slider.js_on_change(
'value',
sliderCallback(source_sample, depth_sample, slider, slider_af,
syngroup, unique_longitudinal_muts))
syngroup.js_on_change(
'active',
sliderCallback(source_sample, depth_sample, slider, slider_af,
syngroup, unique_longitudinal_muts))
ose.js_on_change(
'value',
sliderCallback2(source_sample, depth_sample, slider, slider_af,
syngroup, ose, unique_longitudinal_muts))
# When mousing over Bokeh plot, allele frequency updated to user input.
if (user_af != -123):
slider_af.value = user_af
## g.js_on_event(events.PlotEvent, sliderCallback(source_sample, depth_sample, slider, slider_af, syngroup))
genome_plot.js_on_event(
events.MouseEnter,
sliderCallback(source_sample, depth_sample, slider, slider_af,
syngroup, unique_longitudinal_muts))
# Creates labels with read information
reads_info = (reads.loc[reads['Sample'] == name.strip()])
div = Div(text="""<font size="2" color="gray"><b>Total reads: </b>""" +
str(reads_info.iloc[0]["Total"]) +
"""<br><b>Total reads mapped: </b>""" +
str(reads_info.iloc[0]["Mapped"]) +
"""<br><b>Percentage of reads mapped: </b>
""" + str(reads_info.iloc[0]["Percentage"]) + """</font>""",
width=300,
def interactive_hist(adata, keys=['n_counts', 'n_genes'],
bins='auto', max_bins=100,
groups=None, fill_alpha=0.4,
palette=None, display_all=True,
tools='pan, reset, wheel_zoom, save',
legend_loc='top_right',
plot_width=None, plot_height=None, save=None,
*args, **kwargs):
"""Utility function to plot distributions with variable number of bins.
Params
--------
adata: AnnData object
annotated data object
keys: list(str), optional (default: `['n_counts', 'n_genes']`)
keys in `adata.obs` or `adata.var` where the distibutions are stored
bins: int; str, optional (default: `auto`)
number of bins used for plotting or str from numpy.histogram
max_bins: int, optional (default: `1000`)
maximum number of bins possible
groups: list(str), (default: `None`)
keys in `adata.obs.obs_keys()`, groups by all possible combinations of values, e.g. for
3 plates and 2 time points, we would create total of 6 groups
fill_alpha: float[0.0, 1.0], (default: `0.4`)
alpha channel of the fill color
palette: list(str), optional (default: `None`)
palette to use
display_all: bool, optional (default: `True`)
display the statistics for all data
tools: str, optional (default: `'pan,reset, wheel_zoom, save'`)
palette of interactive tools for the user
legend_loc: str, (default: `'top_right'`)
position of the legend
legend_loc: str, default(`'top_left'`)
position of the legend
plot_width: int, optional (default: `None`)
width of the plot
plot_height: int, optional (default: `None`)
height of the plot
save: Union[os.PathLike, Str, NoneType], optional (default: `None`)
path where to save the plot
*args, **kwargs: arguments, keyword arguments
addition argument to bokeh.models.figure
Returns
--------
None
"""
if max_bins < 1:
raise ValueError(f'`max_bins` must >= 1')
palette = Set1[9] + Set2[8] + Set3[12] if palette is None else palette
# check the input
for key in keys:
if key not in adata.obs.keys() and \
key not in adata.var.keys() and \
key not in adata.var_names:
raise ValueError(f'The key `{key}` does not exist in `adata.obs`, `adata.var` or `adata.var_names`.')
def _create_adata_groups():
if groups is None:
return [adata], [('all',)]
combs = list(product(*[set(adata.obs[g]) for g in groups]))
adatas= [adata[reduce(lambda l, r: l & r,
(adata.obs[k] == v for k, v in zip(groups, vals)), True)]
for vals in combs] + [adata]
if display_all:
combs += [('all',)]
adatas += [adata]
return adatas, combs
# group_v_combs contains the value combinations
ad_gs = _create_adata_groups()
cols = []
for key in keys:
callbacks = []
fig = figure(*args, tools=tools, **kwargs)
slider = Slider(start=1, end=max_bins, value=0, step=1,
title='Bins')
plots = []
for j, (ad, group_vs) in enumerate(filter(lambda ad_g: ad_g[0].n_obs > 0, zip(*ad_gs))):
if key in ad.obs.keys():
orig = ad.obs[key]
hist, edges = np.histogram(orig, density=True, bins=bins)
elif key in ad.var.keys():
orig = ad.var[key]
hist, edges = np.histogram(orig, density=True, bins=bins)
else:
orig = ad[:, key].X
hist, edges = np.histogram(orig, density=True, bins=bins)
slider.value = len(hist)
# case when automatic bins
max_bins = max(max_bins, slider.value)
# original data, used for recalculation of histogram in JS code
orig = ColumnDataSource(data=dict(values=orig))
# data that we update in JS code
source = ColumnDataSource(data=dict(hist=hist, l_edges=edges[:-1], r_edges=edges[1:]))
legend = ', '.join(': '.join(map(str, gv)) for gv in zip(groups, group_vs)) \
if groups is not None else 'all'
p = fig.quad(source=source, top='hist', bottom=0,
left='l_edges', right='r_edges',
fill_color=palette[j], legend_label=legend if legend_loc is not None else None,
muted_alpha=0,
line_color="#555555", fill_alpha=fill_alpha)
# create callback and slider
callback = CustomJS(args=dict(source=source, orig=orig), code=_inter_hist_js_code)
callback.args['bins'] = slider
callbacks.append(callback)
# add the current plot so that we can set it
# visible/invisible in JS code
plots.append(p)
slider.end = max_bins
# slider now updates all values
slider.js_on_change('value', *callbacks)
button = Button(label='Toggle', button_type='primary')
button.callback = CustomJS(
args={'plots': plots},
code='''
for (var i = 0; i < plots.length; i++) {
plots[i].muted = !plots[i].muted;
}
'''
)
if legend_loc is not None:
fig.legend.location = legend_loc
fig.legend.click_policy = 'mute'
fig.xaxis.axis_label = key
fig.yaxis.axis_label = 'normalized frequency'
_set_plot_wh(fig, plot_width, plot_height)
cols.append(column(slider, button, fig))
if _bokeh_version > (1, 0, 4):
from bokeh.layouts import grid
plot = grid(children=cols, ncols=2)
else:
cols = list(map(list, np.array_split(cols, np.ceil(len(cols) / 2))))
plot = layout(children=cols, sizing_mode='fixed', ncols=2)
if save is not None:
save = save if str(save).endswith('.html') else str(save) + '.html'
bokeh_save(plot, save)
else:
show(plot)
days_from_death = days_from_death[index]
rate = rate[index]
return ([days_from_death, rate, time])
text1 = Select(title='Sate1', options=list(covid_states.keys()))
text2 = Select(title='Param', options=['Positive', 'Active', 'Deaths'])
inf_time = Slider(title="Days Spent Infecting Others",
value=10.0,
start=2,
end=20,
step=1)
text1.value = 'California'
text2.value = 'Active'
inf_time.value = 12
state = text1.value
param = text2.value
infection_time = inf_time.value
days_from_death, y, time = transform(state, param, infection_time)
pred, forcast, forcast_L, forcast_U, y_new, confi = gam_results(
days_from_death, y, covid_states[state], 'Active', infection_time)
for_dates = time.iloc[-1] + pd.to_timedelta(np.arange(10), unit='d')
plot = figure(plot_height=400,
plot_width=900,
title="Active Infections",
tools="crosshair,pan,reset,save,wheel_zoom",
x_axis_type="datetime")
