def run_item_clicked(self, item):
logging.info('Run item %s clicked' % item.text(0))
output = io.StringIO()
if item.parent() is not None:
if item.parent().text(0) == 'Variables':
cpt = self._concrete_model.find_component(item.text(0))
# create ggplot
df = mo.get_entity(cpt)
if item.text(0) == 'S':
ff = pn.ggplot(df, pn.aes('T', item.text(0))) + pn.ggtitle(cpt.doc) +\
pn.geom_step(pn.aes(color='States'), direction='hv') + pn.facet_wrap('States')
elif item.text(0) == 'Q':
ff = pn.ggplot(df, pn.aes('T', item.text(0))) + pn.ggtitle(cpt.doc) +\
pn.geom_step(pn.aes(color='J'), direction='hv') + pn.facet_grid('J~')
else:
ff = pn.ggplot(df, pn.aes('T', item.text(0))) + pn.ggtitle(cpt.doc) +\
pn.geom_step(pn.aes(color='J'), direction='hv') + pn.facet_grid('I~')
size = self.canvas.size()
ff += pn.theme(figure_size=(size.width() / 100, size.height() / 100))
# update to the new figure
fig = ff.draw()
self.canvas.figure = fig
self.canvas.draw()
output.close()
def test_step():
p = (ggplot(df, aes('x')) +
geom_step(aes(y='y'), size=4) +
geom_step(aes(y='y+2'), color='red',
direction='vh', size=4))
assert p == 'step'
def test_step():
p = (ggplot(df, aes('x')) +
geom_step(aes(y='y'), size=4) +
geom_step(aes(y='y+2'), color='red',
direction='vh', size=4))
assert p == 'step'
def test_step_mid():
df = pd.DataFrame({'x': range(9), 'y': range(9)})
p = (ggplot(df, aes('x', 'y'))
+ geom_point(size=4)
+ geom_step(direction='mid', size=2)
)
assert p == 'step_mid'
def test_line():
df2 = df.copy()
# geom_path plots in given order. geom_line &
# geom_step sort by x before plotting
df2['x'] = df['x'].values[::-1]
p = (ggplot(df2, aes('x')) + geom_path(aes(y='y'), size=4) +
geom_line(aes(y='y+2'), color='blue', size=4) +
geom_step(aes(y='y+4'), color='red', size=4))
assert p == 'path_line_step'
def test_line():
df2 = df.copy()
# geom_path plots in given order. geom_line &
# geom_step sort by x before plotting
df2['x'] = df['x'].values[::-1]
p = (ggplot(df2, aes('x')) +
geom_path(aes(y='y'), size=4) +
geom_line(aes(y='y+2'), color='blue', size=4) +
geom_step(aes(y='y+4'), color='red', size=4))
assert p == 'path_line_step'
sv = scale_predictors(df, predictor='SVC')
# ld = scale_predictors(df, predictor='LDA')
nb = scale_predictors(df, predictor='naive_bayes')
rn = scale_predictors(df, predictor='Random')
ac = scale_predictors(df, predictor='acg_ip_risk')
rf = scale_predictors(df, predictor='RandmForest')
ct = scale_predictors(df, predictor='cheating')
df2 = pd.concat([nb, rn, ac, rf, sv, ct])
# df2 = pd.concat([nb, rn, ac, rf, ct])
print(df2.head(20))
print(df2.describe())
p = pn.ggplot(df2, pn.aes(x='num_examined', y='num_detected', group='classifier', colour='classifier')) +\
pn.geom_step() +\
pn.ggtitle("How Many ppl would we need to intervene on to prevent Y hospitalizations?")
# pn.scales.scale_x_reverse()
p.save(HOME_DIR + 'all_together_d.png', height=8, width=10, units='in', verbose=False)
p2 = pn.ggplot(df2, pn.aes(x='num_examined', y='num_detected', group='classifier', colour='classifier')) +\
pn.geom_step() +\
pn.ggtitle("How Many ppl would we need to intervene on to prevent Y hospitalizations?") +\
pn.xlim(0, 300) + pn.ylim(0, 300)
# pn.scales.scale_x_reverse()
p2.save(HOME_DIR + 'all_together_trunc.png', height=8, width=10, units='in', verbose=False)
print("Finished!")