array(totalsNoOpto))
# OPTO TRIALS
maskOnlyOpto = df[(df['trialType'] == 'maskOnlyOpto')
& (df['ignoreTrial'] == False)]
targetOnlyOpto = df[(df['trialType'] == 'targetOnlyOpto')
& (df['ignoreTrial'] == False)]
maskingOpto = df[(df['trialType'] == 'maskOpto')
& (df['ignoreTrial'] == False)]
catchOpto = df[(df['trialType'] == 'catchOpto')
& (df['ignoreTrial'] == False)]
# separate out by onset and count correct by onset
optoOnset = d['optoOnset'][:]
hits = [[] for i in range(4)] # list for each df above
resps = [[] for i in range(4)]
totals = [[] for i in range(4)]
for i, frame in enumerate(
[maskOnlyOpto, targetOnlyOpto, maskingOpto, catchOpto]):
for on in optoOnset:
hits[i].append(
np.sum((frame['optoOnset'] == on) & (frame['resp'] == 1)))
resps[i].append(
np.sum((frame['optoOnset'] == on)
& (np.isfinite(frame['respDir']))))
totals[i].append(np.sum(frame['optoOnset'] == on))
totalTrialsOpto = np.array([
len(frame)
for frame in [maskOnlyOpto, targetOnlyOpto, maskingOpto, catchOpto]
])
hits = np.squeeze(np.array(hits))
resps = np.squeeze(np.array(resps))
totals = np.squeeze(np.array(totals))
# plot the percent correct against the opto onset on the xaxis
xNoOpto = max(optoOnset) + 1
yText = [1.2, 1.15, 1.1, 1.05]
colors = ['k', 'c', 'k', 'm']
alphas = [.3, .5, .8, 1]
for num, denom, title in zip(
[resps, hits], [totals, resps],
['Response Rate', 'Fraction Correct Given Response']):
fig, ax = plt.subplots()
for i, (al, c, text, lbl, noOptoResp, noOptoCorr) in enumerate(
zip(alphas, colors, yText,
['Mask only', 'Target only', 'Masked trials', 'Catch'],
list(respsNoOpto / totalsNoOpto),
list(hitsNoOpto / respsNoOpto))
): # this needs to be resps over totals for no opto
#for fraction correct there should only be target and masked plotted
# or the plot needs to be different, like the ones sam made
if title == 'Response Rate':
ax.plot(optoOnset,
num[i] / denom[i],
'o-',
color=c,
lw=3,
alpha=al,
label=lbl)
ax.plot(xNoOpto, noOptoResp, 'o', color=c, alpha=al)
for x, n in zip(optoOnset, np.transpose(denom[i])):
fig.text(x,
text,
str(n),
transform=ax.transData,
color=c,
alpha=al,
fontsize=10,
ha='center',
va='bottom')
else:
if lbl == 'Mask only' or lbl == 'Catch':
pass
else:
ax.plot(optoOnset,
num[i] / denom[i],
'o-',
color=c,
lw=3,
alpha=al,
label=lbl)
ax.plot(xNoOpto, noOptoCorr, 'o', color=c, alpha=al)
for x, n in zip(optoOnset, np.transpose(denom[i])):
fig.text(x,
text,
str(n),
transform=ax.transData,
color=c,
alpha=al,
fontsize=10,
ha='center',
va='bottom')
# for r in respsNoOpto[1:3]:
# fig.text(xNoOpto, text, str(r), transform=ax.transData, color=c, alpha=al, fontsize=10, ha='center', va='bottom')
#
if title == 'Response Rate':
for t, y, cl, alp in zip(totalsNoOpto, yText, colors, alphas):
fig.text(xNoOpto,
y,
str(t),
transform=ax.transData,
color=cl,
alpha=alp,
fontsize=10,
ha='center',
va='bottom')
elif title == 'Fraction Correct Given Response':
for r, y, cl, alp in zip(respsNoOpto[1:3], yText[1:3], colors[1:3],
alphas[1:3]):
fig.text(xNoOpto,
y,
str(r),
transform=ax.transData,
color=cl,
alpha=alp,
fontsize=10,
ha='center',
va='bottom')
## format the plots
formatFigure(fig,
ax,
xLabel='Opto onset from target onset (ms)',
yLabel=title)
fig.suptitle((mouse + ', ' + date + ' ' +
'Masking with optogenetic silencing of V1'),
fontsize=13)
xticks = list(optoOnset)
xticklabels = [
int(np.round((tick / framerate) * 1000)) - 17 for tick in optoOnset
] # remove 17 ms of latency
x, lbl = ([xNoOpto], ['no\nopto'])
xticks = np.concatenate((xticks, x))
xticklabels = xticklabels + lbl
ax.set_xticks(xticks)
ax.set_xticklabels(xticklabels)
ax.set_xlim([min(optoOnset) - 1, max(xticks) + 1])
ax.set_ylim([0, 1.05])
ax.spines['right'].set_visible(False)
ax.spines['top'].set_visible(False)
ax.tick_params(direction='out', top=False, right=False)
plt.subplots_adjust(top=0.815,
bottom=0.135,
left=0.1,
right=0.92,
hspace=0.2,
wspace=0.2)
plt.legend(loc='best', fontsize='small', numpoints=1)
@author: chelsea.strawder
"""
import fileIO, h5py
import numpy as np
import pandas as pd
import matplotlib
import matplotlib.pyplot as plt
from nogoData import nogo_turn
from dataAnalysis import import_data, create_df
matplotlib.rcParams['pdf.fonttype'] = 42
#def plot_by_param(param): #turn into callable function - IN dataAnalysis?
df = create_df(import_data())
param = 'targetLength'
nonzeroRxns = df[(df['trialLength'] != df['trialLength'].max())
& (df['ignoreTrial'] != True) & (df['resp'] != 0)]
corrNonzero = nonzeroRxns[(nonzeroRxns['resp'] == 1)
& (nonzeroRxns['rewDir'] != 0)]
missNonzero = nonzeroRxns[(nonzeroRxns['resp'] == -1)
& (nonzeroRxns['rewDir'] != 0)]
miss = missNonzero.pivot_table(values='trialLength',
index='{}'.format(param),
columns='rewDir',
margins=True,
dropna=True)
hit = corrNonzero.pivot_table(values='trialLength',
# -*- coding: utf-8 -*-
"""
Created on Tue Jun 30 16:24:15 2020
@author: chelsea.strawder
"""
from dataAnalysis import import_data, create_df
import matplotlib.pyplot as plt
import numpy as np
d = import_data()
df = create_df(d)
trialStart = d['trialStartFrame'][:]
trialEnd = d['trialEndFrame'][:]
wheelPos = d['deltaWheelPos'][:]
bigWheel = [
i for i, (start, end) in enumerate(zip(trialStart, trialEnd))
if abs(np.max(wheelPos[start:end])) > .2
]
for i in bigWheel[:10]:
plt.figure()
plt.plot(wheelPos[trialStart[i]:trialEnd[i]], label='wheel')
plt.plot(df.loc[i, 'trialFrameIntervals'], label='fi')
plt.title(df.mouse + ' ' + str(i))
for j, x in enumerate(df.loc[i, 'trialFrameIntervals']):
if x > .016:
plt.vlines(j, -.5, .5, ls='--')
# -*- coding: utf-8 -*-
"""
Created on Thu Mar 26 15:09:56 2020
@author: chelsea.strawder
"""
from dataAnalysis import import_data, create_df
import scipy.stats as stats
dn = {}
for i, f in enumerate(files[-3:]): #change index for desired files
d = h5py.File(f)
dn['df_{}'.format(i)] = create_df(d) #creates keys in dict dn with format df_#
dictget = lambda x, *k: [x[i] for i in k]
df1, df2, df3 = dictget(dn, 'df_0', 'df_1', 'df_2') # assigns each dataframe to a new variable - essentially unpacks dict dn
dfall = pd.concat([df1,df2,df3], ignore_index=True)
def extract_vars(data):
return {key: data[str(key)][()] for key in data.keys()}
def create_vars(dn):
for key,val in dn.items(): #work in progress, if even possible
globals()
exec (key + '= val')
# IN PLOTTING BY PARAM