def get_layout(self):
# Set up plot
plot = self.plot
plot.line('x', 'y', source=self.source, line_width=3, line_alpha=0.6)
plot.renderers.append(self.vline)
self.trigger_plot = plot.line('x',
'trigger',
source=self.source,
line_width=3,
line_alpha=0.6)
self.trigger_plot.glyph.line_alpha = 0
self.trigger_plot.glyph.line_color = "#15bd42"
self.trigger_plot.glyph.line_width = 1
trigger_slider = md.Slider(start=-2.5,
end=2.5,
value=2,
step=0.01,
title="Trigger Value")
trigger_slider.on_change("value", self.on_trigger_moved)
width_slider = md.Slider(start=1,
end=1000,
value=10,
step=1,
title='Width (ms)')
width_slider.on_change("value", self.on_width_moved)
scale_slider = md.Slider(start=0.1,
end=10,
value=1,
step=0.01,
title='Scale')
scale_slider.on_change("value", self.on_scale_moved)
self.trigger_button = md.buttons.Toggle(label="Trigger")
self.trigger_button.on_click(self.trigger_clicked)
raise_or_fall = md.RadioButtonGroup(labels=["Raising", "Falling"],
active=self.raise_or_fall_value)
raise_or_fall.on_change('active', self.raise_or_fall_clicked)
scope = row(
plot,
column(
row(
self.trigger_button,
raise_or_fall,
),
trigger_slider,
width_slider,
scale_slider,
),
)
self.layour = scope
return scope
def _build_int_option(self, option_name, title):
current = self.model.options[option_name]
select_ui = mdl.Slider(title=title, value=current, start=1, end=30)
def on_change(attr, old_state, new_state):
assert attr == 'value'
del old_state
self.controller.set_option(option_name, new_state)
select_ui.on_change('value', on_change)
self.option_uis[option_name] = select_ui
def get_layout(self):
step = self.parameter.step
if step is None:
step = (self.parameter.max - self.parameter.min) / 1000
self.slider = slider = md.Slider(start=self.parameter.min,
end=self.parameter.max,
title=self.parameter.name,
value=self.parameter.value,
step=step)
slider.on_change("value", self.on_value_change)
return slider
class FittedParamWidgets:
"""Collection of Bokeh Widgets for fitting a light-curve parameters"""
top_section_div = models.Div(text=(
'<h2>Fitted Parameters</h2>'
'<p>Current parameter values are used as an initial guess when fitting. '
'Note that redshift values are not varied as part of the fit.</p>'))
# User input widgets for setting model parameters
fit_t0_slider = models.Slider(start=-10,
end=10,
value=0,
step=1E-4,
title='t0')
fit_x0_slider = models.Slider(start=0.001,
end=2,
value=.1,
step=1E-4,
title='x0')
fit_x1_slider = models.Slider(start=-1,
end=1,
value=0,
step=1E-4,
title='x1')
fit_c_slider = models.Slider(start=-1,
end=1,
value=0,
step=1E-4,
title='c')
plot_model_button = models.Button(label='Plot Model',
button_type='warning')
fit_button = models.Button(label='Fit Light-Curve', button_type='success')
fitted_params_widgets_list = [
top_section_div, fit_t0_slider, fit_x0_slider, fit_x1_slider,
fit_c_slider, plot_model_button, fit_button
]
class SimulatedParamWidgets:
"""Collection of Bokeh Widgets for specifying simulation parameters"""
top_section_div = models.Div(text='<h2>Simulated Parameters</h2>')
# User input widgets for setting model parameters
sim_z_slider = models.Slider(start=0.001,
end=1,
value=.55,
step=.01,
title='z')
sim_t0_slider = models.Slider(start=-10,
end=10,
value=-2,
step=.01,
title='t0')
sim_x0_slider = models.Slider(start=0.001,
end=2,
value=.25,
step=.01,
title='x0')
sim_x1_slider = models.Slider(start=-1,
end=1,
value=0.11,
step=.01,
title='x1')
sim_c_slider = models.Slider(start=-1,
end=1,
value=-.05,
step=.01,
title='c')
sampling_input = models.TextInput(value='4', title='Sampling (Days):')
sim_pwv_slider = models.Slider(start=-0,
end=15,
value=7,
step=.1,
title='PWV')
plot_button = models.Button(label='Plot Light-Curve',
button_type='success')
snr_input = models.TextInput(value='10.0', title='SNR:', default_size=220)
checkbox = models.CheckboxGroup(
labels=["Plot SNR", 'Subtract Reference Star'], active=[0])
# Having all inputs as a list is useful when constructing layouts
# as it establishes the default column order
sim_params_widgets_list = [
top_section_div, sim_z_slider, sim_t0_slider, sim_x0_slider,
sim_x1_slider, sim_c_slider, sim_pwv_slider, sampling_input, snr_input,
checkbox, plot_button
]
def example_doc_modifier():
df = sea_surface_temperature.copy()
source = bkm.ColumnDataSource(data=df)
plot = bkp.figure(x_axis_type='datetime',
y_range=(0, 25),
y_axis_label='Temperature (Celsius)',
title="Sea Surface Temperature at 43.18, -70.43")
plot.line('time', 'temperature', source=source)
def callback(attr, old, new):
if new == 0:
data = df
else:
data = df.rolling('{0}D'.format(new)).mean()
source.data = bkm.ColumnDataSource(data=data).data
slider = bkm.Slider(start=0,
end=30,
value=0,
step=1,
title="Smoothing by N Days")
slider.on_change('value', callback)
return slider, plot
def make_slider(prop, start, end, value):
slider = models.Slider(title=prop, start=start, end=end, value=value)
add_callback(slider, prop)
return slider
def __init__(self):
self.numeric_eqn = sp.lambdify(self.all_symbols,
sp.Matrix(self.symbolic_eqns))
ys = np.linspace(*self.y_range, self.vector_density[0])
dys = np.linspace(*self.dy_range, self.vector_density[1])
y_grid, dy_grid = np.meshgrid(ys, dys)
self._ys = y_grid.flatten()
self._dys = dy_grid.flatten()
self.vector_source = models.ColumnDataSource(data=dict(
xs=[],
ys=[],
))
self.trajectories_source = models.ColumnDataSource(data=dict(
xs=[],
ys=[],
))
self.trajectory_starts_source = models.ColumnDataSource(data=dict(
x=[],
y=[],
))
self.plot = figure(
width=FIGURE_WIDTH,
height=FIGURE_HEIGHT,
x_range=self.y_range,
y_range=self.dy_range,
title='Phase portrait',
toolbar_location=None,
)
self.plot.xaxis.axis_label = 'y'
self.plot.yaxis.axis_label = 'ẏ'
self.plot.axis.axis_label_text_font_size = AXIS_TITLE_FONT_SIZE
self.plot.axis.major_label_text_font_size = AXIS_TICK_FONT_SIZE
self.plot.on_event(events.Tap, self.plot_clicked)
self.plot.scatter(x=self._ys, y=self._dys)
self.plot.multi_line(
xs='xs',
ys='ys',
source=self.vector_source,
line_width=VECTOR_LINE_WIDTH,
)
self.setup_trajectory_glyphs(self.trajectory_starts_source,
self.trajectories_source)
self.t_slider = models.Slider(
title='Trajectory duration',
width=SLIDER_WIDTH,
**TIME_SLIDER_PARAMS,
)
self.param_sliders = {}
for param, range_, step, value in zip(
self.param_symbols,
self.param_ranges,
self.param_steps,
self.param_defaults,
):
slider = models.Slider(
start=range_[0],
end=range_[1],
step=step,
value=value,
title=str(param),
width=SLIDER_WIDTH,
)
slider.on_change('value', self.on_slider_change)
self.param_sliders[param] = slider
def make_responsive_plot(prot, tag):
#output_file("test_deg_def.html", title="line deg")
title = tag
fig_fit = get_base_plot(prot, title)
#fig_fit.set(y_range=bkm.Range1d(-0.1, 1.1))
x_lim = prot.X[-1] + 0.1
X = np.arange(-0.1, x_lim, 0.1)
Ys = model_syn(X, prot.N0, prot.tau, prot.offset)
Yd = model_deg(X, prot.N0, prot.tau, prot.offset)
params_start = [prot.N0, prot.tau, prot.offset, title]
params = [prot.N0, prot.tau, prot.offset, title]
diff = len(Yd) - len(params)
add = [np.nan] * diff
params_start += add
params += add
source = bkm.ColumnDataSource(data=dict(
X=X,
Ys=Ys,
Yd=Yd,
params_start=params_start,
params=params,
))
code = """
var data = source.get('data');
var params_start = data['params_start']
var X = data['X'];
var Ys = data['Ys'];
var Yd = data['Yd'];
var params = data['params'];
var title = params[3];
var f = cb_obj.get('value');
if (cb_obj.get('title') == 'N0') {N0 = cb_obj.get('value'); params[0] = N0;};
if (cb_obj.get('title') == 'tau') {tau = cb_obj.get('value'); params[1] = tau; };
if (cb_obj.get('title') == 'offset') {offset = cb_obj.get('value'); params[2] = offset;};
var N0 = parseFloat(params[0]);
var tau = parseFloat(params[1]);
var offset = parseFloat(params[2]);
//console.log(params, 'here parms');
//console.log(N0, 'here N0');
//console.log(tau, 'here tau');
//console.log(offset, 'here offset');
//console.log(Ys, 'here Ys before');
//console.log(Yd, 'here Yd before');
for (i = 0; i < X.length; i++) {
Ys[i] = 1 - ((N0 * Math.exp( -X[i]/tau )) + offset) ;
}
for (i = 0; i < X.length; i++) {
Yd[i] = (N0 * Math.exp( -X[i]/tau )) + offset ;
}
var last_element = X[X.length - 1];
var residual = (N0 * Math.exp( -last_element/tau )) + offset
//console.log(Ys, 'here Ys after');
//console.log(Yd, 'here Yd after');
source.trigger('change');
console.log(N0, 'here N0 after');
console.log(tau, 'here tau after');
console.log(offset, 'here offset after');
console.log( tau / Math.log(2) ,'here half life')
//$("field_name_1").update(tau/Math.log(2));
//$("field_name_2").update(tau);
//$("field_name_3").update(offset);
$('#field_name_1').text(Math.round(tau/Math.log(2)* 100)/100);
$('#field_name_2').text(Math.round(residual * 100)/100);
//$('#field_name_3').text(offset);
//document.getElementById("field_name_1").innerHTML = N0;
// $( ".hello" ).remove();
//var $new = $('<div class="hello"></div>').appendTo('.insert_body');
//http://stackoverflow.com/questions/247483/http-get-request-in-javascript
//function httpGet(theUrl)
//{
//var xmlHttp = new XMLHttpRequest();
//xmlHttp.open( "GET", theUrl, false ); // false for synchronous request
//xmlHttp.send( null );
//return xmlHttp.responseText;
//}
//var address = '{web_address}find_intersection?N0='+N0+'&tau='+tau+'&c='+offset
//var res = httpGet(address);
//console.log(res, 'test');
//if (title == 'Bloodstrem'){
//console.log('new_intersection_bs');
//$("#new_intersection_bs").html('new intersection: '+res);
//}
//if (title == 'Procyclic'){
// $("#new_intersection_pc").html('new intersection: '+res);
//}
//res = httpGet(address)
//console.log(res, 'test');
//ras = $.get("address");
//req.open('GET', address, false);
//
//console.log(address);
"""
#code = code.replace('{web_address}', web_adress)
callback = bkm.CustomJS(args=dict(source=source), code=code)
sliderN0 = bkm.Slider(start=0,
end=1,
value=prot.N0,
step=.01,
title="N0",
name='N0',
callback=callback)
sliderTau = bkm.Slider(start=0.1,
end=50,
value=prot.tau,
step=.01,
title="tau",
name='tau',
callback=callback)
sliderOffSet = bkm.Slider(start=0,
end=1,
value=prot.offset,
step=.01,
title="offset",
name='offset',
callback=callback)
syn = fig_fit.line(x='X', y='Ys', source=source, color='green')
deg = fig_fit.line(x='X', y='Yd', source=source, color='red')
legend = Legend(items=[("syn", [syn]), ("deg", [deg])], location=(10, -30))
#fig_fit.legend.location = (0, -30)
#layout = vform(fig_fit)
#script, div = components(layout)
#layout = vform()
#script_bar, div_bar = components(layout)
#show(layout(
#[[fig_fit],[sliderN0],[sliderTau],[sliderOffSet]]
#))
fig_fit.add_layout(legend, 'right')
#show(layouts.column(fig_fit, sliderN0, sliderTau, sliderOffSet))
script, div = components(
layouts.column(fig_fit, sliderN0, sliderTau, sliderOffSet))
return script, div
# Grab the data from the file, into a pandas array
bokeh_data = read_data(fname)
# Create a bokeh figure
p = bkp.figure(title=title, plot_width=1600, plot_height=700)
# Add data to it
p.scatter(x='phase', y='flux', source=bokeh_data, color=color)
# Format some stuff
p.xaxis.axis_label = 'Phase'
p.yaxis.axis_label = 'Flux'
# Plot the figure
fig = p
#### Lets add a slider
slider = bkm.Slider(title='Flux offset',
start=-1,
end=1,
value=0.0,
width=600,
step=0.01)
#TODO:
slider.on_change('value', alter_data)
# Make a layout, and add to document
layout = bkl.column([fig, slider])
bkp.curdoc().add_root(layout)
bkp.curdoc().title = __name__
def generate_plot(data, df_all_prediction):
COLORS = d3["Category10"][10]
tooltips = f"""
<div>
<p>
<span style="font-size: 12px; color: black;font-family:century gothic;">Nombre de cas : </span>
<span style="font-size: 12px; color: {COLORS[0]}; font-weight: bold;font-family:century gothic;">@total_cases</span>
<br>
<span style="font-size: 12px; color: black;font-family:century gothic;">Nombre de nouveaux cas : </span>
<span style="font-size: 12px; color: {COLORS[1]}; font-weight: bold;font-family:century gothic;">@new_cases</span>
<br>
<span style="font-size: 12px; color: black;font-family:century gothic;">Nombre de deces : </span>
<span style="font-size: 12px; color: {COLORS[3]}; font-weight: bold;font-family:century gothic;">@total_deaths</span>
<br>
<span style="font-size: 12px; color: black;font-family:century gothic;">Nombre nouveaux deces : </span>
<span style="font-size: 12px; color: {COLORS[5]}; font-weight: bold;font-family:century gothic;">@new_deaths</span>
<br>
<span style="font-size: 12px; color: black;font-family:century gothic;">Date : </span>
<span style="font-size: 12px; color: black; font-weight: bold;font-family:century gothic;">@date_str</span>
</p>
</div>
"""
tooltips_predictions = (f"""
<div>
<p>
<span style="font-size: 12px; color: black;font-family:century gothic;">Prédiction nombre de cas : </span>
<span style="font-size: 12px; color: {COLORS[0]}; font-weight: bold;font-family:century gothic;">@median_display</span>
<br>
<span style="font-size: 12px; color: black;font-family:century gothic;">Date : </span>
<span style="font-size: 12px; color: black; font-weight: bold;font-family:century gothic;">"""
+ """@date_str</span>
</p>
</div>
""")
hover = bkm.tools.HoverTool(names=["line_total"],
tooltips=tooltips,
mode="vline")
hover_prediction = bkm.tools.HoverTool(
names=["prediction"],
tooltips=tooltips_predictions,
)
# --- define all DataSource needed --- #
source_all = bkm.ColumnDataSource(data)
country = "World"
source = bkm.ColumnDataSource(get_country(data, country))
source_all_prediction = bkm.ColumnDataSource(df_all_prediction)
source_prediction = bkm.ColumnDataSource(
get_country(df_all_prediction, country))
date_end_training = np.unique(
get_country(df_all_prediction, country)["date_end_train"])[-1]
source_prediction_end_date = bkm.ColumnDataSource(
get_country(df_all_prediction, country)[get_country(
df_all_prediction, country).date_end_train == date_end_training])
slider = bkm.Slider(
start=0,
end=len(
np.unique(
get_country(df_all_prediction, country)["date_end_train"])) -
1,
value=0,
step=1,
title="Days dropped for prediction",
)
# ----------- #
p = bkp.figure(
y_axis_type="linear",
x_axis_type="datetime",
sizing_mode="stretch_both",
title=f"Covid 19 evolution: {country}",
x_axis_label="date",
y_axis_label="Total number of Covid 19 cases",
tools=[hover, "pan", "wheel_zoom", "reset"],
x_range=[
get_country(data, country).date.min(),
get_country(data, country).date.max() + datetime.timedelta(days=1),
],
y_range=[
-get_country(data, country).total_cases.max() * 0.05,
get_country(data, country).total_cases.max() * 1.1,
],
)
p.yaxis.formatter = bkm.formatters.NumeralTickFormatter(format="0,0")
p.xaxis.formatter = bkm.formatters.DatetimeTickFormatter(
days=["%d/%m", "%d%a"], months=["%m/%Y", "%b %Y"])
p.add_tools(hover_prediction)
p.toolbar.active_drag = None
# p.toolbar.active_scroll = p.select_one(bkm.WheelZoomTool)
y_extra_range_max = np.max([
np.max(get_country(data, country).new_cases.values),
np.max(get_country(data, country).total_deaths.values),
])
p.extra_y_ranges = {
"Number of deaths":
bkm.Range1d(start=-0.05 * y_extra_range_max,
end=1.1 * y_extra_range_max)
}
p.add_layout(
bkm.LinearAxis(
y_range_name="Number of deaths",
axis_label="New Covid 19 cases",
formatter=bkm.formatters.NumeralTickFormatter(format="0,0"),
),
"right",
)
# --- plot total cases --- #
p.line(
source=source,
x="date",
y="total_cases",
name="line_total",
color=COLORS[0],
legend_label="total cases",
muted_alpha=0.1,
)
p.circle(source=source,
x="date",
y="total_cases",
color=COLORS[0],
muted_alpha=0.1)
# --- plot new cases --- #
p.vbar(
source=source,
x="date",
top="new_cases",
color=COLORS[1],
width=50e6,
alpha=0.5,
name="bar",
y_range_name="Number of deaths",
legend_label="new cases",
muted_alpha=0.1,
)
# --- plot total death --- #
p.line(
source=source,
x="date",
y="total_deaths",
color=COLORS[3],
y_range_name="Number of deaths",
name="line_death",
legend_label="total deaths",
muted_alpha=0.1,
)
p.circle(
source=source,
x="date",
y="total_deaths",
color=COLORS[3],
y_range_name="Number of deaths",
muted_alpha=0.1,
)
# --- plot new death --- #
p.vbar(
source=source,
x="date",
top="new_deaths",
color=COLORS[5],
width=50e6,
alpha=0.5,
y_range_name="Number of deaths",
legend_label="new deaths",
muted_alpha=0.1,
)
button_click_count = bkm.ColumnDataSource({"clicks": [0]})
select = bkm.Select(title="Country: ",
value=country,
options=list(data.location.unique()))
button_log = bkm.Button(label="Log Scale", button_type="primary")
# --- Predictions --- #
median_prediction = p.line(
source=source_prediction_end_date,
x="date",
y="median",
line_color=COLORS[0],
name="prediction",
)
prediction_cases_line = p.line(
source=source_prediction_end_date,
x="date",
y="derivative",
color=COLORS[1],
y_range_name="Number of deaths",
)
band_low = bkm.Band(
source=source_prediction_end_date,
base="date",
lower="25%",
upper="median",
fill_color=COLORS[0],
level="underlay",
fill_alpha=0.1,
line_width=0.5,
line_color="black",
)
band_high = bkm.Band(
source=source_prediction_end_date,
base="date",
lower="median",
upper="75%",
fill_color=COLORS[0],
level="underlay",
fill_alpha=0.1,
line_width=0.5,
line_color="black",
)
median_prediction.visible = False
prediction_cases_line.visible = False
band_low.visible = False
band_high.visible = False
p.add_layout(band_low)
p.add_layout(band_high)
button_prediction = bkm.Button(label="Show predictions",
button_type="primary")
# -- Callback -- #
callback = bkm.CustomJS(
args=dict(
source=source,
source_all=source_all,
select=select,
x_range=p.x_range,
y_range_left=p.y_range,
y_range_right=p.extra_y_ranges["Number of deaths"],
title=p.title,
button_click_count=button_click_count,
slider=slider,
source_all_prediction=source_all_prediction,
source_prediction=source_prediction,
source_prediction_end_date=source_prediction_end_date,
median_prediction=median_prediction,
band_low=band_low,
prediction_cases_line=prediction_cases_line,
band_high=band_high,
),
code="""
var country = select.value
var date = source_all.data['date']
var date_str = source_all.data['date_str']
var location = source_all.data['location']
var total_cases = source_all.data['total_cases']
var new_cases = source_all.data['new_cases']
var total_deaths = source_all.data['total_deaths']
var new_deaths = source_all.data['new_deaths']
var new_date = []
var new_date_str = []
var new_total_cases = []
var new_new_cases = []
var new_total_deaths = []
var new_new_deaths = []
for(var i=0; i < date.length; i++){
if(location[i]==country){
new_date.push(date[i]);
new_date_str.push(date_str[i])
new_total_cases.push(total_cases[i]);
new_new_cases.push(new_cases[i]);
new_total_deaths.push(total_deaths[i]);
new_new_deaths.push(new_deaths[i]);
}
}
source.data['date']=new_date;
source.data['date_str']=new_date_str;
source.data['total_cases']=new_total_cases;
source.data['new_cases']=new_new_cases;
source.data['total_deaths']=new_total_deaths;
source.data['new_deaths']=new_new_deaths;
const new_cases_no_Nan = new_new_cases.filter(function (value) {
return !Number.isNaN(value);
});
const cases_no_Nan = new_total_cases.filter(function (value) {
return !Number.isNaN(value);
});
y_range_right.setv({"start": -0.05*Math.max.apply(Math, new_cases_no_Nan.concat(new_total_deaths)),
"end": 1.1*Math.max.apply(Math, new_cases_no_Nan.concat(new_total_deaths))})
y_range_left.setv({"start": -0.05*Math.max.apply(Math, cases_no_Nan),
"end": 1.1*Math.max.apply(Math, cases_no_Nan)})
x_range.setv({"start": Math.min.apply(Math, new_date), "end": 1.0001*Math.max.apply(Math, new_date)})
title.text = "Evolution du nombre de cas en " + country
source.change.emit();
// change value of predictions
button_click_count.data.clicks = 0
median_prediction.visible = false
band_low.visible = false
band_high.visible = false
prediction_cases_line.visble = false
var date_end_prediction = source_all_prediction.data['date_end_train']
var location = source_all_prediction.data['location']
var date = source_all_prediction.data['date']
var date_str = source_all_prediction.data['date_str']
var quantile_1 = source_all_prediction.data['25%']
var quantile_2 = source_all_prediction.data['median']
var quantile_3 = source_all_prediction.data['75%']
var new_cases = source_all_prediction.data['derivative']
var median_prediction = source_all_prediction.data['median_display']
var new_date = []
var new_date_str = []
var new_date_end_prediction = []
var new_quantile_1 = []
var new_quantile_2 = []
var new_quantile_3 = []
var new_new_cases = []
var new_median_prediction = []
for(var i=0; i < quantile_1.length; i++){
if(location[i]==country){
new_date.push(date[i])
new_date_str.push(date_str[i])
new_date_end_prediction.push(date_end_prediction[i])
new_quantile_1.push(quantile_1[i]);
new_quantile_2.push(quantile_2[i]);
new_quantile_3.push(quantile_3[i]);
new_new_cases.push(new_cases[i]);
new_median_prediction.push(median_prediction[i]);
}
}
source_prediction.data['date']=new_date
source_prediction.data['date_str']=new_date_str
source_prediction.data['date_end_train']=new_date_end_prediction
source_prediction.data['25%']=new_quantile_1;
source_prediction.data['median']=new_quantile_2;
source_prediction.data['75%']=new_quantile_3;
source_prediction.data['derivative']=new_new_cases;
source_prediction.data['median_display']=new_median_prediction;
var n = new_date.length
var max_date = Math.max.apply(Math, new_date_end_prediction)
var new_date_bis = []
var new_date_str_bis = []
var new_date_end_prediction_bis = []
var new_quantile_1_bis = []
var new_quantile_2_bis = []
var new_quantile_3_bis = []
var new_new_cases_bis = []
var new_median_prediction_bis = []
for(var i=0; i < n; i++){
if(new_date_end_prediction[i]==max_date){
new_date_bis.push(new_date[i])
new_date_str_bis.push(new_date_str[i])
new_date_end_prediction_bis.push(new_date_end_prediction[i])
new_quantile_1_bis.push(new_quantile_1[i]);
new_quantile_2_bis.push(new_quantile_2[i]);
new_quantile_3_bis.push(new_quantile_3[i]);
new_new_cases_bis.push(new_new_cases[i]);
new_median_prediction_bis.push(new_median_prediction[i]);
}
}
var n = new_date_bis.length
var max_date = Math.max.apply(Math, new_date_end_prediction_bis)
source_prediction_end_date.data['date']=new_date_bis
source_prediction_end_date.data['date_str']=new_date_str_bis
source_prediction_end_date.data['date_end_train']=new_date_end_prediction_bis
source_prediction_end_date.data['25%']=new_quantile_1_bis;
source_prediction_end_date.data['median']=new_quantile_2_bis;
source_prediction_end_date.data['75%']=new_quantile_3_bis;
source_prediction_end_date.data['derivative']=new_new_cases_bis;
source_prediction_end_date.data['median_display']=new_median_prediction_bis;
source_prediction.change.emit();
source_prediction_end_date.change.emit()
const unique = (value, index, self) => {
return self.indexOf(value) === index
}
// change slider value
slider.setv({"end": new_date_end_prediction.filter(unique).length - 1, "value": 0})
""",
)
callback_button = bkm.CustomJS(
args=dict(y_axis=p.left, title=p.title),
code="""
console.log(y_axis)
y_axis = LogAxis()
""",
)
select.js_on_change("value", callback)
button_log.js_on_click(callback_button)
callback_button = bkm.CustomJS(
args=dict(
source=source,
source_prediction=source_prediction,
source_all_prediction=source_all_prediction,
source_prediction_end_date=source_prediction_end_date,
select=select,
button_prediction=button_prediction,
median_prediction=median_prediction,
band_low=band_low,
prediction_cases_line=prediction_cases_line,
band_high=band_high,
button_click_count=button_click_count,
x_range=p.x_range,
y_range_left=p.y_range,
y_range_right=p.extra_y_ranges["Number of deaths"],
),
code="""
// function to get unique value of an array
const unique = (value, index, self) => {
return self.indexOf(value) === index
}
var date = source.data['date'];
var total_cases = source.data['total_cases'];
var new_cases = source.data['new_cases'];
var total_deaths = source.data['total_deaths'];
var date_prediction = source_prediction.data['date'];
var total_cases_prediction = source_prediction.data['75%'];
const new_cases_no_Nan = new_cases.filter(function (value) {
return !Number.isNaN(value);
});
const cases_no_Nan = total_cases.filter(function (value) {
return !Number.isNaN(value);
});
var country = select.value
button_click_count.data.clicks ++
var show_prediction = (button_click_count.data.clicks % 2) == 1
var locations_predicted = source_all_prediction.data['location'].filter(unique)
if (locations_predicted.includes(country) == false){
window.alert("This country doesn't have prediction: Available countries are: " + locations_predicted);
}
else{
if (show_prediction == true){
median_prediction.visible = true
band_low.visible = true
band_high.visible = true
prediction_cases_line.visble = true
const y_range_right_values = [].concat([].slice.call(new_cases_no_Nan), [].slice.call(total_deaths))
y_range_left.setv({"start": -0.05*Math.max.apply(Math, total_cases_prediction), "end": 1.1 * Math.max.apply(Math, total_cases_prediction)})
y_range_right.setv({"start": -0.05*Math.max.apply(Math, y_range_right_values) * Math.max.apply(Math, total_cases_prediction) / Math.max.apply(Math, cases_no_Nan),
"end": 1.1*Math.max.apply(Math, y_range_right_values) * Math.max.apply(Math, total_cases_prediction) / Math.max.apply(Math, cases_no_Nan)})
x_range.setv({"start": Math.min.apply(Math, date_prediction), "end": 1.0001*Math.max.apply(Math, date_prediction)})
}
else{
median_prediction.visible = false
band_low.visible = false
band_high.visible = false
prediction_cases_line.visble = false
const y_range_right_values = [].concat([].slice.call(new_cases_no_Nan), [].slice.call(total_deaths))
y_range_left.setv({"start": -0.05*Math.max.apply(Math, cases_no_Nan), "end": 1.1*Math.max.apply(Math, cases_no_Nan)})
y_range_right.setv({"start": -0.05*Math.max.apply(Math, y_range_right_values), "end": 1.1*Math.max.apply(Math, y_range_right_values)})
x_range.setv({"start": Math.min.apply(Math, date), "end": 1.0001*Math.max.apply(Math, date)})
}
}
""",
)
button_prediction.js_on_click(callback_button)
callback_slider = bkm.CustomJS(
args=dict(
source=source,
source_prediction=source_prediction,
source_all_prediction=source_all_prediction,
source_prediction_end_date=source_prediction_end_date,
select=select,
prediction_cases_line=prediction_cases_line,
slider=slider,
button_click_count=button_click_count,
x_range=p.x_range,
y_range_left=p.y_range,
y_range_right=p.extra_y_ranges["Number of deaths"],
),
code="""
// function to get unique value of an array
const unique = (value, index, self) => {
return self.indexOf(value) === index
}
var slider_value = slider.value
var country = select.value
var date_prediction = source_prediction.data['date']
var date_str = source_prediction.data['date_str']
var date_end_prediction = source_prediction.data['date_end_train']
var quantile_1 = source_prediction.data['25%'];
var quantile_2 = source_prediction.data['median']
var quantile_3 = source_prediction.data['75%']
var new_cases = source_prediction.data['derivative'];
var median_prediction = source_prediction.data['median_display']
var unique_end_prediction = date_end_prediction.filter(unique)
var show_prediction = (button_click_count.data.clicks % 2) == 1
var locations_predicted = source_all_prediction.data['location'].filter(unique)
if (show_prediction == true && locations_predicted.includes(country)){
var new_date_prediction = []
var new_date_str = []
var new_date_end_prediction = []
var new_quantile_1 = []
var new_quantile_2 = []
var new_quantile_3 = []
var new_new_cases = []
var new_median_prediction = []
for(var i=0; i < quantile_1.length; i++){
if(date_end_prediction[i]==unique_end_prediction[slider.end - slider_value]){
new_date_prediction.push(date_prediction[i])
new_date_str.push(date_str[i])
new_date_end_prediction.push(date_end_prediction[i])
new_quantile_1.push(quantile_1[i]);
new_quantile_2.push(quantile_2[i]);
new_quantile_3.push(quantile_3[i]);
new_new_cases.push(new_cases[i]);
new_median_prediction.push(median_prediction[i]);
}
}
source_prediction_end_date.data['date']=new_date_prediction
source_prediction_end_date.data['date_str']=new_date_str
source_prediction_end_date.data['date_end_train']=new_date_end_prediction
source_prediction_end_date.data['25%']=new_quantile_1;
source_prediction_end_date.data['median']=new_quantile_2;
source_prediction_end_date.data['75%']=new_quantile_3;
source_prediction_end_date.data['derivative']=new_new_cases;
source_prediction_end_date.data['median_display']=new_median_prediction;
source_prediction_end_date.change.emit();
var date_prediction = source_prediction_end_date.data['date'];
var total_cases_prediction = source_prediction_end_date.data['75%'];
const new_cases_no_Nan = new_cases.filter(function (value) {
return !Number.isNaN(value);
});
const cases_no_Nan = quantile_2.filter(function (value) {
return !Number.isNaN(value);
});
// y_range_left.setv({"start": -0.05*Math.max.apply(Math, total_cases_prediction), "end": 1.1*Math.max.apply(Math, total_cases_prediction)})
// x_range.setv({"start": Math.min.apply(Math, date_prediction), "end": 1.0001*Math.max.apply(Math, date_prediction)})
}
""",
)
slider.js_on_change("value", callback_slider)
p.legend.location = "top_left"
p.legend.click_policy = "mute"
return select, button_prediction, slider, p
def add_numerical_filter(self, filter_mode='>=', title='Value', numeric_name='Altitude', step=50, height_policy='min', callback_policy='value_throttled', callback_class=None, **kwargs):
"""
Add a Numerical Filter to the Canvas
Parameters
----------
filter_mode: str (default: >=)
The operator that the filter will use on the loaded dataset (allowed operators: '<', '<=', '>', '>=', '==', '!=', 'range')
title: str (default: 'Value')
The title of the categorical filter
numeric_name: str (default: ```'Altitude'```)
The column name of the loaded dataset that contains the numeric information
step: float (default: 50)
The step in which the filter will move within the numeric range of the dataset.
height_policy: str (default: 'min')
Describes how the component should maintain its height (accepted values: 'auto', 'fixed', 'fit', 'min', 'max')
From: https://docs.bokeh.org/en/1.1.0/docs/reference/models/layouts.html#bokeh.models.layouts.LayoutDOM.height_policy
callback_policy: str (default: 'value_throttled')
Describes when the callback will be triggered. If callback_policy == 'value', the callback will be fired with each change,
while if callback_policy == 'value_throttled' the callback will be executed only when the value is set (i.e., on mouseup).
callback_class: callbacks.BokehFilters (default: None)
Allows custom callback methods to be set. If None, the baseline callback method is used.
**kwargs: Dict
Other parameters related to the filter creation
"""
kwargs.pop('value', None)
if filter_mode not in list(ALLOWED_FILTER_OPERATORS.keys()):
raise ValueError(f'filter_mode must be one of the following: {list(ALLOWED_FILTER_OPERATORS.keys())}')
start, end = self.data[numeric_name].agg([np.min, np.max])
if filter_mode != 'range':
# value = start if value is None else value
value = start
num_filter = bokeh_mdl.Slider(start=start, end=end, step=step, value=value, title=title, height_policy=height_policy, **kwargs)
else:
# value = (start, end) if value is None else value
value = (start, end)
num_filter = bokeh_mdl.RangeSlider(start=start, end=end, step=step, value=value, title=title, height_policy=height_policy, **kwargs)
if callback_class is None:
class Callback(callbacks.BokehFilters):
def __init__(self, vsn_instance, widget):
super().__init__(vsn_instance, widget)
def callback(self, attr, old, new):
self.callback_filter_data()
num_value = new
new_pts = self.get_data()
if filter_mode == 'range':
new_pts = new_pts.loc[new_pts[numeric_name].between(num_value[0], num_value[1], inclusive=True)]
else:
new_pts = new_pts.loc[ALLOWED_FILTER_OPERATORS[filter_mode](new_pts[numeric_name], num_value)]
self.callback_prepare_data(new_pts, self.widget.id==self.vsn_instance.aquire_canvas_data)
callback_class = Callback
num_filter.on_change(callback_policy, callback_class(self, num_filter).callback)
self.widgets.append(num_filter)
def __init__(self,
phase_y_range=(-2.5, 2.5),
potential_y_range=(-2.5, 2.5)):
phi, b, d_omega = sp.symbols('phi, b, Delta_omega')
self.state_symbol = phi
self.param_symbols = b, d_omega
self.symbolic_eqn = d_omega - sp.sin(phi) - 2 * b * sp.sin(2 * phi)
self.numeric_eqn = sp.lambdify((phi, b, d_omega), self.symbolic_eqn)
self.symbolic_slope = sp.diff(self.symbolic_eqn, self.state_symbol)
self.numeric_slope = sp.lambdify((phi, b, d_omega),
self.symbolic_slope)
self.symbolic_pot = -sp.integrate(self.symbolic_eqn, self.state_symbol)
self.numeric_pot = sp.lambdify((phi, b, d_omega), self.symbolic_pot)
self.roots = self.get_roots()
self.phase_plot, self.pot_plot = self.init_plots(
phase_y_range, potential_y_range)
# Set up empty DataSources. As much as possible, these are shared between the plots.
self.line_source = models.ColumnDataSource(data=dict(
x=[],
phase=[],
v=[],
))
self.attractor_source = models.ColumnDataSource(data=dict(
x=[],
phase=[],
v=[],
))
self.repeller_source = models.ColumnDataSource(data=dict(
x=[],
phase=[],
v=[],
))
self.half_source = models.ColumnDataSource(data=dict(
x=[],
phase=[],
v=[],
))
self.arrow_source = models.ColumnDataSource(data=dict(
x=[],
angle=[],
))
# Draw all the glyphs.
for plot, y in [(self.phase_plot, 'phase'), (self.pot_plot, 'v')]:
if y == 'v':
y_circle = 'v_circle'
else:
y_circle = y
plot.line(
x='x',
y=y,
source=self.line_source,
line_width=LINE_WIDTH,
)
plot.circle(
x='x',
y=y_circle,
source=self.attractor_source,
size=CIRCLE_SIZE,
)
plot.circle(
x='x',
y=y_circle,
source=self.repeller_source,
size=CIRCLE_SIZE,
fill_color=None,
line_width=CIRCLE_LINE_WIDTH,
)
plot.circle(
x='x',
y=y_circle,
source=self.half_source,
size=CIRCLE_SIZE,
fill_color=None,
line_width=CIRCLE_LINE_WIDTH,
)
plot.wedge(
x='x',
y=y_circle,
source=self.half_source,
radius=CIRCLE_SIZE / 2,
start_angle=-np.pi / 2,
end_angle=np.pi / 2,
radius_units='screen',
line_color=None,
)
self.phase_plot.triangle(
x='x',
y=0,
angle='angle',
source=self.arrow_source,
size=ARROW_SIZE,
color='black',
)
self.b_slider = models.Slider(
start=self.b_values[0],
end=self.b_values[-1],
step=self.b_values[1] - self.b_values[0],
value=0.5,
title='b',
width=FIGURE_WIDTH // 2,
)
self.d_omega_slider = models.Slider(
start=self.d_omega_values[0],
end=self.d_omega_values[-1],
step=self.d_omega_values[1] - self.d_omega_values[0],
value=0,
title='Δω',
width=FIGURE_WIDTH // 2,
)
self.b_slider.on_change('value', self.on_slider_change)
self.d_omega_slider.on_change('value', self.on_slider_change)
self.update_sources()