#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
| `@purpose`: Hub for creating data visualizations using pandas, seaborn, and bokeh.
| `@date`: Created on Sat May 1 15:12:38 2019
| `@author`: Semeon Risom
| `@email`: semeon.risom@gmail.com
| `@url`: https://semeon.io/d/imhr
"""
# available functions
__all__ = ['Plot']
# core
from pdb import set_trace as breakpoint
from distutils import dir_util
import importlib
import datetime
import shutil
import sys
import os
# data
import pandas as pd
import numpy as np
# seaborn
from pathlib import Path
import matplotlib
matplotlib.use('Agg')
import seaborn as sns
import matplotlib.pyplot as plt
# local libraries
from .. import settings
# constants
console = settings.console
debug = settings.debug
[docs]class Plot():
"""Hub for creating data visualizations using pandas, seaborn, and bokeh."""
@classmethod
def __init__(self, isLibrary=False):
"""Hub for creating data visualizations using pandas, seaborn, and bokeh.
Parameters
----------
isLibrary : :obj:`bool`
Check if required libraries are available.
"""
#check libraries
if isLibrary:
settings.library(__required__)
@classmethod
def __font__(cls):
"""Add Helvetica to matplotlib."""
import imhr
from pathlib import Path
from matplotlib import matplotlib_fname, rcParams
import matplotlib.font_manager as font_manager
directory = matplotlib_fname().replace("/matplotlibrc", "")
root = Path(imhr.__file__).parent
source_ = "%s/dist/resources/Helvetica.ttf"%(root)
destination_ = f'{directory}/fonts/ttf'
#add to matplotlib font folder
shutil.copy(source_, destination_)
#add to computer font folder
##if running osx
if sys.platform == "darwin":
shutil.copy(source_, '/Library/Fonts/')
##if running win32
if sys.platform == "win32":
shutil.copy(source_, 'c:\\windows\\fonts')
#rebuild fonts
prop = font_manager.FontProperties(fname=source_)
prop.set_weight = 'light'
rcParams['font.family'] = prop.get_name()
rcParams['font.weight'] = 'light'
font_manager._rebuild()
[docs] @classmethod
def bokeh_trial(cls, config, df, stim_bounds, roi_bounds, flt):
"""Create single subject trial bokeh plots.
Parameters
----------
df : :class:`pandas.DataFrame`
Pandas dataframe of participant sample data.
stim_bounds : :class:`dict`
Stimulus bounds on screen.
roi_bounds : :class:`dict`
ROI bounds on screen.
flt : :class:`str`
Filter type.
"""
#----bokeh
from bokeh.core.properties import value
from bokeh.plotting import reset_output, figure
from bokeh.models import HoverTool, Range1d, BoxAnnotation, ColumnDataSource, CDSView, BooleanFilter
from bokeh.layouts import gridplot
from bokeh.embed import components
#timestamp
_t0 = datetime.datetime.now()
_f = debug(message='t', source="timestamp")
#metadata
emotion = [df['left_mood'][0],df['right_mood'][0]]
#filter xy coordinates
flt_x = '%s_x'%(flt)
flt_y = '%s_y'%(flt)
#get display size
xy = [int(i) for i in (df['monitorSize.px'].values[0]).split('x')]
monitorSize = xy
#subject data
t = df['timestamp'].values
#get fixations
df['fixation_onset'] = df.loc[df.groupby('fix_num')['%s_fix_all'%(flt)].head(1).index, '%s_fix_all'%(flt)]
df['fixation_onset'] = df.apply(lambda x: True if (x['fixation_onset'] == True) else False, axis=1)
#source for syncing data
s = ColumnDataSource(df)
##tools = "box_select,save,pan,box_zoom,reset"
tools="box_select,save,reset"
#figure 1
#sizing_mode: ``"fixed"``, ``"stretch_both"``, ``"scale_width"``, ``"scale_height"``, ``"scale_both"``
p1 = figure(
output_backend="webgl",
tools=tools,
plot_width=1200,
plot_height=400,
x_range=Range1d(-100, (t[-1]+100), bounds="auto"),
y_range=Range1d(0,(monitorSize[0]+100), bounds="auto"),
x_axis_label = 'time (msec)', y_axis_label = 'x/y (px)',
sizing_mode="stretch_both",
)
#hover
hover = HoverTool()
hover.tooltips = [
("name", "$name"),
("sample", "$index"),
("data (x,y)", "($x{(0)},$y{(0)})")
]
p1.tools.append(hover)
p1.hover.line_policy = "nearest"
#graph
###-----------------------------------------------------------------x=time, y=(x or y)
#original x
p1.scatter(x='timestamp', y='x', color='#2965C0', alpha=0, line_width=2,
line_dash='dashed', name='raw_x', source=s)
p1.line(x='timestamp', y='x', color='#2965C0', alpha=0.5, line_width=2,
legend=value('raw_x'), line_dash='dashed', name='raw_x', source=s)
#filtered x
p1.scatter(x='timestamp', y=flt_x, color='#265EB3', alpha=0, line_width=2,
name=flt_x, source=s)
p1.line(x='timestamp', y=flt_x, color='#265EB3', alpha=1, line_width=2,
legend=value(flt_x), name=flt_x, source=s)
#peak removed
#p1.line(x=t, y=sg_x, color='#101f23', alpha=1, line_width=2,legend='x_peaksg', name='x_peaksg')
#original y
p1.scatter(x='timestamp', y='y', color='#C01F1F', alpha=0, line_width=2,
line_dash='dashed', name='raw_y', source=s)
p1.line(x='timestamp', y='y', color='#C01F1F', alpha=0.5, line_width=2,
legend=value('raw_y'), line_dash='dashed', name='raw_y', source=s)
#filtered y
p1.scatter(x='timestamp', y=flt_y, color='#A61B1B', alpha=0, line_width=2,
name=flt_y, source=s)
p1.line(x='timestamp', y=flt_y, color='#A61B1B', alpha=1, line_width=2,
legend=value(flt_y), name=flt_y, source=s)
#task-events (stimulus onset, dotloc onset)
##stimulus onset
time = df.loc[df['marker']=='Stimulus Onset']['timestamp'].iloc[0]
box = BoxAnnotation(left=time, right=time + 10, fill_alpha=1, fill_color='#434e54', level='glyph')
p1.add_layout(box)
##dotloc onset
time = df.loc[df['marker']=='Dotloc Onset']['timestamp'].iloc[0]
box = BoxAnnotation(left=time, right=time + 10, fill_alpha=1, fill_color='#434e54', level='glyph')
p1.add_layout(box)
#fixations
for index, row in df[~df['sg_fix_all'].isin([False])].groupby('fix_num'):
start = row.iloc[0]['timestamp']
end = row.iloc[-1]['timestamp']
box = BoxAnnotation(left=start, right=end,fill_alpha=0.2, fill_color='#9e9e9e', level='glyph')
p1.add_layout(box)
del index, row
#legend
p1.toolbar.logo = None
p1.legend.orientation = "horizontal"
p1.legend.location = "top_left"
p1.legend.click_policy = "hide"
###-----------------------------------------------------------------window display
#sizing_mode: ``"fixed"``, ``"stretch_both"``, ``"scale_width"``, ``"scale_height"``, ``"scale_both"``
#display what participant sees
tools = "box_select,save,pan,box_zoom,reset"
p2 = figure(
output_backend="webgl",
tools=tools,
plot_width=1200,
plot_height=600,
x_axis_label = 'x (px)', y_axis_label = 'y (px)',
x_range=Range1d(0, monitorSize[0], bounds="auto"),
y_range=Range1d(0, monitorSize[1], bounds="auto"),
sizing_mode="stretch_both"
)
#hover
hover = HoverTool()
hover.tooltips = [
("name", "$name"),
("sample", "$index"),
("data (x,y)", "($x{(0)},$y{(0)})")
]
p2.tools.append(hover)
p2.hover.line_policy = "nearest"
###---------line plots
#------gaze xy
p2.cross(x=flt_x, y=flt_y, size=10, color="firebrick", alpha=1, fill_alpha=0.6,
fill_color="firebrick", source=s, name='gaze', legend='gaze')
#------fixation onset within and outside roi
#get booleanfilter
l_fix_all = [True if x !=False else False for x in s.data['fixation_onset']]
fix_all = BooleanFilter(l_fix_all)
view = CDSView(source=s, filters=[fix_all])
##draw square ##2222b2
p2.square(x=flt_x, y=flt_y, size=10, color="#2222b2", alpha=1, fill_alpha=0.6,
fill_color="white", source=s, view=view, name='fixation', legend=value("fixation"))
#------fixation within and outside roi
#get booleanfilter
l_fix_false = [True if x !=False else False for x in s.data['sg_fix_all']]
fix_false = BooleanFilter(l_fix_false)
view = CDSView(source=s, filters=[fix_false])
##draw square ##2222b2
p2.square(x=flt_x, y=flt_y, size=10, color="black", alpha=1, fill_alpha=0.6,
fill_color="white", source=s, view=view, name='false-fixation', legend=value("false-fixation"))
#------only fixation within roi
#get booleanfilter
l_fix_roi = [True if (y==True and z==True) else False for y,z in zip(s.data['fixation_onset'], s.data['dwell'])]
fix_roi = BooleanFilter(l_fix_roi)
view = CDSView(source=s, filters=[fix_roi])
##draw triangle ##198D40
p2.triangle(x=flt_x, y=flt_y, size=10, color="#198D40", alpha=1, fill_alpha=0.6,
fill_color="white", source=s, view=view, name='roi', legend=value("roi"))
#------dwell-time gaze within roi
l_dwell_roi = [True if x !=False else False for x in s.data['dwell']]
dwell_roi = BooleanFilter(l_dwell_roi)
view = CDSView(source=s, filters=[dwell_roi])
##draw circle ##ff9800
p2.circle(x=flt_x, y=flt_y, size=10, color="#ff9800", alpha=1, fill_alpha=0.6,
fill_color="white", source=s, view=view, name='dwell', legend=value("dwell"))
#------stim bounds
stimBounds = stim_bounds
#------roi bounds
roiBounds = roi_bounds
#emotion
left, right = emotion
if left == 'Sad':
left_color = '#ff0000'
right_color = '#9e9e9e'
else:
right_color = '#ff0000'
left_color = '#9e9e9e'
#roi
#left
p2.add_layout(BoxAnnotation(left=roiBounds[0]['bx1'], right=roiBounds[0]['bx2'],
top=roiBounds[0]['by1'], bottom=roiBounds[0]['by2'],
fill_alpha=0.05, fill_color='blue', level='glyph'))
#right
p2.add_layout(BoxAnnotation(left=roiBounds[1]['bx1'], right=roiBounds[1]['bx2'],
top=roiBounds[1]['by1'], bottom=roiBounds[1]['by2'],
fill_alpha=0.05, fill_color='blue', level='glyph'))
#--------stim bounds
#left
p2.add_layout(BoxAnnotation(left=stimBounds[0]['bx1'], right=stimBounds[0]['bx2'],
top=stimBounds[0]['by1'], bottom=stimBounds[0]['by2'],
fill_alpha=0.3, fill_color=left_color, level='glyph'))
#right
p2.add_layout(BoxAnnotation(left=stimBounds[1]['bx1'], right=stimBounds[1]['bx2'],
top=stimBounds[1]['by1'], bottom=stimBounds[1]['by2'],
fill_alpha=0.3, fill_color=right_color, level='glyph'))
#hover
hover = HoverTool()
hover.tooltips = [
("time", "@timestamp"),
("sample", "$index"),
("data (x,y)", "($%s_x{(0)},$%s_y{(0)})"%(flt,flt))
]
#p2.tools.append(hover)
#p2.hover.line_policy = "nearest"
#legend
p2.toolbar.logo = None
p2.legend.orientation = "horizontal"
p2.legend.location = "top_left"
p2.legend.click_policy = "hide"
#reverse axis
p2.x_range = Range1d(0, monitorSize[0])
p2.y_range = Range1d(monitorSize[1], 0)
###-----------------------------------------------------------------slider
#sizing_mode: ``"fixed"``, ``"stretch_both"``, ``"scale_width"``, ``"scale_height"``, ``"scale_both"``
start = df.iloc[0]['timestamp']
end = df.iloc[-1]['timestamp']
# callback = CustomJS(args=dict(source=s), code="""
# data = source.data;
# t = cb_obj;
# f = cb_obj.value;
# x = data['x'];
# y = data['y'];
# source.change.emit();
# console.log('test');
# """)
# p3 = Slider(
# title="time",
# start=start,
# end=end,
# value=1, step=1,
# callback=callback,
# bar_color='red',
# width=1200,
# show_value=True,
# sizing_mode="stretch_both"
# )
###-----------------------------------------------------------------combining plots
#sizing_mode: ``"fixed"``, ``"stretch_both"``, ``"scale_width"``, ``"scale_height"``, ``"scale_both"``
# grid = gridplot(children=[[p1], [p2], [p3]],
grid = gridplot(children=[[p1], [p2]],
merge_tools=True,
toolbar_location='right',
sizing_mode="fixed")
#get html
script, div = components(grid)
##convert seperate plots and div to single string
plots = (''.join(map(str, [div, '\n', script])))
#clear cache
reset_output()
#--------finished
console('%s finished in %s msec'%(_f,((datetime.datetime.now()-_t0).total_seconds()*1000)), 'blue')
return plots
[docs] @classmethod
def bokeh_calibration(cls, config, df, cxy, event, monitorSize=[1920,1080]):
"""Create calibration matrix, using pandas and bokeh.
Parameters
----------
config : :obj:`dict`
Configuration data.
df : :class:`pandas.DataFrame`
Pandas dataframe of raw data.
cxy : :class:`pandas.DataFrame`
Pandas dataframe of calibration points.
event : :obj:`string`
calibration, or validation.
monitorSize : :obj:`list`
Monitor size, in pixels.
"""
from bokeh.plotting import reset_output, figure
from bokeh.models import Range1d, ColumnDataSource
from bokeh.embed import components
#timestamp
_t0 = datetime.datetime.now()
_f = debug(message='t', source="timestamp")
#subject data
event_t = df['event_trial'].values
#source for syncing data
s = ColumnDataSource(df)
##tools = "box_select,save,pan,box_zoom,reset"
#figure 1
#sizing_mode: ``"fixed"``, ``"stretch_both"``, ``"scale_width"``, ``"scale_height"``, ``"scale_both"``
##tools = "box_select,save,pan,box_zoom,reset"
tools = "box_select,reset"
p1 = figure(
output_backend="webgl",
tools=tools,
plot_width=1200,
plot_height=150,
x_range=Range1d(0 - 0.1, (event_t[-1] + 0.1), bounds="auto"),
y_range=Range1d(0,(monitorSize[0]+100), bounds="auto"),
x_axis_label = 'calibration point', y_axis_label = 'x/y (px)',
sizing_mode="scale_width",
)
#graph
###-----------------------------------------------------------------x=time, y=(x or y)
#original x
p1.scatter(x='event_trial', y='cx', color='#2965C0', alpha=0.5, line_width=2,
line_dash='dashed', name='cx', source=s)
# p1.line(x='event_trial', y='cx', color='#2965C0', alpha=0.5, line_width=2,
# line_dash='dashed', name='cx', source=s,legend=value('cx'))
#original y
p1.scatter(x='event_trial', y='cy', color='#C01F1F', alpha=0.5, line_width=2,
line_dash='dashed', name='cy', source=s)
# p1.line(x='event_trial', y='cy', color='#C01F1F', alpha=0.5, line_width=2,
# line_dash='dashed', name='cy', source=s,legend=value('cy'))
###-----------------------------------------------------------------window display
#sizing_mode: ``"fixed"``, ``"stretch_both"``, ``"scale_width"``, ``"scale_height"``, ``"scale_both"``
##tools = "box_select,save,pan,box_zoom,reset"
tools="save,reset"
p2 = figure(
output_backend="webgl",
tools=tools,
plot_width=1200,
plot_height=600,
x_axis_label = 'x (px)', y_axis_label = 'y (px)',
x_range=Range1d(0, monitorSize[0], bounds="auto"),
y_range=Range1d(0, monitorSize[1], bounds="auto"),
sizing_mode="scale_width"
)
###---------line plots
#----gaze xy
# cross = p2.cross(x='gx', y='gy', size=10, color="firebrick", alpha=1, fill_alpha=0.6,
p2.cross(x='gx', y='gy', size=10, color="firebrick", alpha=1, fill_alpha=0.6,
fill_color="firebrick", source=s, name='gaze')
# fill_color="firebrick", source=s, name='gaze', legend=value("gaze"))
#----circle xy
p2.circle(x=cxy['cx'], y=cxy['cy'], size=12, color="#198D40", alpha=1, fill_alpha=0.6,
fill_color="white", name='circle')
#fill_color="white", source=s, name='circle', legend=value("circle"))
# draw mouse click for calibration
if event=='calibration':
#----mouse xy
p2.square(x='mx', y='my', size=10, color="#2222b2", alpha=1, fill_alpha=0.6,
fill_color="white", source=s, name='mouse')
#fill_color="white", source=s, name='mouse', legend=value("mouse"))
#----legend
p1.toolbar.logo = None
p2.toolbar.logo = None
#----reverse axis
p2.x_range = Range1d(0, monitorSize[0])
p2.y_range = Range1d(monitorSize[1], 0)
#----grid
#p1
p1.grid.grid_line_alpha = 0.5
p1.grid.grid_line_color = "white"
#----font
# p2
## title
p2.axis.axis_label_text_font_style = "normal"
## labels
p2.axis.axis_label_text_font_size= "11.5pt"
p2.axis.axis_label_text_font = "Helvetica"
p2.axis.axis_label_text_color = "#1c1d1e"
p2.axis.major_label_text_font_size= "11.5pt"
p2.axis.major_label_text_font = "Helvetica"
p2.axis.major_label_text_color = "#1c1d1e"
# p1
## title
p1.axis.axis_label_text_font_style = "normal"
## labels
p1.axis.axis_label_text_font_size = "9.5pt"
p1.axis.axis_label_text_font = "Helvetica"
p1.axis.axis_label_text_color = "#1c1d1e"
p1.axis.major_label_text_font_size = "9.5pt"
p1.axis.major_label_text_font = "Helvetica"
p1.axis.major_label_text_color = "#1c1d1e"
p1.axis.major_tick_line_color = None
p1.axis.minor_tick_line_color = None
#----padding
p2.min_border_top = 16
p1.min_border_bottom = 0
#----axis line
# p2
p2.axis.axis_line_width = 0
p2.axis.axis_line_color = None
p2.outline_line_color = "#1c1d1e"
# p1
p1.background_fill_color = "#dfe1e4"
p1.axis.axis_line_width = 0
p1.axis.axis_line_color = None
p1.outline_line_color = "#1c1d1e"
#----get html
#get componants
plot = {'top': p2, 'bottom': p1}
script, div = components(plot)
#combine
div = ''.join(list(div.values()))
plot = div + script
#----clear cache
reset_output()
#--------finished
console('%s finished in %s msec'%(_f,((datetime.datetime.now()-_t0).total_seconds()*1000)), 'blue')
return plot
[docs] @classmethod
def onset_diff_plot(cls, config, df, meta, drop, y, clip=None):
"""Plot onset differences using pandas and seaborn.
Parameters
----------
config : :obj:`dict`
Configuration data.
df : :class:`pandas.DataFrame`
Pandas dataframe of raw data.
meta : :class:`pandas.DataFrame`
Metadata for each chart.
drop : :class:`pandas.DataFrame`
Participants to be dropped.
y : :obj:`str`
Variable of interest.
clip : :obj:`int`
Clip value for single subject plot.
Returns
-------
odp
Bokeh or seaborn plot.
"""
#----initiate fonts
__font__()
#setting clip
clip = 200 if clip == None else clip
#timestamp
_t0 = datetime.datetime.now()
_f = debug(message='t', source="timestamp")
importlib.reload(plt); importlib.reload(sns)
console('running onset_diff_plot(%s)'%(y), 'blue')
html_plots = []
#-----------------participant trend line
individual = meta['individual']
sns.set(style=config['style']['seaborn'], font_scale=1.25, font="Helvetica")
#sns.despine(offset=10, trim=True)
fig, ax = plt.subplots(figsize=(10,5))
#plot
df["%s (msec)"%(y)] = df.groupby('participant')[y].rolling(window=5).mean().reset_index(0,drop=True)
##if diff_dotloc and diff_stim < 500 msec
sns.color_palette("RdBu",6)
ax_ = sns.lineplot(x="TrialNum_", y="%s (msec)"%(y), ax=ax, legend=False, lw=1, estimator=None, palette="ch:2.5,.25",
hue="participant", units="participant", data=df.loc[df['onset_greater'] == False])
##if diff_dotloc and diff_stim > 500 msec
ax_ = sns.lineplot(x="TrialNum_", y="%s (msec)"%(y), ax=ax, legend=False, lw=1, estimator=None, alpha=0.2,
hue="participant", units="participant", data=df.loc[df['onset_greater'] == True])
#settings
ax_.set(xlim=(0, 1), ylim=(0, clip), ylabel='| %s (msec) |'%(y), xlabel='trial')
#append
html_plots.append({"title":individual['title'],"file":"%s.png"%(individual['file']),"footnote":individual['footnote'],"anchor":"itl"})
#tight layout
fig.tight_layout()
#save
fig.savefig(individual['path'], dpi=300)
#clear figure from memory
plt.close(fig)
#-----------------diff plot
group = meta['group']
#for each plot
importlib.reload(plt); importlib.reload(sns)
fig, ax = plt.subplots(figsize=(10,10), ncols=2, nrows=2, sharey=False)
ax1, ax2, ax3, ax4 = ax.flatten()
sns.set(style=config['style']['seaborn'], font_scale=1.1, font="Helvetica")
#sns.despine(offset=10, trim=True)
#binned
##iaps
type_='iaps'
df_1 = df.loc[df['trialType'] == type_].reset_index(level=0, drop=True)
sns.regplot(x="TrialNum_",y=y,color="#5874a6",data=df_1,ax=ax1,
fit_reg=True,x_bins=group['binned']['bins'],scatter_kws={"s": 40},line_kws={'color': 'red'})
ax1.set(ylim=(0, 200), xlabel='', ylabel='| %s (msec) |'%(y), title='trial_type=%s'%(type_))
##pofa
type_='pofa'
df_1 = df.loc[df['trialType'] == type_].reset_index(level=0, drop=True)
sns.regplot(x="TrialNum_",y=y,color="#5d9f6c",data=df_1,ax=ax2,
fit_reg=True,x_bins=group['binned']['bins'],scatter_kws={"s": 40},line_kws={'color': 'red'})
ax2.set(ylim=(0, 200), xlabel='', ylabel='', title='trial_type=%s'%(type_))
#unbinned
##iaps
type_='iaps'
df_1 = df.loc[df['trialType'] == type_].reset_index(level=0, drop=True)
sns.regplot(x="TrialNum_",y=y,color="#5874a6",data=df_1,ax=ax3,
fit_reg=True,x_bins=group['unbinned']['bins'],scatter_kws={"s": 5},line_kws={'color': 'red'})
ax3.set(ylim=(0, 1000), xlabel='trial', ylabel='| %s (msec) |'%(y), title='trial_type=%s'%(type_))
##pofa
type_='pofa'
df_1 = df.loc[df['trialType'] == type_].reset_index(level=0, drop=True)
sns.regplot(x="TrialNum_",y=y,color="#5d9f6c",data=df_1,ax=ax4,
fit_reg=True,x_bins=group['unbinned']['bins'],scatter_kws={"s": 5},line_kws={'color': 'red'})
ax4.set(ylim=(0, 1000), xlabel='trial', ylabel='', title='trial_type=%s'%(type_))
#append
html_plots.append({"title":group['title'],"file":"%s.png"%(group['file']),"footnote":group['footnote'],"anchor":"gtl"})
#tight layout
fig.tight_layout()
#save
fig.savefig(group['path'], dpi=300)
#clear figure from memory
plt.close(fig)
#--------finished
#timestamp
console('%s finished in %s msec'%(_f,((datetime.datetime.now()-_t0).total_seconds()*1000)), 'blue')
return html_plots
[docs] @classmethod
def density_plot(cls, config, df, title):
"""
Create density plot (draws kernel density estimate), using seaborn and pandas.
Parameters
----------
config : :obj:`dict`
Configuration data.
df : :class:`pandas.DataFrame`
Pandas dataframe of raw data.
title : :obj:`str`
Chart title.
Returns
-------
cm
Bokeh or seaborn plot.
"""
#----initiate fonts
__font__()
#timestamp
_t0 = datetime.datetime.now()
_f = debug(message='t', source="timestamp")
#seaborn
#https://seaborn.pydata.org/generated/seaborn.distplot.html
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import seaborn as sns
importlib.reload(plt); importlib.reload(sns)
console('running density_plot()', 'blue')
#remove cesd_group, trialType_ column
df = df.drop(['trialType_','cesd_group'], axis=1)
#create metadata to be used for html file
html_plots = []
#for each column
for item in (df.columns.tolist()):
#----plot type
sns.set(style=config['style']['seaborn'], font_scale=1.1, font="Helvetica")
sns.despine(offset=10, trim=True)
#----plot
fig, ax = plt.subplots(figsize=(10,10))
density = sns.distplot(df[item], ax=ax, norm_hist=True,
#bins - number of bins for histogram
hist=True, bins=20,
#bw (bandwidth) - width of the kernels
kde=True, kde_kws={'clip':None,'bw':'scott','kernel':'gau'}).set(xlim=(None, None))
ax.set(ylabel='Density')
#----design
#tight layout
fig.tight_layout()
#----save
#check if path exists
file = "density_%s.png"%(item)
path = config['path']['output'] + "/analysis/html/img/density/"
if not os.path.exists(path):
os.makedirs(path)
#save
fig.savefig(path + file, dpi=300)
#append
html_plots.append({"title":'%s for %s.'%(title, item),"file":'density/'+file,"footnote":""})
#clear figure from memory
plt.close(fig)
#--------finished
console('%s finished in %s msec'%(_f,((datetime.datetime.now()-_t0).total_seconds()*1000)), 'blue')
return density, html_plots
[docs] @classmethod
def corr_matrix(cls, config, df, path, title, method, footnote=None):
"""
Create correlation matrix using bokeh and pandas.
Parameters
----------
config : :obj:`dict`
Configuration data.
df : :class:`pandas.DataFrame`
Pandas dataframe of raw data.
path : :obj:`str`
The directory path to save the bokeh or plot.
method : :obj:`str`
Spearman or Pearsons correlation coefficient.
title : :obj:`str`
Chart title.
footnote : :obj:`str`
Chart footnote.
Returns
-------
cm
Bokeh plot.
"""
#timestamp
_t0 = datetime.datetime.now()
_f = debug(message='t', source="timestamp")
console('running bokeh corr_matrix()', 'blue')
# math
import bisect
from math import pi
from numpy import arange
from itertools import chain
from collections import OrderedDict
# bokeh
from bokeh.models import ColorBar, LinearColorMapper, TapTool, HoverTool, Range1d, ColumnDataSource
from bokeh.models.callbacks import CustomJS
from bokeh.plotting import reset_output, figure
from bokeh.embed import components
#create color palette
matplotlib.use('Agg')
import matplotlib.colors
from matplotlib import cm as mpl_cmap
#p-values
if method == "pearson":
from scipy.stats import pearsonr as cf
elif method == "spearman":
from scipy.stats import spearmanr as cf
#calculate p-values correlation coefficent
def get_pvalue(df):
df = df.dropna()._get_numeric_data()
dfcols = pd.DataFrame(columns=df.columns)
pvalues = dfcols.transpose().join(dfcols, how='outer')
for row in df.columns:
for column in df.columns:
pvalues[row][column] = round(cf(df[row], df[column])[1], 4)
return pvalues
#Gets bounds for quads with n features
def get_bounds(n):
bottom = list(chain.from_iterable(
[[ii]*nlabels for ii in range(nlabels)]))
top = list(chain.from_iterable(
[[ii+1]*nlabels for ii in range(nlabels)]))
left = list(chain.from_iterable(
[list(range(nlabels)) for ii in range(nlabels)]))
right = list(chain.from_iterable(
[list(range(1, nlabels+1)) for ii in range(nlabels)]))
corr_items = list(chain.from_iterable(
[[ii+1]*nlabels for ii in range(nlabels)]))
return top, bottom, left, right, corr_items
#Aligns color values from palette with the correlation coefficient and p-values
def get_colors_corr(corr_array, p_array, colors):
c_corr = arange(-1, 1, 1/(len(colors)/2))
corr_color = []
p_color = []
corr_list = []
p_list = []
factor_list = []
itr = 0
for corr, pvalue in zip(corr_array,p_array):
ind = bisect.bisect_left(c_corr, corr)
#colors
##corr_color
if (itr)%(nlabels+1)==0:
corr_color.append('#022f62')
else:
corr_color.append(colors[ind-1])
##pcolor
if pvalue<=0.05:
p_color.append("#F44336")
else:
p_color.append(colors[ind-1])
#append corr and pvalues
corr_list.append(corr)
p_list.append(pvalue)
factor_list.append(['[%s, %s]'%(l_labels[itr][0],l_labels[itr][1])])
itr = itr + 1
return corr_color, corr_list, p_color, p_list, factor_list
#create seaborn plots for each possible correlation combinations
def get_corr_plot(df, x, y):
importlib.reload(plt); importlib.reload(sns)
#----parameters
#figure
fig, ax = plt.subplots(1, 1, figsize=(10,10))
#bounds
plt.subplots_adjust(top=0.95, bottom=0.085, left=0.125, right=0.975, hspace=0.2, wspace=0.2)
#----title
title = "x=%s, y=%s"%(x,y)
#----set style
#pal = sns.color_palette("Set1", n_colors=2, desat=.5)
sns.set(style=config['style']['seaborn'], font="Helvetica", font_scale=1.1)
sns.despine(offset=10, trim=True)
#----plot
#if grouping #g = sns.lmplot(x, y, hue=groupby, data=df, palette=pal).set_title(title)
sns.regplot(x, y, data=df, scatter=True, fit_reg=True, ci=95).set_title(title)
#----add rho, p-value to plot
#get values
corr_r = cf(df[x],df[y])
corr_t = "r = %.2f, p < %.2f"%(corr_r[0],corr_r[1])
#place text 90% of max-x and 90% of max-y
fig.text(0.85, 0.05, corr_t, size=16, horizontalalignment='center',
verticalalignment='center', transform=ax.transAxes)
#----save
#check if path exists
file = "%s.png"%(title)
path = config['path']['output'] + "/analysis/html/img/corr/"
if not os.path.exists(path):
os.makedirs(path)
#save
fig.savefig(path + file, dpi=300)
#reset
reset_output()
#get color map
RdYlBu = mpl_cmap.get_cmap('RdBu', 32)
colors_np = np.vstack(RdYlBu(np.linspace(0, 1, 32)))
cmap = matplotlib.colors.ListedColormap(colors_np, name='RedBlue')
colors = [matplotlib.colors.rgb2hex(cmap(x/32)[:3]) for x in range(32)]
# calculate correlation coefficients
#coeff
corr_coeff = df.corr(method=method)
#p-value
p_value = get_pvalue(df)
# get list and number of variables
labels = df.columns
nlabels = len(corr_coeff)
#create list of labels
##x
l_labels_x = labels.tolist() * nlabels
##y
l_labels_y = labels.tolist() * nlabels
###split into chunks
l_labels_y_split = np.hsplit(np.array(np.array_split(l_labels_y, nlabels)), nlabels)
l_labels_y = np.concatenate(l_labels_y_split).ravel().tolist()
##combine
l_labels = list(zip(l_labels_x, l_labels_y))
#create coeff label coordinates
num = np.arange(0, nlabels, 0.5).tolist()
num = [x for x in num if x not in np.arange(0, nlabels).tolist()]
##x
num_x = [(x - 0.025) for x in num]
num_x = [x for x in num_x if x not in np.arange(0, nlabels).tolist()] * nlabels
##y
num_y = [(x + .125) for x in num]
num_y = [y for y in num_y if y not in np.arange(0, nlabels).tolist()] * nlabels
###split into chunks
num_y_split = np.hsplit(np.array(np.array_split(num_y, nlabels)), nlabels)
num_y = np.concatenate(num_y_split).ravel().tolist()
##combine
#num_labels = list(zip(num_x, num_y))
#plot
tools="box_select,save,reset"
cm = figure(tools=tools, plot_width=1200, plot_height=1200, x_range=(0, nlabels), y_range=(0, nlabels),
output_backend="webgl")
#hover
hover = HoverTool()
hover.tooltips = [
("factor", "@factor"),
("rho", "@coeff_num"),
("p", "@p_num"),
]
#cm.tools.append(hover)
cm.hover.line_policy = "nearest"
#legend
#cm.toolbar.logo = None
#grid
cm.xgrid.grid_line_color = None
cm.ygrid.grid_line_color = None
cm.xaxis.major_label_orientation = pi/4
cm.yaxis.major_label_orientation = pi/4
#turn off scientific notation
cm.left[0].formatter.use_scientific = False
# prepare squares for plot
top, bottom, left, right, corr_items = get_bounds(nlabels)
color_list,corr,p_color,pvalues,factor = get_colors_corr(corr_coeff.values.flatten(),p_value.values.flatten(),colors)
#prepare corr_coeff label for plot
coeff_color = ["#444444"] * (nlabels * nlabels)
coeff_color = ['#f9f9f9' if cr >= np.float64(0.74) else cof for cof, cr in zip(coeff_color, corr)]
coeff_num = ["%.2f"%(x) for x in corr]
#prepare pvalue for plot
p_color = ["#444444"] * (nlabels * nlabels)
p_color = ['#F44336' if p <= np.float64(0.05) else cof for pcol, cof, p in zip(p_color, coeff_color, pvalues)]
p_num = ["(%.2f)"%(y) for y in pvalues]
#data source
source = ColumnDataSource(data=dict(left=left, right=right, top=top, bottom=bottom,
x=num_x, y=num_y, coeff_num=coeff_num, coeff_color=coeff_color,
p_num=p_num, factor=factor, square_color=color_list, p_color=p_color))
#create squares
cm.quad(top='top', bottom='bottom', left='left', right='right',
line_color='white', color='square_color', source=source)
#create text
#correlation coefficient
cm.text(x='x', y='y', source=source, text_color="coeff_color", text='coeff_num',
text_font_style='normal', text_font_size = '1.5em', text_align='center',
y_offset=-5, x_offset=1, text_line_height=1)
#p-value
cm.text(x='x', y='y', source=source, text_color="p_color", text='p_num',
text_font_style='normal', text_font_size = '1.0em', text_align='center',
y_offset=20, x_offset=1, text_line_height=1)
#callback
link = CustomJS(args = dict(source = source), code="""
console.log("second")
obj = cb_obj
s = source
data = source.data
attr = source.attributes
d = attr.data
//selected item
id = source.selected['1d'].indices
item = d.factor[id][0].replace(/[[\]]/g,'').match(/(".*?")|(\S+)/g)
console.log('id: ' + id + ', item: ' + item)
//url
path = window.location.href
current = path.split("/").slice(0,-1).join("/")
url = current + '/reg.html?x='+ item[0] + 'y=' + item[1]
window.open(url)
//window.open(url, 'newwindow', config='height=720, width=1280')
""")
cm.add_tools(TapTool(callback=link))
#reverse y-axis
cm.y_range = Range1d(nlabels, 0)
#p.y_range = Range1d(monitorSize[1], 0)
# Set ticks with labels
ticks = [tick+0.5 for tick in list(range(nlabels))]
tick_dict = OrderedDict([[tick, labels[ii]] for ii, tick in enumerate(ticks)])
# Create the correct number of ticks for each axis
cm.xaxis.ticker = ticks
cm.yaxis.ticker = ticks
# Override the labels
cm.xaxis.major_label_overrides = tick_dict
cm.yaxis.major_label_overrides = tick_dict
#color bar
mapper = LinearColorMapper(
palette=colors,
low=-1, high=1
)
color_bar = ColorBar(color_mapper=mapper, location=(0, 0))
cm.add_layout(color_bar, 'right')
#create link for each comparison
all_columns = [[x,y] for x in list(labels) for y in list(labels)]
for idx, clm in enumerate(all_columns):
get_corr_plot(df=df, x=clm[0], y=clm[1])
#get html
script, div = components(cm)
##convert seperate plots and div to single string
plots = (''.join(map(str, [div, '\n', script])))
#create html
html(cls, config, path=path, plots=plots, source='bokeh', title=title, footnote=footnote)
#reset
reset_output()
#--------finished
#timestamp
console('%s finished in %s msec'%(_f,((datetime.datetime.now()-_t0).total_seconds()*1000)), 'blue')
return cm
[docs] @classmethod
def single_subject(cls, config, df, path):
"""Create single subject scatterplot using seaborn and pandas.
Parameters
----------
df : :class:`pandas.DataFrame`
Pandas dataframe of raw data.
path : :obj:`str`
The directory path to save the bokeh or seaborn plot.
Returns
-------
cm
Bokeh or seaborn plot.
"""
#----initiate fonts
__font__()
#timestamp
_t0 = datetime.datetime.now()
_f = debug(message='t', source="timestamp")
console('running single_subject()', 'blue')
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import seaborn as sns
importlib.reload(plt); importlib.reload(sns)
#styling plot
sns.set(style="darkgrid", font_scale=1, font="Helvetica")
sns.set_context("paper")
#events
events = self.config['events']
#get first pofa and iaps
first = df.drop_duplicates('trialType', keep='first')['TrialNum'].tolist()
#get display size
xy = [int(i) for i in (df['monitorSize.px'].values[0]).split('x')]
#drop x - outside coordinates
df = df.loc[~(df['sg_x']<0)] #drop all gaze_x < 0
df = df.loc[~(df['sg_x']>xy[0])] #drop all gaze_x > width
#drop timestamp - outside coordinates
df = df.loc[~(df['timestamp']>8000)]
#drop all non fixations
df = df.loc[~(df['sg_fix_all']==False)]
#loop to draw event markers
def xline(x, **kwargs):
data = kwargs.pop("data")
trialType = data['trialType'].values[0]
TrialNum = data['TrialNum'].values[0]
#check if first iaps or first pofa
if TrialNum in first:
#fixation onset
time = events['fixation']
plt.axvline(x=0, c='red', linewidth=2)
##annotation
plt.text(s="Fixation Onset", x=(0 + 100), y=(xy[0] - 25), fontsize=12)
#fixation offset
time = events['fixation']
plt.axvline(x=time, c='red', linewidth=2)
##annotation
plt.text(s="Stimulus Offset", x=(time + 100), y=(xy[0] - 25), fontsize=12)
#stimulus offset
time = events['fixation'] + events['stimulus'][trialType]
#line
plt.axvline(x=time, c='red', linewidth=2)
#annotation
plt.text(s="Dotloc Offset", x=(time + 100), y=(xy[0] - 25), fontsize=12)
#plot
pal = sns.color_palette("RdBu", n_colors=192)
sp = sns.FacetGrid(col="trialType", data=df, height=5, palette=pal, hue="TrialNum")
sp.map(plt.scatter, "timestamp", "sg_x", alpha=0.3, s=10)
sp.map_dataframe(xline, x="timestamp")
#post
ax = sp.axes.flatten()
ax[0].set_ylabel("Gaze Coordinates (x)", fontdict={'fontsize':12})
ax[0].set_title("trialType = IAPS", fontdict={'fontsize':12})
ax[0].tick_params(labelsize=12)
ax[1].set_title("trialType = POFA", fontdict={'fontsize':12})
ax[1].tick_params(labelsize=12)
#size
sp.fig.set_size_inches(20,10)
#save
sp.savefig(path, dpi=300)
#clear figure from memory
plt.close()
#--------finished
#timestamp
console('%s finished in %s msec'%(_f,((datetime.datetime.now()-_t0).total_seconds()*1000)), 'blue')
return plt
[docs] @classmethod
def boxplot(cls, config, df, path=None, x=None, y=None, title=None, plots=None, cat='analysis'):
"""
Creates boxplot using seaborn and pandas.
Parameters
----------
config : :obj:`dict`
Configuration data.
df : :class:`pandas.DataFrame`
Pandas dataframe of raw data.
path : :obj:`str`
Path to save data.
drift : :obj:`str`
X-axis.
drift : :obj:`str`
Y-axis.
title : :obj:`str`
Plot title.
plots : :obj:`dict`
Dictionary of plots metadata.
cat : :obj:`str`
Type of plot.
"""
#----initiate fonts
__font__()
#timestamp
_t0 = datetime.datetime.now()
_f = debug(message='t', source="timestamp")
console('running boxplot(%s)'%(y), 'blue')
# start
matplotlib.use('Agg')
import matplotlib.pyplot as plt
importlib.reload(plt); importlib.reload(sns)
sns.set(style=config['style']['seaborn'], font_scale=1.5, font="Helvetica")
sns.despine(offset=10, trim=True)
short_ = config['metadata']['short']
#if looking at dwell time
if cat=='bias':
#plot
fig, (ax1, ax2) = plt.subplots(figsize=(10,10),ncols=2,sharey=True)
#design
##for each unique value in x-axis variable, assign color
cat_x = df[x].unique().tolist()
color=['b','r','g','y','o']
pal = dict(zip(cat_x, color))
#create plot
for ax, trialType in zip([ax1,ax2],['iaps','pofa']):
df_ = df.loc[df['trialType'].isin([trialType])]
sns.boxplot(x=x, y='.fitted', data=df_, ax=ax, palette=pal)
sns.swarmplot(x=x, y='.fitted', data=df_, color=".25", ax=ax)
ax.set_title(trialType)
ax.set_xlabel(short_[x])
if trialType == 'pofa':
ax.set_ylabel('')
#if looking at dwell time
if cat=='dwell':
#plot
fig, (ax1, ax2) = plt.subplots(figsize=(10,10),ncols=2,sharey=True)
#design
##for each unique value in x-axis variable, assign color
cat_x = df[x].unique().tolist()
color=['b','r','g','y','o']
pal = dict(zip(cat_x, color))
#create plot
for ax, trialType in zip([ax1,ax2],['iaps','pofa']):
df_ = df.loc[df['trialType'].isin([trialType])]
sns.boxplot(x=x, y='.fitted', data=df_, ax=ax, palette=pal)
sns.swarmplot(x=x, y='.fitted', data=df_, color=".25", ax=ax)
ax.set_title(trialType)
ax.set_xlabel(short_['aoi'])
if trialType == 'pofa':
ax.set_ylabel('')
#if looking at variables like gaze_bias, dp_bias
elif cat=='analysis':
#plot
fig, (ax1, ax2, ax3, ax4) = plt.subplots(figsize=(19.2, 10.8),ncols=4,sharey=False)
#design
##for each unique value in x-axis variable, assign color
cat_x = df[x].unique().tolist()
color=['b','r','g','y','o']
pal = dict(zip(cat_x, color))
#create plot
for ax, itm in zip([ax1,ax2,ax3,ax4],y):
#dp_bias
sns.boxplot(x=x, y=itm, data=df, ax=ax, palette=pal)
sns.swarmplot(x=x, y=itm, data=df, color=".25", ax=ax)
ax.set(xlabel=x)
#if looking at variables like race, gender, os
elif cat=='demographics':
#layout
sns.set(style=config['style']['seaborn'], palette="deep", font_scale=1.1, font="Helvetica")
sns.despine(offset=10, trim=True)
fig, (ax1,ax2,ax3,ax4) = plt.subplots(figsize=(20, 10),ncols=4,sharey=True)
palette = sns.color_palette(["#e74c3c", "#3f51b5", "#2ecc71","#9b59b6","#ffb400"])
#counter
counter = 0
for ax, x_, in zip([ax1,ax2,ax3,ax4],x):
##exclude smaller race categories for display purposees
if x_=='gender':
df_ = df.loc[~df['gender'].isin(["other"])].drop_duplicates(subset="participant",keep="first")
##exclude smaller race categories for display purposes
elif x_=='race':
df_ = df.loc[~df['race'].isin(["American Indian or Alaska Native","Two or more races",
"Black or African American","None of the above"])].drop_duplicates(subset="participant",keep="first")
else:
df_ = df.drop_duplicates(subset="participant",keep="first")
#plot
sns.boxplot(x=x_, y=y, palette=palette, ax=ax, data = df_)
#remove label
if counter!=0:
ax.set_ylabel('')
#counter
counter = counter + 1
#layer
sns.despine(offset=10, trim=True)
#tight layout
fig.tight_layout()
##save
fig.savefig(path, dpi=300)
#clear from memory
plt.close(fig)
#--------finished
console('%s finished in %s msec'%(_f,((datetime.datetime.now()-_t0).total_seconds()*1000)), 'blue')
[docs] @classmethod
def cooks_plot(cls, config, y, model, path, df):
"""Create cooks distance plot plot using seaborn, pandas, and statsmodel.
Parameters
----------
config : :obj:`dict`
Configuration data.
df : :class:`pandas.DataFrame`
Pandas dataframe of raw data.
path : :obj:`str`
The directory path save the seaborn plot.
y : :obj:`str`
The predictor variable.
model : :obj:`dict`
statsmodel model.
Returns
-------
res
seaborn plot.
"""
#----initiate fonts
__font__()
#timestamp
_t0 = datetime.datetime.now()
_f = debug(message='t', source="timestamp")
console('running cooks_plot(%s)'%(y), 'blue')
#-------plot
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import seaborn as sns
importlib.reload(plt); importlib.reload(sns)
#plot design
sns.set(style=config['style']['seaborn'], font_scale=1.1, font="Helvetica")
sns.despine(offset=10, trim=True)
#plot
fig, ax = plt.subplots(figsize=(10,10))
# influence
influence = model.get_influence()
# cook's distance, from statsmodels internals
## c is the distance and p is p-value
(c, p) = influence.cooks_distance
#plot
plt.stem(np.arange(len(c)), c, markerfmt=",")
plt.xlabel('Observation Number')
plt.ylabel('Cooks Distance')
# annotations
particpants = list(df['participant'])
leverage_lines = np.flip(np.argsort(c), 0)[:3]
for i in leverage_lines:
plt.annotate(particpants[i], xy=(i, c[i]))
#tight layout
fig.tight_layout()
#save
fig.savefig(path, dpi=300)
#clear figure from memory
plt.close(fig)
#--------finished
console('%s finished in %s msec'%(_f,((datetime.datetime.now()-_t0).total_seconds()*1000)), 'blue')
return plt
[docs] @classmethod
def residual_plot(cls, config, y, residuals, path):
"""Create probability plot using seaborn, pandas, and rpy2.
Parameters
----------
config : :obj:`dict`
Configuration data.
y : :obj:`str`
Predictor variable.
residuals : :class:`pandas.DataFrame`
Pandas dataframe of residuals vs fitted, qq data, and raw data.
path : :obj:`str`
The directory path save the seaborn plot.
Returns
-------
lmp
seaborn plot.
"""
#----initiate fonts
__font__()
#timestamp
_t0 = datetime.datetime.now()
_f = debug(message='t', source="timestamp")
console('running residual_plot(%s)'%(y), 'blue')
#-------plot
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import seaborn as sns
importlib.reload(plt); importlib.reload(sns)
#-------plot design
sns.set(style=config['style']['seaborn'], font_scale=1.1, font="Helvetica")
sns.despine(offset=10, trim=True)
#-------plot
fig, ax = plt.subplots(figsize=(10,10))
#residual vs fitted plot
cmap=sns.color_palette(["#bd282f","#226db2"])
sns.scatterplot(x=".fitted", y=".resid", data=residuals, ax=ax, hue="trialType", palette=cmap, legend=None, s=60, edgecolor=None)
#loess smoothing curve
sns.regplot(x=".fitted", y=".resid", data=residuals, ax=ax, lowess=True, scatter=False, color="blue")
#horizontal line
plt.axhline(0, zorder=1, color='black', linewidth=2.25)
#annotations
##max
max_ = residuals[residuals['.resid']==residuals['.resid'].max()]
max_x = max_['.fitted'].iloc[0]
max_y = max_['.resid'].iloc[0]
max_subject = max_['participant'].iloc[0]
plt.text(max_x, max_y, max_subject, color='black', weight='semibold')
##min
min_ = residuals[residuals['.resid']==residuals['.resid'].min()]
min_x = min_['.fitted'].iloc[0]
min_y = min_['.resid'].iloc[0]
min_subject = min_['participant'].iloc[0]
plt.text(min_x, min_y, min_subject, color='black', weight='semibold')
#-------plot design
#label
plt.xlabel('.fitted')
plt.ylabel('.resid')
#tight layout
fig.tight_layout()
#-------save
fig.savefig(path, dpi=300)
#-------clear figure from memory
plt.close(fig)
#--------finished
console('%s finished in %s msec'%(_f,((datetime.datetime.now()-_t0).total_seconds()*1000)), 'blue')
return plt
[docs] @classmethod
def qq_plot(cls, config, y, residuals, path):
"""Create probability plot using seaborn, pandas, and rpy2.
Parameters
----------
config : :obj:`dict`
Configuration data.
y : :obj:`str`
Predictor variable.
residuals : :class:`pandas.DataFrame`
Pandas dataframe of residuals vs fitted, qq data, and raw data.
path : :obj:`str`
The directory path save the seaborn plot.
Returns
-------
lmp
seaborn plot.
"""
#----initiate fonts
__font__()
#timestamp
_t0 = datetime.datetime.now()
_f = debug(message='t', source="timestamp")
console('running qq_plot(%s)'%(y), 'blue')
#-------plot
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import seaborn as sns
importlib.reload(plt); importlib.reload(sns)
#plot design
sns.set(style=config['style']['seaborn'], font_scale=1.1, font="Helvetica")
sns.despine(offset=10, trim=True)
#plot
fig, ax = plt.subplots(figsize=(10,10))
#qq residuals
cmap=sns.color_palette(["#bd282f","#226db2"])
sns.scatterplot(x="theoretical", y=".resid", data=residuals, ax=ax, hue="trialType", palette=cmap, legend=None, s=60, edgecolor=None)
sns.regplot(x="theoretical", y=".resid", data=residuals, ax=ax, scatter=None, color="black", ci=None)
#label
plt.xlabel('Theoretical Quantiles')
plt.ylabel('Sample Quantiles')
#annotations
##max
max_ = residuals[residuals['.resid']==residuals['.resid'].max()]
max_x = max_['theoretical'].iloc[0]
max_y = max_['.resid'].iloc[0]
max_subject = max_['participant'].iloc[0]
plt.text(max_x, max_y, max_subject, color='black', weight='semibold')
##min
min_ = residuals[residuals['.resid']==residuals['.resid'].min()]
min_x = min_['theoretical'].iloc[0]
min_y = min_['.resid'].iloc[0]
min_subject = min_['participant'].iloc[0]
plt.text(min_x, min_y, min_subject, color='black', weight='semibold')
#tight layout
fig.tight_layout()
#save
fig.savefig(path, dpi=300)
#clear figure from memory
plt.close(fig)
#--------finished
console('%s finished in %s msec'%(_f,((datetime.datetime.now()-_t0).total_seconds()*1000)), 'blue')
return plt
[docs] @classmethod
def logit_plot(cls, config, df, path, param):
"""Create logistic regression plot using seaborn and pandas
Parameters
----------
df : :class:`pandas.DataFrame`
Pandas dataframe of raw data.
path : :obj:`str`
The directory path save the seaborn plot.
param : :obj:`dict`
x, y, groupby parameters.
Returns
-------
lmp
seaborn plot.
"""
#----initiate fonts
__font__()
#----for timestamp
_t0 = datetime.datetime.now()
_f = debug(message='t', source="timestamp")
console('processing.logit_plot()', 'blue')
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import seaborn as sns
importlib.reload(plt); importlib.reload(sns)
#plot design
p=param
sns.set(style=config['style']['seaborn'], font_scale=1.1, font="Helvetica")
sns.despine(offset=10, trim=True)
console('running logit_plot(), (y=%s, x=%s)'%(p['y'],p['x']), 'blue')
#plot
fig, (ax1,ax2) = plt.subplots(figsize=(19.2, 10.8), ncols=2, sharey=True)
#iaps
type_ = 'iaps'
df_ = df.loc[(df['trialType_']==type_)]
lmp = sns.regplot(x=p['x'], y=p['y'], data=df_, logistic=True, y_jitter=.02, ax=ax1, color="#5874a6")
ax1.set(xlabel='trial type', title='trialType=%s'%(type_))
#pofa
type_ = 'pofa'
df_ = df.loc[(df['trialType_']==type_)]
lmp = sns.regplot(x=p['x'], y=p['y'], data=df_, logistic=True, y_jitter=.02, ax=ax2, color="#5d9f6c")
ax2.set(xlabel='trial type', title='trialType=%s'%(type_))
#tight layout
fig.tight_layout()
#save
fig.savefig(path, dpi=300)
#clear figure from memory
plt.close(fig)
#--------finished
console('%s finished in %s msec'%(_f,((datetime.datetime.now()-_t0).total_seconds()*1000)), 'blue')
return lmp
[docs] @classmethod
def html(cls, destination=None, df=None, raw_data=None, name=None, path=None, plots=None, source=None, title=None, intro=None, footnote=None, script="", **kwargs):
"""
Create HTML output.
Parameters
----------
destination : :obj:`str`
Path to save file to.
df : :class:`pandas.DataFrame`
Pandas dataframe of analysis results data.
raw_data : :class:`pandas.DataFrame`
Pandas dataframe of raw data.
name : :obj:`str`
(py::`if source is logit`) The name of csv file created.
path : :obj:`str`
The directory path of the html file.
plots : :obj:`dict`
If generating seaborn images, the list of plots used.
source : :obj:`str`
The type of data being recieved.
trial : :obj:`str`
(If Bokeh) Trial Number.
session : :obj:`str`
(If Bokeh) Session Number.
bokeh_type : :obj:`str`
(If Bokeh) Control directory location. If trial, create trial plots.
title : :obj:`str`
The title of the table or figure.
intro : :obj:`str`
The introduction of the group of figures or tables.
footnote : :obj:`str`
The footnote of the table or figure.
metadata : :obj:`dict`
Additional data to be included.
**kwargs : :obj:`str`, :obj:`int`, or :obj:`None`, optional
Additional properties, relevent for specific content types. Here's a list of available properties:
.. list-table::
:class: kwargs
:widths: 25 50
:header-rows: 1
* - Property
- Description
* - **short**, **long** : :obj:`str`
- Short (aoi) and long form (Area of Interest) label of html page. This is primarily used for constructing metadata tags in html.
* - **display** : :obj:`str`
- (For bokeh) The type of calibration/validation display.
* - **trial** : :obj:`str`
- (For bokeh) The trial number for the eyetracking task.
* - **session** : :obj:`int`
- (For bokeh) The session number for the eyetracking task.
* - **day** : :obj:`str`
- (For bokeh) The day the eyetracking task was run.
Returns
-------
html : :obj:`str`
String of html code.
"""
#----initiate fonts
cls.__font__()
#timestamp
date = datetime.datetime.now().replace(microsecond=0).strftime('%Y-%m-%d %H:%M:%S')
_t0 = datetime.datetime.now()
_f = debug(message='t', source="timestamp")
console('running plot.html()', 'blue')
#---copy csv, imhr, js files to each model
#copy files
for folder in ['css','imhr','js']:
import imhr
p_ = str(Path(imhr.__file__).parent.parent)
s_ = '%s/docs/source/_templates/html/%s/'%(p_, folder)
d_ = '%s/%s/'%(os.path.dirname(path), folder)
#shutil.move(s_, d_)
import distutils
distutils.dir_util.copy_tree(s_, d_)
#----if creating html file from bokeh
if ((source=='bokeh') or (source=='methods')):
#get short and long names
short = kwargs['short'] if 'short' in kwargs else source
long = kwargs['long'] if 'long' in kwargs else source
#if trial plot
if (source=='bokeh'):
display = kwargs['display'] if 'display' in kwargs else None
trial = kwargs['trial'] if 'trial' in kwargs else None
session = kwargs['session'] if 'session' in kwargs else None
day = kwargs['day'] if 'day' in kwargs else None
if ((display=='trial') or (display=='calibration') or (display=='validation')):
parent='../'
else:
parent=''
else:
parent=''
#build header
link = []; css = []; js = []
link.append('<title>IMHR [webgazer] %s</title>'%(long))
link.append('<meta name="title" content="mdl[webgazer]: %s">'%(long))
link.append('<meta name="description" content="Results from our feasibility study for in-browser eyetracking.">')
link.append('<meta name="author" content="Semeon Risom">')
link.append('<meta name="email" content="semeon.risom@gmail.com">')
link.append('<meta name="robots" content="index,follow">')
link.append('<meta name="AdsBot-Google" content="noindex">')
link.append('<!--size-->')
link.append('<meta name="viewport" content="width=device-width, initial-scale=1.0" charset="utf-8"/>')
link.append('<!--no cache-->')
link.append('<meta http-equiv="Cache-Control" content="no-cache, no-store, must-revalidate"/>')
link.append('<meta http-equiv="Pragma" content="no-cache"/>')
link.append('<meta http-equiv="Expires" content="0"/>')
link.append('<!--icons-->')
link.append('<link rel="shortcut icon" href="'+parent+'imhr/favicon/favicon.ico">')
link.append('<link rel="icon" sizes="16x16 32x32 64x64" href="'+parent+'imhr/favicon/favicon.ico">')
link.append('<link rel="icon" type="image/png" sizes="196x196" href="'+parent+'imhr/favicon/favicon-192.png">')
link.append('<link rel="icon" type="image/png" sizes="160x160" href="'+parent+'imhr/favicon/favicon-160.png">')
link.append('<link rel="icon" type="image/png" sizes="96x96" href="'+parent+'imhr/favicon/favicon-96.png">')
link.append('<link rel="icon" type="image/png" sizes="64x64" href="'+parent+'imhr/favicon/favicon-64.png">')
link.append('<link rel="icon" type="image/png" sizes="32x32" href="'+parent+'imhr/favicon/favicon-32.png">')
link.append('<link rel="icon" type="image/png" sizes="16x16" href="'+parent+'imhr/favicon/favicon-16.png">')
link.append('<link rel="apple-touch-icon" href="'+parent+'imhr/favicon/favicon-57.png">')
link.append('<link rel="apple-touch-icon" sizes="114x114" href="'+parent+'imhr/favicon/favicon-114.png">')
link.append('<link rel="apple-touch-icon" sizes="72x72" href="'+parent+'imhr/favicon/favicon-72.png">')
link.append('<link rel="apple-touch-icon" sizes="144x144" href="'+parent+'imhr/favicon/favicon-144.png">')
link.append('<link rel="apple-touch-icon" sizes="60x60" href="'+parent+'imhr/favicon/favicon-60.png">')
link.append('<link rel="apple-touch-icon" sizes="120x120" href="'+parent+'imhr/favicon/favicon-120.png">')
link.append('<link rel="apple-touch-icon" sizes="76x76" href="'+parent+'imhr/favicon/favicon-76.png">')
link.append('<link rel="apple-touch-icon" sizes="152x152" href="'+parent+'imhr/favicon/favicon-152.png">')
link.append('<link rel="apple-touch-icon" sizes="180x180" href="'+parent+'imhr/favicon/favicon-180.png">')
link.append('<meta name="msapplication-TileColor" content="#FFFFFF">')
link.append('<meta name="msapplication-TileImage" content="'+parent+'imhr/favicon/favicon-144.png">')
link.append('<meta name="msapplication-config" content="'+parent+'docs/favicon/browserconfig.xml">')
css.append('<link rel="stylesheet" type="text/css" href="'+parent+'css/user.css">')
js.append('<script type="text/javascript" language="javascript" src="'+parent+'js/jquery/jquery-3.2.1.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="'+parent+'js/user.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="'+parent+'js/lodash.min.js"></script>')
css.append('<!--bootstrap-->')
css.append('<link rel="stylesheet" type="text/css" href="'+parent+'css/bootstrap/bootstrap.css">')
css.append('<!--bokeh-->')
css.append('<link rel="stylesheet" type="text/css" href="'+parent+'css/bokeh/bokeh-1.0.4.min.css">')
css.append('<link rel="stylesheet" type="text/css" href="'+parent+'css/bokeh/bokeh-widgets-1.0.4.min.css">')
css.append('<link rel="stylesheet" type="text/css" href="'+parent+'css/bokeh/bokeh-tables-1.0.4.min.css">')
js.append('<script type="text/javascript" language="javascript" src="'+parent+'js/bokeh/bokeh-1.0.4.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="'+parent+'js/bokeh/bokeh-widgets-1.0.4.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="'+parent+'js/bokeh/bokeh-tables-1.0.4.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="'+parent+'js/bokeh/bokeh-gl-1.0.4.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="'+parent+'js/bokeh/bokeh-api-1.0.4.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="'+parent+'js/popper.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="'+parent+'js/bootstrap/bootstrap.js"></script>')
# finish header
head = ('\n\t'.join(map(str, link))) + ('\n\t'.join(map(str, css))) + ('\n\t'.join(map(str, js)))
#body, footer
if source=='bokeh':
#build body
body = ['<div class="container" style="">',
'<div class="container bokeh" style="">',
'<div class="figure %s" trial-id="%s" session-id="%s" day-id="%s" type-id="%s">\
'%(display, trial, session, day, display),
'<div class="title"><b>Figure 1.</b> <a>%s</a></div>'%(title),
'<div class="date"> Last Updated: ' + '%s'%(date) + '</div>',
plots,
'<div class="footnote">%s</div>'%(footnote),
'</div>',
'</div>',
'</div>']
# finish body
body = ('\n\t'.join(map(str, body)))
#build footer
if ((display=='trial') or (display=='calibration') or (display=='validation')):
foot = ['<script>',
'$(document).ready(function() {getinput();});',
'</script>']
# finish footer
foot = ('\n\t'.join(map(str, foot)))
else:
foot = ''
elif source=='methods':
#build body
body = ['<div class="container" style="">',
'<div class="dashboard-main">',
'<div class="table-container">',
'<div class="title"><a><b>%s</b></a></div>'%(title) ,
plots,
'</div>',
'</div>',
'</div>']
# finish body
body = ('\n\t'.join(map(str, body)))
#build footer
foot = ''
#build html
html = ['<html short="%s" long="%s">'%(short, long),
'<head>',
head,
'</head>',
'<body id="body" class="%s" style="margin: 0px; background-color: #434e54 !important;">'%(display),
body,
'</body>',
'<footer>',
foot,
'</footer>',
'</html>']
# finish html
html = ('\n'.join(map(str, html)))
#all other html
else:
#get short and long names
short = kwargs['short'] if 'short' in kwargs else source
long = kwargs['long'] if 'long' in kwargs else source
var = kwargs['var'] if 'var' in kwargs else source
#build header
link = []; css = []; js = []
link.append('<title>IMHR [webgazer] %s</title>'%(long))
link.append('<meta name="title" content="mdl[webgazer]: %s">'%(long))
link.append('<meta name="description" content="Results from our feasibility study for in-browser eyetracking.">')
link.append('<meta name="author" content="Semeon Risom">')
link.append('<meta name="email" content="semeon.risom@gmail.com">')
link.append('<meta name="robots" content="index,follow">')
link.append('<meta name="AdsBot-Google" content="noindex">')
link.append('<!--size-->')
link.append('<meta name="viewport" content="width=device-width, initial-scale=1.0" charset="utf-8"/>')
link.append('<!--no cache-->')
link.append('<meta http-equiv="Cache-Control" content="no-cache, no-store, must-revalidate"/>')
link.append('<meta http-equiv="Pragma" content="no-cache"/>')
link.append('<meta http-equiv="Expires" content="0"/>')
link.append('<meta name="msapplication-TileColor" content="#FFFFFF">')
link.append('<meta name="msapplication-TileImage" content="imhr/favicon/favicon-144.png">')
link.append('<meta name="msapplication-config" content="docs/favicon/browserconfig.xml">')
link.append('<!--icons-->')
link.append('<link rel="shortcut icon" href="imhr/favicon/favicon.ico">')
link.append('<link rel="icon" sizes="16x16 32x32 64x64" href="imhr/favicon/favicon.ico">')
link.append('<link rel="icon" type="image/png" sizes="196x196" href="imhr/favicon/favicon-192.png">')
link.append('<link rel="icon" type="image/png" sizes="160x160" href="imhr/favicon/favicon-160.png">')
link.append('<link rel="icon" type="image/png" sizes="96x96" href="imhr/favicon/favicon-96.png">')
link.append('<link rel="icon" type="image/png" sizes="64x64" href="imhr/favicon/favicon-64.png">')
link.append('<link rel="icon" type="image/png" sizes="32x32" href="imhr/favicon/favicon-32.png">')
link.append('<link rel="icon" type="image/png" sizes="16x16" href="imhr/favicon/favicon-16.png">')
link.append('<link rel="apple-touch-icon" href="imhr/favicon/favicon-57.png">')
link.append('<link rel="apple-touch-icon" sizes="114x114" href="imhr/favicon/favicon-114.png">')
link.append('<link rel="apple-touch-icon" sizes="72x72" href="imhr/favicon/favicon-72.png">')
link.append('<link rel="apple-touch-icon" sizes="144x144" href="imhr/favicon/favicon-144.png">')
link.append('<link rel="apple-touch-icon" sizes="60x60" href="imhr/favicon/favicon-60.png">')
link.append('<link rel="apple-touch-icon" sizes="120x120" href="imhr/favicon/favicon-120.png">')
link.append('<link rel="apple-touch-icon" sizes="76x76" href="imhr/favicon/favicon-76.png">')
link.append('<link rel="apple-touch-icon" sizes="152x152" href="imhr/favicon/favicon-152.png">')
link.append('<link rel="apple-touch-icon" sizes="180x180" href="imhr/favicon/favicon-180.png">')
js.append('<!--all-->')
js.append('<script type="text/javascript" language="javascript" src="js/jquery/jquery-3.2.1.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="js/lodash.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="js/bootstrap/bootstrap.bundle.js"></script>')
link.append('<!--dataTables-->')
css.append('<link rel="stylesheet" type="text/css" href="css/bootstrap/bootstrap.min.css">')
css.append('<link rel="stylesheet" type="text/css" href="css/datatables/dataTables.bootstrap.min.css">')
css.append('<link rel="stylesheet" type="text/css" href="css/datatables/responsive.bootstrap.min.css">')
css.append('<link rel="stylesheet" type="text/css" href="css/datatables/buttons.bootstrap.min.css">')
js.append('<!--user-->')
link.append('<link rel="stylesheet" type="text/css" href="css/user.css">')
js.append('<!--dataTables-->')
js.append('<script type="text/javascript" language="javascript" src="js/datatables/jquery.dataTables.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="js/datatables/dataTables.rowsGroup.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="js/datatables/dataTables.rowGroup.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="js/datatables/dataTables.bootstrap.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="js/datatables/dataTables.buttons.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="js/datatables/buttons.bootstrap.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="js/datatables/dataTables.responsive.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="js/datatables/responsive.bootstrap.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="js/datatables/buttons.html5.min.js"></script>')
js.append('<script type="text/javascript" language="javascript" src="js/datatables/jszip.min.js"></script>')
js.append('<!--user-->')
js.append('<script type="text/javascript" language="javascript" src="js/user.js"></script>')
js.append('<script>source="%s"</script>'%(source))
# additional libraries
if ((source == "logit") or (source == "onset") or (source == "anova")):
link.append('<!--prism-->')
css.append('<link rel="stylesheet" type="text/css" href="css/prism/prism.css">')
css.append('<link rel="stylesheet" type="text/css" href="css/prism/prism-line-numbers.css">')
css.append('<script type="text/javascript" language="javascript" src="js/prism/prism.js"></script>')
css.append('<script type="text/javascript" language="javascript" src="js/prism/prism-r.js"></script>')
css.append('<script type="text/javascript" language="javascript" src="js/prism/prism-python.js"></script>')
css.append('<script type="text/javascript" language="javascript" src="js/prism/prism-line-numbers.js"></script>')
css.append('<script type="text/javascript" language="javascript" src="js/prism/prism-normalize-whitespace.js"></script>')
# finish header
head = ('\n\t'.join(map(str, link))) + ('\n\t'.join(map(str, css))) + ('\n\t'.join(map(str, js)))
#if displaying seaborn plots
if source == "plots":
html_plots = []
#for each plot
for idx, p in enumerate(plots):
html_p = ['<div class="figure" style="max-width:800px;">',
#title
'<div class="title"><b>Figure %s.</b> %s</div>'%((idx+1), p['title']),
#date
'<div class="date"> Last Updated: ' + '%s'%(date) + '</div>',
#figure
'<a href="img\\' + p['file'] + '">\n<img class="figure" src="img\\' + p['file'] + '">\n</a>',
#footnote
'<div class="footnote">%s</div>'%(p['footnote']),
'</div>']
html_plots.append(('\n\t'.join(map(str, html_p))))
#build body
##add none as class only if there is no intro
if intro!=None:
intro_class = ''
else:
intro_class = None
body = ['<div class="container" style="">',
'<div class="container-figure %s" style="">'%(display),
'<div class="intro %s">%s</div>'%(intro_class, intro),
(''.join(map(str, html_plots))),
'</div>',
'</div>',
'</div>']
body = ('\n\t'.join(map(str, body)))
#build html
html = ['<html>',
'<head>',
head,
'</head>',
'<body id="body">',
body,
'</body>',
'</html>']
html = ('\n' .join(map(str, html)))
#save
with open(path,'w') as html_:
html_.write(html)
#--------finished
console('%s finished in %s msec'%(_f,((datetime.datetime.now()-_t0).total_seconds()*1000)), 'blue')
return html
#if displaying dataframe tables
else:
#build body
##save table to html for datatable
if ((source == "demographic") or (source == "device") or (source == "task") or (source == "summary")):
#allow word wrapping
if(source == "demographic") or (source == "task"):
wrap = "wrap"
else:
wrap = "nowrap"
# save table to csv for downloading
## check if paths exist
_folder = '%s/csv'%(Path(path).parent)
p_ = '%s/%s.csv'%(_folder, name)
if not os.path.exists(_folder):
os.makedirs(_folder)
df.to_csv(p_, index=False)
#get table
table = df.to_html(index=False, index_names=True).replace('<table border="1" class="dataframe">',
'<table id="table" class="table '+source+' table-striped table-bordered \
hover dt-responsive '+wrap+'" cellspacing="0" width="100%">').replace('>','>').replace('<','<')
elif (source == "definitions"):
#wordwrap
wrap = "nowrap"
#prevent trucating strings
width = pd.get_option('display.max_colwidth')
pd.set_option('display.max_colwidth', -1)
##save table to csv for downloading
_folder = '%s/csv'%(Path(path).parent)
p_ = '%s/%s.csv'%(_folder, name)
if not os.path.exists(_folder):
os.makedirs(_folder)
df.to_csv(p_, index=False)
#get table
table = df.to_html(index=False, index_names=True).replace('<table border="1" class="dataframe">',
'<table id="table" class="table '+source+' table-striped table-bordered \
hover dt-responsive '+wrap+'" cellspacing="0" width="100%">')
#reset
pd.set_option('display.max_colwidth', width)
elif ((source == "logit") or (source == "onset") or (source == "anova")):
#prevent trucating strings
width = pd.get_option('display.max_colwidth')
pd.set_option('display.max_colwidth', -1)
##save table to csv for downloading #here saving raw data for future analysis
_folder = '%s/csv'%(Path(path).parent)
p_ = '%s/%s.csv'%(_folder, name)
if not os.path.exists(_folder):
os.makedirs(_folder)
raw_data.to_csv(p_, index=False)
#get table
table = df.to_html(index=True, index_names=True).replace('<table border="1" class="dataframe">',
'<table id="table" data-file="'+ name +'" class="table '+source+' table-striped table-bordered \
hover dt-responsive nowrap" cellspacing="0" width="100%">')
#reset
pd.set_option('display.max_colwidth', width)
elif (source == "full_summary"):
#wordwrap
wrap = "nowrap"
#prevent trucating strings
width = pd.get_option('display.max_colwidth')
pd.set_option('display.max_colwidth', -1)
##save table to csv for downloading
_folder = '%s/csv'%(Path(path).parent)
p_ = '%s/%s.csv'%(_folder, name)
if not os.path.exists(_folder):
os.makedirs(_folder)
df.to_csv(p_, index=False)
#get table
table = df.to_html(index=False, index_names=True).replace('<table border="1" class="dataframe">',
'<table id="table" class="table '+source+' table-striped table-bordered \
hover dt-responsive '+wrap+'" cellspacing="0" width="100%">')
#reset
pd.set_option('display.max_colwidth', width)
else:
##check if paths exist
_folder = '%s/csv'%(Path(path).parent)
p_ = '%s/%s.csv'%(_folder, name)
if not os.path.exists(_folder):
os.makedirs(_folder)
# save table to csv for downloading
df.to_csv(p_ + '%s.csv'%(name), index=False)
#get table
table = df.to_html(index=True, index_names=True).replace('<table border="1" class="dataframe">',
'<table id="table" data-file="'+ name +'" class="table '+source+' table-striped table-bordered \
hover dt-responsive nowrap" cellspacing="0" width="100%">')
#add images to bottom of model
if ((source == "logit") or (source == "onset") or (source == "anova")):
html_plots = []
#for each plot
if plots != None:
fig_num = 0
for idx, p in enumerate(plots):
if (source == "anova"):
#if supplimentary table
if (p['type'] == 'table'):
if 'anchor' in p:
anchor = p['anchor']
_table = p['df']
#prevent trucating strings
width = pd.get_option('display.max_colwidth')
pd.set_option('display.max_colwidth', -1)
#convert df to html table
_table = _table.to_html(index=True, index_names=True).replace('<table border="1" class="dataframe">',
'<table id="table2" data-file="'+ name +'" class="table '+source+' table-striped table-bordered \
hover dt-responsive nowrap"\ cellspacing="0" width="100%">')
#create table
html_p = [
'<div class="figure %s" id="%s">'%(source, anchor),
'<div class="title">%s</div>'%(p['title']),
'<div class="date"> Last Updated: ' + '%s'%(date) + '</div>',
'<div class="dataTables_wrapper form-inline dt-bootstrap">',
'<!--script-->' + '\n',
_table,
'</div>',
'<div class="footnote">%s</div>'%(p['footnote']),
'</div>']
html_p = ('\n\t'.join(map(str, html_p)))
#reset
pd.set_option('display.max_colwidth', width)
#else is plot
else:
fig_num = fig_num + 1
if 'anchor' in p:
anchor = p['anchor']
else:
anchor = ""
html_p = ['<div class="figure %s" id="%s">'%(source, anchor),
#title
'<div class="title"><b>Figure %d.</b> %s</div>'%(fig_num, p['title']),
#date
'<div class="date"> Last Updated: ' + '%s'%(date) + '</div>',
#figure
'<a class="img" href="img\\'+p['file']+'">\n<img class="img" src="img\\'+p['file']+'">\n</a>',
#footnote
'<div class="footnote">%s</div>'%(p['footnote']),
'</div>']
#else is plot
else:
fig_num = fig_num + 1
if 'anchor' in p:
anchor = p['anchor']
else:
anchor = ""
html_p = ['<div class="figure %s" id="%s">'%(source, anchor),
#title
'<div class="title"><b>Figure %d.</b> %s</div>'%(fig_num, p['title']),
#date
'<div class="date"> Last Updated: ' + '%s'%(date) + '</div>',
#figure
'<a class="img" href="img\\'+p['file']+'">\n<img class="img" src="img\\'+p['file']+'">\n</a>',
#footnote
'<div class="footnote">%s</div>'%(p['footnote']),
'</div>']
#----append to html plots
html_plots.append(('\n\t'.join(map(str, html_p))))
html_plots = '\n\t'.join(map(str, html_plots))
else:
html_plots = ""
##body
body = ['<div class="container-large %s" style="">'%(source),
'<div class="container" style="">',
'<div class="dashboard-main">',
'<div class="table-container %s">'%(source),
'<div class="title">'+'%s'%(title)+'</div>',
'<div class="date"> Last Updated: '+'%s'%(date),
'<span class="link" id="code">code</span>',
'<span class="link" id="csv">csv</span>',
'</div>',
'<div class="dataTables_wrapper form-inline dt-bootstrap">',
'<!--script-->',
table,
'</div>',
'<div class="footnote">' + '\n' + '%s'%(footnote) + '</div>',
html_plots,
'</div>',
'</div>',
'</div>',
'</div>']
#replace script within body, then join as string
if ((source == "logit") or (source == "onset") or (source == "anova")):
body = [script if x==( '<!--script-->') else x for x in body]
# join
body = ('\n\t'.join(map(str, body)))
#build footer
if ((source == "logit") or (source == "onset") or (source == "anova")):
foot = ['<script>',
'$(document).ready(function(){',
'getLogitTable();',
'additionalTable();',
'popover();',
'codeButton();',
'})',
'</script>']
elif (source == "task"):
foot = ['<script>',
'$(document).ready(function(){',
'getLogitTable();',
'popover();',
'})',
'</script>']
else:
foot = ['<script>',
'$(document).ready(function(){',
'getLogitTable();',
'})',
'</script>']
foot = ('\n\t'.join(map(str, foot)))
#build html
html = ['<html short="%s" long="%s">'%(short, long),
'<head>',
head,
'</head>',
'<body id="body">',
body,
'</body>',
'<foot>',
foot,
'</foot>',
'</html>']
html = ('\n'.join(map(str, html)))
#----save
# check if paths exist
if not os.path.exists(os.path.dirname(path)):
os.makedirs(os.path.dirname(path))
# tidy html
#html = re.sub("\s\s+", " ", html) #multiple spaces
# save
with open(path,'w') as html_:
html_.write(html)
console('saved html file at %s'%(path), 'blue')
#--------finished
console('%s finished in %s msec'%(_f,((datetime.datetime.now()-_t0).total_seconds()*1000)), 'blue')
return html