def plot_compare_accuracy(self, expo=False):
if expo:
return (
ggplot(self.acc_df) + facet_wrap('position')
+ aes(x='guesser', y='accuracy', fill='Dataset')
+ geom_bar(stat='identity', position='dodge')
+ xlab('Guessing Model')
+ ylab('Accuracy')
)
else:
return (
ggplot(self.acc_df) + facet_wrap('position')
+ aes(x='guesser', y='accuracy')
+ geom_bar(stat='identity')
)
def test_ribbon_facetting():
p = (ggplot(df, aes('x', ymin='ymin', ymax='ymax',
fill='factor(z)')) +
geom_ribbon() +
facet_wrap('~ z')
)
assert p + _theme == 'ribbon_facetting'
def plot_char_percent_vs_accuracy_histogram(self, category=False):
if category:
return (
ggplot(self.char_plot_df) + facet_wrap('category_jmlr')
+ aes(x='char_percent', fill='Outcome')
+ geom_histogram(binwidth=.05)
)
else:
return (
ggplot(self.char_plot_df)
+ aes(x='char_percent', fill='Outcome')
+ geom_histogram(binwidth=.05)
)
def create_confidence_plot(conf_df):
plt = (
ggplot(conf_df)
+ aes(x='x', color='Method', fill='Method')
+ geom_density(alpha=.45)
+ facet_wrap('Task', nrow=4)
+ xlab('Confidence')
+ scale_color_manual(values=COLORS)
+ scale_fill_manual(values=COLORS)
+ theme_fs()
+ theme(
axis_text_y=element_blank(),
axis_ticks_major_y=element_blank(),
axis_title_y=element_blank(),
legend_title=element_blank(),
legend_position='top',
legend_box='horizontal',
)
)
return plt
def create_length_plot(len_df, legend_position='right', legend_box='vertical'):
mean_len_df = len_df.groupby(['Task', 'Method']).mean().reset_index()
mean_len_df[' '] = 'Mean Length'
plt = (
ggplot(len_df)
+ aes(x='x', fill='Method', y='..density..')
+ geom_histogram(binwidth=2, position='identity', alpha=.6)
+ geom_text(
aes(x='x', y=.22, label='x', color='Method'),
mean_len_df,
inherit_aes=False,
format_string='{:.1f}',
show_legend=False
)
+ geom_segment(
aes(x='x', xend='x', y=0, yend=.205, linetype=' '),
mean_len_df,
inherit_aes=False, color='black'
)
+ scale_linetype_manual(['dashed'])
+ facet_wrap('Task')
+ xlim(0, 20) + ylim(0, .23)
+ xlab('Example Length') + ylab('Frequency')
+ scale_color_manual(values=COLORS)
+ scale_fill_manual(values=COLORS)
+ theme_fs()
+ theme(
aspect_ratio=1,
legend_title=element_blank(),
legend_position=legend_position,
legend_box=legend_box,
)
)
return plt
def test_facet_wrap_one_var():
p = g + facet_wrap('~var1')
p2 = g + facet_wrap('~class') # python keyword in formula
assert p == 'facet_wrap_one_var'
assert p2 == 'facet_wrap_one_var'
def test_aslabeller_func_hashtag():
func = as_labeller(lambda s: '#{}'.format(s))
p = g + facet_wrap('~ gear + am', labeller=func)
assert p == 'aslabeller_func_hashtagit'
def test_label_context_wrap2vars():
p = g + facet_wrap('~ gear + am', labeller='label_context')
assert p == 'label_context_wrap2vars'
def test_non_mapped_facetting():
p = (g + geom_abline(intercept=0, slope=1, size=1) + facet_wrap('var1'))
assert p == 'non_mapped_facetting'
def syntactic_diversity_plots():
with open('data/external/syntactic_diversity_table.json') as f:
rows = json.load(f)
parse_df = pd.DataFrame(rows)
parse_df['parse_ratio'] = parse_df['unique_parses'] / parse_df['parses']
melt_df = pd.melt(
parse_df,
id_vars=['dataset', 'depth', 'overlap', 'parses'],
value_vars=['parse_ratio', 'unique_parses'],
var_name='metric',
value_name='y'
)
def label_facet(name):
if name == 'parse_ratio':
return 'Average Unique Parses per Instance'
elif name == 'unique_parses':
return 'Count of Unique Parses'
def label_y(ys):
formatted_ys = []
for y in ys:
y = str(y)
if y.endswith('000.0'):
formatted_ys.append(y[:-5] + 'K')
else:
formatted_ys.append(y)
return formatted_ys
p = (
ggplot(melt_df)
+ aes(x='depth', y='y', color='dataset')
+ facet_wrap('metric', scales='free_y', nrow=2, labeller=label_facet)
+ geom_line() + geom_point()
+ xlab('Parse Truncation Depth') + ylab('')
+ scale_color_discrete(name='Dataset')
+ scale_y_continuous(labels=label_y)
+ scale_x_continuous(
breaks=list(range(1, 11)),
minor_breaks=list(range(1, 11)),
limits=[1, 10])
+ theme_fs()
)
p.save(path.join(output_path, 'syn_div_plot.pdf'))
p = (
ggplot(parse_df)
+ aes(x='depth', y='unique_parses', color='dataset')
+ geom_line() + geom_point()
+ xlab('Parse Truncation Depth')
+ ylab('Count of Unique Parses')
+ scale_color_discrete(name='Dataset')
+ scale_x_continuous(
breaks=list(range(1, 11)),
minor_breaks=list(range(1, 11)),
limits=[1, 10])
+ theme_fs()
)
p.save(path.join(output_path, 'n_unique_parses.pdf'))
p = (
ggplot(parse_df)
+ aes(x='depth', y='parse_ratio', color='dataset')
+ geom_line() + geom_point()
+ xlab('Parse Truncation Depth')
+ ylab('Average Unique Parses per Instance')
+ scale_color_discrete(name='Dataset')
+ scale_x_continuous(breaks=list(range(1, 11)), minor_breaks=list(range(1, 11)), limits=[1, 10])
+ scale_y_continuous(limits=[0, 1])
+ theme_fs()
)
p.save(path.join(output_path, 'parse_ratio.pdf'))
def test_facet_wrap_direction_v():
p = g + facet_wrap('~var1', dir='v')
assert p == 'facet_wrap_direction_v'
def test_facet_wrap_expression():
p = g + facet_wrap('pd.cut(var1, (0, 2, 4), include_lowest=True)')
assert p == 'facet_wrap_expression'
def plot_tag_repartition(self, data, options):
tag_df = data["tags"]
if not "background" in tag_df.columns:
tag_df["background"] = False
test = tag_df[["tag", "matched", "background", "id"]].copy()
test.loc[:, "prop_matched"] = -1
test.loc[:, "prop_background"] = -1
test.loc[:, "lbl_matched"] = ""
test.loc[:, "lbl_background"] = ""
test.loc[:, "n_tags"] = -1
n_total = test.shape[0]
n_matched = test.matched.value_counts()
tags_summary = (
test.groupby("tag")
.apply(self.get_proportions, by=["matched", "background"])
.reset_index(drop=True)
)
tags_summary = tags_summary.sort_values(["tag", "matched", "background"])
plt = ggplot(
data=tags_summary,
mapping=aes(
x="tag", # "factor(species, ordered=False)",
y="n_tags",
fill="background",
ymax=max(tags_summary.n_tags) + 35, # "factor(species, ordered=False)",
),
)
plot_width = 10 + len(tags_summary.tag.unique()) * 0.75
plt = (
plt
+ geom_bar(stat="identity", show_legend=True, position=position_dodge())
+ facet_wrap(
"matched",
nrow=1,
ncol=2,
scales="fixed",
labeller=(lambda x: self.get_matched_label(x, n_total, n_matched)),
)
+ xlab("Species")
+ ylab("Number of annotations")
+ geom_text(
mapping=aes(label="lbl_background"), position=position_dodge(width=0.9),
)
+ geom_text(
mapping=aes(y=max(tags_summary.n_tags) + 30, label="lbl_matched",)
)
+ theme_classic()
+ theme(
axis_text_x=element_text(angle=90, vjust=1, hjust=1, margin={"r": -30}),
plot_title=element_text(
weight="bold", size=14, margin={"t": 10, "b": 10}
),
figure_size=(plot_width, 10),
text=element_text(size=12, weight="bold"),
)
+ ggtitle(
(
"Tag repartition for model {}, database {}, class {}\n"
+ "with detector options {}"
).format(
options["scenario_info"]["model"],
options["scenario_info"]["database"],
options["scenario_info"]["class"],
options,
)
)
)
return plt
.mean()
.reset_index()
)
for plate in platelist:
os.makedirs(output_figuresdir, exist_ok=True)
by_well_gg = (
gg.ggplot(
cell_count_totalcells_df.loc[
cell_count_totalcells_df["site"].str.contains(plate)
],
gg.aes(x="x_loc", y="y_loc"),
)
+ gg.geom_point(gg.aes(fill="total_cell_count"), shape="s", size=6)
+ gg.geom_text(gg.aes(label="site_location"), color="lightgrey", size=6)
+ gg.facet_wrap("~well")
+ gg.coord_fixed()
+ gg.theme_bw()
+ gg.ggtitle(f"Total Cells/Well\n{plate}")
+ gg.theme(
axis_text=gg.element_blank(),
axis_title=gg.element_blank(),
strip_background=gg.element_rect(colour="black", fill="#fdfff4"),
)
+ gg.labs(fill="Cells")
+ gg.scale_fill_cmap(name="Number of Cells")
)
output_file = pathlib.Path(
output_figuresdir, f"plate_layout_cells_count_per_well_{plate}.png"
)
knnResultsSimplified = pd.DataFrame(
[(x['p'], x['k'], x['cvAccuracy'], x['testAccuracy'])
for x in repeatedKnnResults],
columns=['p', 'k', 'cvAccuracy', 'testAccuracy'])
ggdata = pd.concat([
pd.DataFrame({
'p': knnResultsSimplified.p,
'k': knnResultsSimplified.k.apply(int),
'type': 'cv',
'Accuracy': knnResultsSimplified.cvAccuracy
}),
pd.DataFrame({
'p': knnResultsSimplified.p,
'k': knnResultsSimplified.k.apply(int),
'type': 'test',
'Accuracy': knnResultsSimplified.testAccuracy
})
],
axis=0)
ggo = gg.ggplot(
ggdata,
gg.aes(x='p', y='Accuracy', color='type', group='type', linetype='type'))
ggo += gg.scale_x_log10()
ggo += gg.geom_point(alpha=0.6)
ggo += gg.stat_smooth()
ggo += gg.facet_wrap('~ k')
ggo += gg.theme_bw()
print(ggo)
# In[46]:
all_data_df.columns = [
'PC1', 'PC2', 'num_partitions', 'comparison', 'No. of partitions',
'Comparison'
]
# In[52]:
# Plot all comparisons in one figure
panel_B = ggplot(all_data_df[all_data_df['Comparison'] != '1'],
aes(x='PC1', y='PC2')) \
+ geom_point(aes(color='No. of partitions'),
alpha=0.2) \
+ facet_wrap('~Comparison') \
+ labs(x = "PC 1",
y = "PC 2",
title = "PCA of partition 1 vs multiple partitions") \
+ theme_bw() \
+ theme(
legend_title_align = "center",
plot_background=element_rect(fill='white'),
legend_key=element_rect(fill='white', colour='white'),
legend_text=element_text(family='sans-serif', size=12),
plot_title=element_text(family='sans-serif', size=15),
axis_text=element_text(family='sans-serif', size=12),
axis_title=element_text(family='sans-serif', size=15)
) \
+ guides(colour=guide_legend(override_aes={'alpha': 1})) \
+ scale_color_manual(['#bdbdbd', '#b3e5fc']) \
std_accuracy=pd.NamedAgg('accuracy', np.std),
)
summary.reset_index(inplace=True)
print('Summary for all participants')
print(summary)
summary_per_participant = trials.groupby(
by=['participant_id', 'freq_category']).agg(np.mean)
summary_per_participant.reset_index(inplace=True)
print('Summary per participant')
print(summary_per_participant)
plot = (
ggplot(gg.aes(x='freq_category', y='reaction_time'), data=trials) +
gg.geom_boxplot(gg.aes(fill='freq_category')) + # of jitter
gg.facet_wrap('participant_id'))
plot.draw()
plt.show()
plot = (ggplot(gg.aes(x="freq_category", weight='accuracy'),
summary_per_participant) + gg.geom_bar() +
gg.facet_wrap(['participant_id']))
plot.draw()
plt.show()
plot = (
ggplot(gg.aes(x='freq_category', y='reaction_time'), data=trials) +
gg.geom_boxplot(gg.aes(fill='freq_category')) # of jitter
)
plot.draw()
plt.show()
from plotnine import (ggplot, aes, geom_point, facet_wrap,
stat_smooth, theme_xkcd)
from plotnine.data import mtcars
kwargs = dict(width=6, height=4)
p1 = (ggplot(mtcars, aes('wt', 'mpg'))
+ geom_point())
p1.save('readme-image-1.png', **kwargs)
p2 = p1 + aes(color='factor(gear)')
p2.save('readme-image-2.png', **kwargs)
p3 = p2 + stat_smooth(method='lm')
p3.save('readme-image-3.png', **kwargs)
p4 = p3 + facet_wrap('~gear')
p4.save('readme-image-4.png', **kwargs)
p5 = p4 + theme_xkcd()
p5.save('readme-image-5.png', **kwargs)
def plot_char_percent_vs_accuracy_smooth(self,
expo=False,
no_models=False,
columns=False):
if self.y_max is not None:
limits = [0, float(self.y_max)]
eprint(f'Setting limits to: {limits}')
else:
limits = [0, 1]
if expo:
if os.path.exists('data/external/all_human_gameplay.json'
) and not self.no_humans:
with open('data/external/all_human_gameplay.json') as f:
all_gameplay = json.load(f)
frames = []
for event, name in [('parents', 'Intermediate'),
('maryland', 'Expert'),
('live', 'National')]:
if self.merge_humans:
name = 'Human'
gameplay = all_gameplay[event]
if event != 'live':
control_correct_positions = gameplay[
'control_correct_positions']
control_wrong_positions = gameplay[
'control_wrong_positions']
control_positions = control_correct_positions + control_wrong_positions
control_positions = np.array(control_positions)
control_result = np.array(
len(control_correct_positions) * [1] +
len(control_wrong_positions) * [0])
argsort_control = np.argsort(control_positions)
control_x = control_positions[argsort_control]
control_sorted_result = control_result[
argsort_control]
control_y = control_sorted_result.cumsum(
) / control_sorted_result.shape[0]
control_df = pd.DataFrame({
'correct': control_y,
'char_percent': control_x
})
control_df['Dataset'] = 'Regular Test'
control_df['Guessing_Model'] = f' {name}'
frames.append(control_df)
adv_correct_positions = gameplay[
'adv_correct_positions']
adv_wrong_positions = gameplay['adv_wrong_positions']
adv_positions = adv_correct_positions + adv_wrong_positions
adv_positions = np.array(adv_positions)
adv_result = np.array(
len(adv_correct_positions) * [1] +
len(adv_wrong_positions) * [0])
argsort_adv = np.argsort(adv_positions)
adv_x = adv_positions[argsort_adv]
adv_sorted_result = adv_result[argsort_adv]
adv_y = adv_sorted_result.cumsum(
) / adv_sorted_result.shape[0]
adv_df = pd.DataFrame({
'correct': adv_y,
'char_percent': adv_x
})
adv_df['Dataset'] = 'IR Adversarial'
adv_df['Guessing_Model'] = f' {name}'
frames.append(adv_df)
if len(gameplay['advneural_correct_positions']) > 0:
adv_correct_positions = gameplay[
'advneural_correct_positions']
adv_wrong_positions = gameplay[
'advneural_wrong_positions']
adv_positions = adv_correct_positions + adv_wrong_positions
adv_positions = np.array(adv_positions)
adv_result = np.array(
len(adv_correct_positions) * [1] +
len(adv_wrong_positions) * [0])
argsort_adv = np.argsort(adv_positions)
adv_x = adv_positions[argsort_adv]
adv_sorted_result = adv_result[argsort_adv]
adv_y = adv_sorted_result.cumsum(
) / adv_sorted_result.shape[0]
adv_df = pd.DataFrame({
'correct': adv_y,
'char_percent': adv_x
})
adv_df['Dataset'] = 'RNN Adversarial'
adv_df['Guessing_Model'] = f' {name}'
frames.append(adv_df)
human_df = pd.concat(frames)
human_vals = sort_humans(
list(human_df['Guessing_Model'].unique()))
human_dtype = CategoricalDtype(human_vals, ordered=True)
human_df['Guessing_Model'] = human_df[
'Guessing_Model'].astype(human_dtype)
dataset_dtype = CategoricalDtype(
['Regular Test', 'IR Adversarial', 'RNN Adversarial'],
ordered=True)
human_df['Dataset'] = human_df['Dataset'].astype(
dataset_dtype)
if no_models:
p = ggplot(human_df) + geom_point(shape='.')
else:
df = self.char_plot_df
if 1 not in self.rounds:
df = df[df['Dataset'] != 'Round 1 - IR Adversarial']
if 2 not in self.rounds:
df = df[df['Dataset'] != 'Round 2 - IR Adversarial']
df = df[df['Dataset'] != 'Round 2 - RNN Adversarial']
p = ggplot(df)
if self.save_df is not None:
eprint(f'Saving df to: {self.save_df}')
df.to_json(self.save_df)
if os.path.exists('data/external/all_human_gameplay.json'
) and not self.no_humans:
eprint('Loading human data')
p = p + geom_line(data=human_df)
if columns:
facet_conf = facet_wrap('Guessing_Model', ncol=1)
else:
facet_conf = facet_wrap('Guessing_Model', nrow=1)
if not no_models:
if self.mvg_avg_char:
chart = stat_smooth(method='mavg',
se=False,
method_args={'window': 400})
else:
chart = stat_summary_bin(fun_data=mean_no_se,
bins=20,
shape='.',
linetype='None',
size=0.5)
else:
chart = None
p = (p + facet_conf +
aes(x='char_percent', y='correct', color='Dataset'))
if chart is not None:
p += chart
p = (
p + scale_y_continuous(breaks=np.linspace(0, 1, 6)) +
scale_x_continuous(breaks=[0, .5, 1]) +
coord_cartesian(ylim=limits) +
xlab('Percent of Question Revealed') + ylab('Accuracy') +
theme(
#legend_position='top', legend_box_margin=0, legend_title=element_blank(),
strip_text_x=element_text(margin={
't': 6,
'b': 6,
'l': 1,
'r': 5
})) + scale_color_manual(
values=['#FF3333', '#66CC00', '#3333FF', '#FFFF33'],
name='Questions'))
if self.title != '':
p += ggtitle(self.title)
return p
else:
if self.save_df is not None:
eprint(f'Saving df to: {self.save_df}')
df.to_json(self.save_df)
return (ggplot(self.char_plot_df) + aes(
x='char_percent', y='correct', color='Guessing_Model') +
stat_smooth(
method='mavg', se=False, method_args={'window': 500}) +
scale_y_continuous(breaks=np.linspace(0, 1, 6)) +
coord_cartesian(ylim=limits))
# In[11]:
# Side by side original input vs simulated data
# Add label for input or simulated dataset
input_data_UMAPencoded_df['dataset'] = 'original'
simulated_data_UMAPencoded_df['dataset'] = 'simulated'
# Concatenate input and simulated dataframes together
combined_data_df = pd.concat(
[input_data_UMAPencoded_df, simulated_data_UMAPencoded_df])
# Plot
ggplot(combined_data_df, aes(
x='1', y='2')) + geom_point(alpha=0.3) + facet_wrap('~dataset') + labs(
x="UMAP 1", y="UMAP 2", title="UMAP of original and simulated data")
# In[12]:
# Overlay original input vs simulated data
# Add label for input or simulated dataset
input_data_UMAPencoded_df['dataset'] = 'original'
simulated_data_UMAPencoded_df['dataset'] = 'simulated'
# Concatenate input and simulated dataframes together
combined_data_df = pd.concat(
[input_data_UMAPencoded_df, simulated_data_UMAPencoded_df])
# Plot
def test_non_mapped_facetting():
p = (g
+ geom_abline(intercept=0, slope=1, size=1)
+ facet_wrap('var1')
)
assert p == 'non_mapped_facetting'
# In[12]:
for level in levels:
all_feature_results_subset_df = all_feature_results_df.query(
"level == @level").reset_index(drop=True)
output_dir = pathlib.Path(f"{output_fig_dir}/{level}")
output_dir.mkdir(exist_ok=True)
# Figure 1 - Per plate feature differences
per_plate_feature_gg = (
gg.ggplot(all_feature_results_subset_df,
gg.aes(x="plate", y="metric_value")) +
gg.geom_point(size=0.1, alpha=0.5) +
gg.facet_wrap("~metric", scales="free", nrow=len(metrics)) +
gg.xlab("Plate") + gg.ylab("Feature Difference\nBetween Tools") +
gg.ggtitle(f"Plate Summary\n{level}") + theme_summary)
output_file = pathlib.Path(f"{output_dir}/{level}_metrics_per_plate.png")
per_plate_feature_gg.save(output_file, dpi=dpi, height=height, width=width)
print(per_plate_feature_gg)
del per_plate_feature_gg
# In[13]:
for level in levels:
all_feature_results_subset_df = all_feature_results_df.query(
"level == @level").reset_index(drop=True)
def test_facet_wrap_label_both():
p = g + facet_wrap('~var1+var2', labeller='label_both')
assert p == 'facet_wrap_label_both'
# print(aci)
res
# res.to_feather("data_glm.feather")
def label_x(dates):
res = [(datetime.datetime(2018, 1, 1) +
datetime.timedelta(x)).strftime("%d-%m") for x in dates]
print(res)
return res
(ggplot(data=res, mapping=aes(x='julian', y='ACI_mean', colour='site')) +
xlab("Day") + ylab("Mean daily ACI (standardized)")
# + facet_grid("site~", scales="free")
+ facet_wrap("site", nrow=2, ncol=3) + geom_point() +
geom_errorbar(aes(ymin="ACI_mean - ACI_std", ymax="ACI_mean + ACI_std")) +
geom_smooth(method="mavg",
se=False,
method_args={
"window": 4,
"center": True,
"min_periods": 1
}) + scale_colour_manual(values=cbbPalette, guide=False) +
scale_x_continuous(labels=label_x)).save("figs/ACI_all_testfacet3.png",
height=10,
width=16,
dpi=150)
#################
### Denoising ###
def plot_portfolio(portfolio_df,
figure_size=(12, 4),
line_size=1.5,
date_text_size=7):
"""
Given a daily snapshot of virtual purchases plot both overall and per-stock
performance. Return a tuple of figures representing the performance as inline data.
"""
assert portfolio_df is not None
#print(portfolio_df)
portfolio_df['date'] = pd.to_datetime(portfolio_df['date'])
avg_profit_over_period = portfolio_df.filter(
items=['stock', 'stock_profit']).groupby('stock').mean()
avg_profit_over_period['contribution'] = [
'positive' if profit >= 0.0 else 'negative'
for profit in avg_profit_over_period.stock_profit
]
avg_profit_over_period = avg_profit_over_period.drop(
'stock_profit',
axis='columns') # dont want to override actual profit with average
portfolio_df = portfolio_df.merge(avg_profit_over_period,
left_on='stock',
right_index=True,
how='inner')
#print(portfolio_df)
# 1. overall performance
df = portfolio_df.filter(items=[
'portfolio_cost', 'portfolio_worth', 'portfolio_profit', 'date'
])
df = df.melt(id_vars=['date'], var_name='field')
plot = (
p9.ggplot(df, p9.aes('date', 'value', group='field', color='field')) +
p9.labs(x='', y='$ AUD') + p9.geom_line(size=1.5) +
p9.facet_wrap('~ field', nrow=3, ncol=1, scales='free_y') +
p9.theme(axis_text_x=p9.element_text(angle=30, size=date_text_size),
figure_size=figure_size,
legend_position='none'))
overall_figure = plot_as_inline_html_data(plot)
df = portfolio_df.filter(
items=['stock', 'date', 'stock_profit', 'stock_worth', 'contribution'])
melted_df = df.melt(id_vars=['date', 'stock', 'contribution'],
var_name='field')
all_dates = sorted(melted_df['date'].unique())
df = melted_df[melted_df['date'] == all_dates[-1]]
df = df[df['field'] == 'stock_profit'] # only latest profit is plotted
df['contribution'] = [
'positive' if profit >= 0.0 else 'negative' for profit in df['value']
]
# 2. plot contributors ie. winners and losers
plot = (p9.ggplot(df, p9.aes('stock', 'value', fill='stock')) +
p9.geom_bar(stat='identity') + p9.labs(x='', y='$ AUD') +
p9.facet_grid('contribution ~ field') +
p9.theme(legend_position='none', figure_size=figure_size))
profit_contributors = plot_as_inline_html_data(plot)
# 3. per purchased stock performance
plot = (
p9.ggplot(melted_df,
p9.aes('date', 'value', group='stock', colour='stock')) +
p9.xlab('') + p9.geom_line(size=1.0) +
p9.facet_grid('field ~ contribution', scales="free_y") + p9.theme(
axis_text_x=p9.element_text(angle=30, size=date_text_size),
figure_size=figure_size,
panel_spacing=
0.5, # more space between plots to avoid tick mark overlap
subplots_adjust={'right': 0.8}))
stock_figure = plot_as_inline_html_data(plot)
return overall_figure, stock_figure, profit_contributors
def test_label_both():
p = g + facet_wrap('~ gear', labeller='label_both')
assert p == 'label_both'
def plot_cellranger_vs_cellbender(samplename, raw_cellranger_mtx,
filtered_cellranger_mtx,
cellbender_unfiltered_h5, fpr,
n_expected_cells, n_total_droplets_included,
out_dir):
"""compare cellranger raw vs cellranger filtered vs cellbender outputs"""
logging.info('samplename ' + str(samplename))
logging.info('raw_cellranger_mtx ' + str(raw_cellranger_mtx))
logging.info('filtered_cellranger_mtx ' + str(filtered_cellranger_mtx))
logging.info('cellbender_unfiltered_h5 ' + str(cellbender_unfiltered_h5))
logging.info('fpr ' + str(fpr))
logging.info('n_expected_cells ' + str(n_expected_cells))
logging.info('n_total_droplets_included ' + str(n_total_droplets_included))
logging.info('out_dir ' + str(out_dir))
# Make the output directory if it does not exist.
if out_dir == '':
out_dir = os.getcwd()
else:
os.makedirs(out_dir, exist_ok=True)
out_dir = out_dir + '/fpr_' + fpr
os.makedirs(out_dir, exist_ok=True)
os.makedirs(out_dir + '/' + samplename, exist_ok=True)
logging.info(out_dir)
# logging.info(df.head())
# Get compression opts for pandas
compression_opts = 'gzip'
if LooseVersion(pd.__version__) > '1.0.0':
compression_opts = dict(method='gzip', compresslevel=9)
# read cellranger raw
adata_cellranger_raw = sc.read_10x_mtx(raw_cellranger_mtx,
var_names='gene_symbols',
make_unique=True,
cache=False,
cache_compression=compression_opts)
# First filter out any cells that have 0 total counts
zero_count_cells_cellranger_raw = adata_cellranger_raw.obs_names[np.where(
adata_cellranger_raw.X.sum(axis=1) == 0)[0]]
# sc.pp.filter_cells(adata, min_counts=1, inplace=True) # Minimum number of counts required for a cell to pass filtering.
logging.info(
"_cellranger_raw: Filtering {}/{} cells with 0 counts.".format(
len(zero_count_cells_cellranger_raw), adata_cellranger_raw.n_obs))
adata_cellranger_raw = adata_cellranger_raw[
adata_cellranger_raw.obs_names.difference(
zero_count_cells_cellranger_raw, sort=False)]
sc.pp.calculate_qc_metrics(adata_cellranger_raw, inplace=True)
logging.info('cellranger raw n barcodes(.obs) x cells(.var) .X.shape:')
logging.info(adata_cellranger_raw.X.shape)
logging.info('cellranger raw .obs:')
logging.info(adata_cellranger_raw.obs)
logging.info('cellranger raw .var:')
logging.info(adata_cellranger_raw.var)
df_total_counts = pd.DataFrame(data=adata_cellranger_raw.obs.sort_values(
by=['total_counts'], ascending=False).total_counts)
df_total_counts['barcode_row_number'] = df_total_counts.reset_index(
).index + 1
df_total_counts['barcodes'] = df_total_counts.index
df_total_counts_cellranger_raw = df_total_counts
df_total_counts_cellranger_raw['dataset'] = 'Cellranger Raw'
logging.info(df_total_counts)
# read cellranger filtered
adata_cellranger_filtered = sc.read_10x_mtx(
filtered_cellranger_mtx,
var_names='gene_symbols',
make_unique=True,
cache=False,
cache_compression=compression_opts)
# First filter out any cells that have 0 total counts
zero_count_cells_cellranger_filtered = adata_cellranger_filtered.obs_names[
np.where(adata_cellranger_filtered.X.sum(axis=1) == 0)[0]]
# sc.pp.filter_cells(adata, min_counts=1, inplace=True) # Minimum number of counts required for a cell to pass filtering.
logging.info(
"_cellranger_filtered: Filtering {}/{} cells with 0 counts.".format(
len(zero_count_cells_cellranger_filtered),
adata_cellranger_filtered.n_obs))
adata_cellranger_filtered = adata_cellranger_filtered[
adata_cellranger_filtered.obs_names.difference(
zero_count_cells_cellranger_filtered, sort=False)]
sc.pp.calculate_qc_metrics(adata_cellranger_filtered, inplace=True)
logging.info(
'cellranger filtered n barcodes(.obs) x cells(.var) .X.shape:')
logging.info(adata_cellranger_filtered.X.shape)
logging.info('cellranger filtered .obs:')
logging.info(adata_cellranger_filtered.obs.columns)
logging.info(adata_cellranger_filtered.obs)
logging.info('cellranger filtered .var:')
logging.info(adata_cellranger_filtered.var)
df_total_counts = pd.DataFrame(
data=adata_cellranger_filtered.obs.sort_values(
by=['total_counts'], ascending=False).total_counts)
df_total_counts['barcodes'] = df_total_counts.index
df_total_counts['barcode_row_number'] = df_total_counts.reset_index(
).index + 1
df_total_counts_cellranger_filtered = df_total_counts
df_total_counts_cellranger_filtered['dataset'] = 'Cellranger Filtered'
logging.info(df_total_counts)
# read cellbender output
adata_cellbender = anndata_from_h5(cellbender_unfiltered_h5,
analyzed_barcodes_only=True)
# First filter out any cells that have 0 total counts
zero_count_cells_cellbender_filtered = adata_cellbender.obs_names[np.where(
adata_cellbender.X.sum(axis=1) == 0)[0]]
# sc.pp.filter_cells(adata, min_counts=1, inplace=True) # Minimum number of counts required for a cell to pass filtering.
logging.info(
"_cellbender_filtered: Filtering {}/{} cells with 0 counts.".format(
len(zero_count_cells_cellbender_filtered), adata_cellbender.n_obs))
adata_cellbender = adata_cellbender[adata_cellbender.obs_names.difference(
zero_count_cells_cellbender_filtered, sort=False)]
sc.pp.calculate_qc_metrics(adata_cellbender, inplace=True)
logging.info(
'cellbender cellbender.n barcodes(.obs) x cells(.var) .X.shape:')
logging.info(adata_cellbender.X.shape)
logging.info('cellbender cellbender.obs:')
logging.info(adata_cellbender.obs)
logging.info('cellbender cellbender.var:')
logging.info(adata_cellbender.var)
df_total_counts = pd.DataFrame(data=adata_cellbender.obs.sort_values(
by=['total_counts'], ascending=False).total_counts)
df_total_counts['barcodes'] = df_total_counts.index
df_total_counts['barcode_row_number'] = df_total_counts.reset_index(
).index + 1
df_total_counts_cellbender = df_total_counts
df_total_counts_cellbender['dataset'] = 'Cellbender'
logging.info(df_total_counts)
# df_total_counts_cellranger_filtered.rename(columns={"total_counts": "cellranger_filtered_total_counts"})
df_cellranger_cellbender = pd.merge(
df_total_counts_cellranger_filtered,
df_total_counts_cellbender,
how='outer',
left_index=True,
right_index=True,
suffixes=('_cellranger',
'_cellbender')).sort_values(by=['total_counts_cellbender'],
ascending=False)
logging.info(df_cellranger_cellbender)
df_cellranger_cellbender[['cellranger', 'cellbender']] = np.where(
df_cellranger_cellbender[[
'total_counts_cellranger', 'total_counts_cellbender'
]].isnull(), 0, 1)
#df_cellranger_cellbender.to_csv('df_cellranger_cellbender.csv', index=True, index_label='barcode')
grouped = df_cellranger_cellbender[['cellranger', 'cellbender']].groupby(
["cellranger", "cellbender"]).size().reset_index(name='counts')
logging.info(grouped.columns)
#grouped.to_csv('cellranger_cellbender.csv', index=False)
df_cellranger_cellbender[
'barcode_row_number'] = df_cellranger_cellbender.reset_index(
).index + 1
### plot UMI counts descending order
df_merged = pd.concat([
df_total_counts_cellranger_raw, df_total_counts_cellranger_filtered,
df_total_counts_cellbender
])
#df_merged.to_csv('df_merged.csv', index=True, index_label='barcode')
df_vline = pd.DataFrame(
data={
'x': [int(n_expected_cells),
int(n_total_droplets_included)],
'color': ['expected-cells', 'total-droplets-included']
})
gplt = ggplot(df_merged, aes(x='barcode_row_number', y='total_counts')) \
+ geom_point() \
+ geom_vline(df_vline, aes(xintercept='x', color='color')) \
+ theme_bw() + facet_wrap('dataset') \
+ labs(x='Barcodes (ordered by descending cell total couts)',color='Cellbender input',
y='Cell total counts', title='Cells filtered out by Cellranger or Cellbender') \
+ scale_y_continuous(trans='log10',minor_breaks=0) + scale_x_continuous(trans='log10',minor_breaks=0)
gplt.save(out_dir + '/' + samplename + '/barcode_vs_total_counts.png',
width=12,
height=5,
dpi=300) # dpi=300,
df_cellranger_cellbender_count = grouped # pd.read_csv('cellranger_cellbender.csv')
df = pd.merge(df_merged,
df_cellranger_cellbender[['cellranger', 'cellbender']],
how='left',
left_index=True,
right_index=True)
df = pd.merge(df,
df_cellranger_cellbender_count,
how='left',
left_on=['cellranger', 'cellbender'],
right_on=['cellranger', 'cellbender'])
df["counts"].fillna(df['counts'].isnull().sum(), inplace=True)
df["counts"] = df["counts"].astype(int)
# df.replace({"counts": {""} }, inplace=True)
df["filtered"] = df["cellranger"].astype(
str) + '-' + df["cellbender"].astype(str)
df.replace(
{
"filtered": {
"nan-nan": 'Cellranger Raw only',
"1.0-1.0": "Cellranger Filtered + Cellbender",
"1.0-0.0": "Cellranger Filtered only",
"0.0-1.0": "Cellbender only",
"0.0-0.0": "0.0-0.0"
}
},
inplace=True)
df["filtered"] = df["filtered"] + ', n=' + df["counts"].astype(str)
df['filtered'].value_counts()
df.replace(
{
"dataset": {
"cellbender": "Cellbender output",
"cellranger_raw": "Cellranger Raw output",
"cellranger_filtered": "Cellranger Filtered output"
}
},
inplace=True)
gplt = ggplot(df, aes(x='filtered', y='total_counts', color='filtered')) \
+ geom_boxplot() \
+ theme_bw() \
+ facet_wrap('dataset') \
+ theme(axis_text_x=element_blank()) \
+ scale_y_continuous(trans='log10',minor_breaks=0) \
+ labs(color='n cells in intersection of datasets', x='', y='Cell total counts', title='Total cell counts compared across datasets (facets)')
gplt.save(out_dir + '/' + samplename +
'/boxplots_cellranger_vs_cellbender.png',
width=12,
height=5,
dpi=300) # dpi=300,
# plot difference cellbender filtered vs cellranger filtered for common cells between the 2 datasets
df_cellranger_cellbender = df_cellranger_cellbender[
df_cellranger_cellbender['cellranger'] == 1]
df_cellranger_cellbender = df_cellranger_cellbender[
df_cellranger_cellbender['cellbender'] == 1]
# Subset the datasets to the relevant barcodes.
adata_cellbender_common = adata_cellbender[
df_cellranger_cellbender.index.values]
adata_cellranger_filtered_common = adata_cellranger_filtered[
df_cellranger_cellbender.index.values]
# Put count matrices into 'layers' in anndata for clarity.
adata = adata_cellbender_common
adata.layers['counts_cellbender'] = adata_cellbender_common.X.copy()
adata.layers['counts_raw'] = adata_cellranger_filtered_common.X.copy()
# Get the differences in counts per cell
X_raw_minus_cb = adata.layers['counts_raw'] - adata.layers[
'counts_cellbender']
X_raw_minus_cb = abs(X_raw_minus_cb)
# Get the top most different genes
df_diff_genes = pd.DataFrame(data=adata.var.gene_symbols.values)
df_diff_genes['ensembl_id'] = adata.var.index
df_diff_genes['gene_symbol'] = adata.var.gene_symbols.values
df_diff_genes['dif_across_cells'] = np.asarray(
X_raw_minus_cb.sum(axis=0)).reshape(-1)
df_diff_genes = df_diff_genes.sort_values('dif_across_cells',
ascending=False).head(n=100)
#df_diff_genes.to_csv('df_diff_genes.csv', index=True)
top_genes = df_diff_genes['ensembl_id']
top_genes_symbols = df_diff_genes['gene_symbol']
logging.info('top_genes:')
logging.info(top_genes)
logging.info(adata_cellbender_common.var.index)
adata_cellbender_common = adata_cellbender[
df_cellranger_cellbender.index.values, top_genes].to_df()
adata_cellbender_common['barcode'] = adata_cellbender_common.index
adata_cellbender_common = pd.melt(adata_cellbender_common,
ignore_index=True,
id_vars=['barcode'],
var_name='ensembl_id',
value_name='count')
adata_cellbender_common = pd.merge(
adata_cellbender_common,
df_diff_genes[['ensembl_id', 'gene_symbol']],
how='left',
left_on='ensembl_id',
right_on='ensembl_id')
adata_cellbender_common = adata_cellbender_common.sort_values(
by=['barcode', 'ensembl_id'], ascending=False)
adata_cellbender_common['dataset'] = 'Cellbender'
#adata_cellbender_common.to_csv('adata_cellbender_common.csv', index=True)
logging.info(adata_cellranger_filtered.var.index)
adata_cellranger_filtered_common = adata_cellranger_filtered[
df_cellranger_cellbender.index.values, top_genes_symbols].to_df()
adata_cellranger_filtered_common[
'barcode'] = adata_cellranger_filtered_common.index
adata_cellranger_filtered_common = pd.melt(
adata_cellranger_filtered_common,
ignore_index=True,
id_vars=['barcode'],
var_name='gene_symbol',
value_name='count')
adata_cellranger_filtered_common = pd.merge(
adata_cellranger_filtered_common,
df_diff_genes[['ensembl_id', 'gene_symbol']],
how='left',
left_on='gene_symbol',
right_on='gene_symbol')
adata_cellranger_filtered_common['dataset'] = 'Cellranger Filtered'
adata_cellranger_filtered_common = adata_cellranger_filtered_common.sort_values(
by=['barcode', 'ensembl_id'], ascending=False)
adata_cellranger_filtered_common = adata_cellranger_filtered_common[
adata_cellbender_common.columns]
#adata_cellranger_filtered_common.to_csv('adata_cellranger_filtered_common.csv', index=True)
logging.info(adata_cellranger_raw.var.index)
adata_cellranger_raw_common = adata_cellranger_raw[
df_cellranger_cellbender.index.values, top_genes_symbols].to_df()
adata_cellranger_raw_common['barcode'] = adata_cellranger_raw_common.index
adata_cellranger_raw_common = pd.melt(adata_cellranger_raw_common,
ignore_index=True,
id_vars=['barcode'],
var_name='gene_symbol',
value_name='count')
adata_cellranger_raw_common = pd.merge(
adata_cellranger_raw_common,
df_diff_genes[['ensembl_id', 'gene_symbol']],
how='left',
left_on='gene_symbol',
right_on='gene_symbol')
adata_cellranger_raw_common['dataset'] = 'Cellranger Raw'
adata_cellranger_raw_common = adata_cellranger_raw_common.sort_values(
by=['barcode', 'ensembl_id'], ascending=False)
adata_cellranger_raw_common = adata_cellranger_raw_common[
adata_cellbender_common.columns]
#adata_cellranger_raw_common.to_csv('adata_cellranger_raw_common.csv', index=True)
logging.info(adata_cellranger_raw_common['gene_symbol'] ==
adata_cellbender_common['gene_symbol'])
logging.info(adata_cellranger_raw_common['ensembl_id'] ==
adata_cellbender_common['ensembl_id'])
adata_filtered_cellbender_diff = adata_cellbender_common.copy()
adata_filtered_cellbender_diff['count'] = adata_cellranger_filtered_common[
'count'] - adata_cellbender_common['count']
adata_filtered_cellbender_diff[
'dataset'] = 'Cellranger Filtered - Cellbender'
adata_raw_cellbender_diff = adata_cellbender_common.copy()
adata_raw_cellbender_diff['count'] = adata_cellranger_raw_common[
'count'] - adata_cellbender_common['count']
adata_raw_cellbender_diff['dataset'] = 'Cellranger Raw - Cellbender'
df_merged = pd.concat([
adata_cellbender_common, adata_cellranger_filtered_common,
adata_cellranger_raw_common, adata_filtered_cellbender_diff,
adata_raw_cellbender_diff
],
ignore_index=True)
gplt = ggplot(df_merged, aes(x='gene_symbol',y='count')) \
+ geom_boxplot() \
+ theme_bw() \
+ theme(axis_text_x = element_text(angle = 90, hjust = 1, size= 6)) \
+ facet_wrap('dataset', scales = 'free', ncol = 1) \
+ labs(x='Genes (top 100 Genes most different between Cellranger Filtered counts and Cellbender filtered counts)', y='Cell total counts', title='Total cell counts compared across most different genes (x-axis) and datasets (facets)')
gplt.save(out_dir + '/' + samplename +
'/boxplot_topgenes_cellranger_vs_cellbender.png',
width=10,
height=20,
dpi=300) # dpi=300,
logging.info('script done.')
def test_facet_wrap_two_vars():
p = g + facet_wrap('~var1+var2')
p2 = g + facet_wrap('~class+var2') # python keyword in formula
assert p == 'facet_wrap_two_vars'
assert p2 == 'facet_wrap_two_vars'
# Add label for input or simulated dataset
input_data_UMAPencoded_df['dataset'] = 'original'
simulated_data_UMAPencoded_df['dataset'] = 'simulated'
# Concatenate input and simulated dataframes together
combined_data_df = pd.concat(
[input_data_UMAPencoded_df, simulated_data_UMAPencoded_df])
# Plot sequentially
#backgrd_data = combined_data_df[combined_data_df['experiment_id'] == 'Not selected']
#select_data = combined_data_df[combined_data_df['experiment_id'] != 'Not selected']
# Plot
ggplot(combined_data_df, aes(x='1', y='2')) + geom_point(
aes(color='experiment_id'),
alpha=0.3) + facet_wrap('~dataset') + xlab('UMAP 1') + ylab(
'UMAP 2') + ggtitle('UMAP of original and simulated data (gene space)')
#+ xlim(3,12) \
#+ ylim(-7,10) \
#+ scale_colour_manual(values=["blue", "purple", "orange", "red", "magenta", "lightgrey"]) \
# In[12]:
# Overlay original and simulated data
ggplot(combined_data_df, aes(x='1', y='2')) + geom_point(
aes(color='dataset'), alpha=0.3) + scale_colour_manual(
values=["grey", "blue"]) + xlab('UMAP 1') + ylab('UMAP 2') + ggtitle(
'UMAP of original and simulated data (gene space)')
# ## Visualize simulated data (gene space) projected into PCA space
def plot_abs_dataframe(self, df: pd.DataFrame) -> p9.ggplot:
facets = []
n_per_facet = {}
print(df)
for col in df.columns:
try:
n_values = df[col].nunique()
if n_values == 1 and col not in [
"TIME_PERIOD",
"value",
"Measure",
"OBS_COMMENT",
]:
self.fixed_datapoints.add(f"{col}={df.at[0, col]}")
elif n_values > 1 and col not in [
"value",
"TIME_PERIOD",
"OBS_COMMENT",
]:
facets.append(col)
n_per_facet[col] = n_values
except:
print(f"Ignoring unusable column: {col}")
continue
extra_args = {}
need_shape = False
if len(facets) > 2:
# can only use two variables as plotting facets, third value will be used as a group on each plot
# any more facets is not supported at this stage
sorted_facets = sorted(n_per_facet.keys(),
key=lambda k: n_per_facet[k])
# print(n_per_facet)
# print(sorted_facets)
facets = sorted_facets[-2:]
extra_args.update({
"group": sorted_facets[0],
"color": facets[0],
"shape": sorted_facets[0],
})
need_shape = True
print(f"Using {facets} as facets, {extra_args} as series")
else:
if len(facets) > 0:
extra_args.update({"color": facets[0]})
# compute figure size to give enough room for each plot
mult = 1
for facet in facets:
mult *= n_per_facet[facet]
mult /= len(facets)
nrow = int(mult + 1)
# facet column names must not have spaces in them as this is not permitted by plotnine facet formulas
if len(facets) > 0:
new_facets = []
for f in facets:
if " " in f:
new_name = f.replace(" ", "_")
df = df.rename(columns={f: new_name})
new_facets.append(new_name)
else:
new_facets.append(f)
facets = new_facets
if "color" in extra_args:
extra_args.update({"color": facets[0]})
print(f"Renamed facet columns due to whitespace: {facets}")
plot = p9.ggplot(df, p9.aes(x="TIME_PERIOD", y="value", **
extra_args)) + p9.geom_point(size=3)
if len(facets) > 0 and len(facets) <= 2:
facet_str = "~" + " + ".join(facets[:2])
print(f"Using facet formula: {facet_str}")
plot += p9.facet_wrap(facet_str, ncol=len(facets), scales="free_y")
plot_theme = {
"figure_size": (12, int(nrow * 1.5)),
}
if (len(facets) == 2
): # two columns of plots? if so, make sure space for axis labels
plot_theme.update({"subplots_adjust": {"wspace": 0.2}})
if need_shape:
plot += p9.scale_shape(guide="legend")
plot += p9.guides(
colour=False
) # colour legend is not useful since it is included in the facet title
plot_theme.update({"legend_position": "right"})
return user_theme(plot, **plot_theme)
def test_label_value():
p = g + facet_wrap('~ gear', labeller='label_value')
assert p == 'label_value'
def test_facet_wrap_one_var():
p = g + facet_wrap('~var1')
assert p == 'facet_wrap_one_var'
def test_label_context():
p = g + facet_wrap('~ gear', labeller='label_context')
assert p == 'label_context'
def test_facet_wrap_expression():
p = g + facet_wrap('pd.cut(var1, (0, 2, 4), include_lowest=True)')
assert p == 'facet_wrap_expression'
def test_labeller_cols_both_wrap():
p = g + facet_wrap('~ gear + am', labeller=labeller_cols_both)
assert p == 'labeller_cols_both_wrap'
def test_facet_wrap_two_vars():
p = g + facet_wrap('~var1+var2')
assert p == 'facet_wrap_two_vars'
def test_aslabeller_dict_0tag():
func = as_labeller({'0': '<tag>0</tag>'})
p = g + facet_wrap('~ gear + am', labeller=func)
assert p == 'aslabeller_dict_0tag'
def test_facet_wrap_label_both():
p = g + facet_wrap('~var1+var2', labeller='label_both')
assert p == 'facet_wrap_label_both'
def plot_char_percent_vs_accuracy_smooth(self, expo=False, no_models=False, columns=False):
if self.y_max is not None:
limits = [0, float(self.y_max)]
eprint(f'Setting limits to: {limits}')
else:
limits = [0, 1]
if expo:
if os.path.exists('data/external/all_human_gameplay.json') and not self.no_humans:
with open('data/external/all_human_gameplay.json') as f:
all_gameplay = json.load(f)
frames = []
for event, name in [('parents', 'Intermediate'), ('maryland', 'Expert'), ('live', 'National')]:
if self.merge_humans:
name = 'Human'
gameplay = all_gameplay[event]
if event != 'live':
control_correct_positions = gameplay['control_correct_positions']
control_wrong_positions = gameplay['control_wrong_positions']
control_positions = control_correct_positions + control_wrong_positions
control_positions = np.array(control_positions)
control_result = np.array(len(control_correct_positions) * [1] + len(control_wrong_positions) * [0])
argsort_control = np.argsort(control_positions)
control_x = control_positions[argsort_control]
control_sorted_result = control_result[argsort_control]
control_y = control_sorted_result.cumsum() / control_sorted_result.shape[0]
control_df = pd.DataFrame({'correct': control_y, 'char_percent': control_x})
control_df['Dataset'] = 'Regular Test'
control_df['Guessing_Model'] = f' {name}'
frames.append(control_df)
adv_correct_positions = gameplay['adv_correct_positions']
adv_wrong_positions = gameplay['adv_wrong_positions']
adv_positions = adv_correct_positions + adv_wrong_positions
adv_positions = np.array(adv_positions)
adv_result = np.array(len(adv_correct_positions) * [1] + len(adv_wrong_positions) * [0])
argsort_adv = np.argsort(adv_positions)
adv_x = adv_positions[argsort_adv]
adv_sorted_result = adv_result[argsort_adv]
adv_y = adv_sorted_result.cumsum() / adv_sorted_result.shape[0]
adv_df = pd.DataFrame({'correct': adv_y, 'char_percent': adv_x})
adv_df['Dataset'] = 'IR Adversarial'
adv_df['Guessing_Model'] = f' {name}'
frames.append(adv_df)
if len(gameplay['advneural_correct_positions']) > 0:
adv_correct_positions = gameplay['advneural_correct_positions']
adv_wrong_positions = gameplay['advneural_wrong_positions']
adv_positions = adv_correct_positions + adv_wrong_positions
adv_positions = np.array(adv_positions)
adv_result = np.array(len(adv_correct_positions) * [1] + len(adv_wrong_positions) * [0])
argsort_adv = np.argsort(adv_positions)
adv_x = adv_positions[argsort_adv]
adv_sorted_result = adv_result[argsort_adv]
adv_y = adv_sorted_result.cumsum() / adv_sorted_result.shape[0]
adv_df = pd.DataFrame({'correct': adv_y, 'char_percent': adv_x})
adv_df['Dataset'] = 'RNN Adversarial'
adv_df['Guessing_Model'] = f' {name}'
frames.append(adv_df)
human_df = pd.concat(frames)
human_vals = sort_humans(list(human_df['Guessing_Model'].unique()))
human_dtype = CategoricalDtype(human_vals, ordered=True)
human_df['Guessing_Model'] = human_df['Guessing_Model'].astype(human_dtype)
dataset_dtype = CategoricalDtype(['Regular Test', 'IR Adversarial', 'RNN Adversarial'], ordered=True)
human_df['Dataset'] = human_df['Dataset'].astype(dataset_dtype)
if no_models:
p = ggplot(human_df) + geom_point(shape='.')
else:
df = self.char_plot_df
if 1 not in self.rounds:
df = df[df['Dataset'] != 'Round 1 - IR Adversarial']
if 2 not in self.rounds:
df = df[df['Dataset'] != 'Round 2 - IR Adversarial']
df = df[df['Dataset'] != 'Round 2 - RNN Adversarial']
p = ggplot(df)
if self.save_df is not None:
eprint(f'Saving df to: {self.save_df}')
df.to_json(self.save_df)
if os.path.exists('data/external/all_human_gameplay.json') and not self.no_humans:
eprint('Loading human data')
p = p + geom_line(data=human_df)
if columns:
facet_conf = facet_wrap('Guessing_Model', ncol=1)
else:
facet_conf = facet_wrap('Guessing_Model', nrow=1)
if not no_models:
if self.mvg_avg_char:
chart = stat_smooth(method='mavg', se=False, method_args={'window': 400})
else:
chart = stat_summary_bin(fun_data=mean_no_se, bins=20, shape='.', linetype='None', size=0.5)
else:
chart = None
p = (
p + facet_conf
+ aes(x='char_percent', y='correct', color='Dataset')
)
if chart is not None:
p += chart
p = (
p
+ scale_y_continuous(breaks=np.linspace(0, 1, 6))
+ scale_x_continuous(breaks=[0, .5, 1])
+ coord_cartesian(ylim=limits)
+ xlab('Percent of Question Revealed')
+ ylab('Accuracy')
+ theme(
#legend_position='top', legend_box_margin=0, legend_title=element_blank(),
strip_text_x=element_text(margin={'t': 6, 'b': 6, 'l': 1, 'r': 5})
)
+ scale_color_manual(values=['#FF3333', '#66CC00', '#3333FF', '#FFFF33'], name='Questions')
)
if self.title != '':
p += ggtitle(self.title)
return p
else:
if self.save_df is not None:
eprint(f'Saving df to: {self.save_df}')
df.to_json(self.save_df)
return (
ggplot(self.char_plot_df)
+ aes(x='char_percent', y='correct', color='Guessing_Model')
+ stat_smooth(method='mavg', se=False, method_args={'window': 500})
+ scale_y_continuous(breaks=np.linspace(0, 1, 6))
+ coord_cartesian(ylim=limits)
)
def test_facet_wrap_not_as_table():
p = g + facet_wrap('~var1', as_table=False)
assert p == 'facet_wrap_not_as_table'
def plot_n_train_vs_accuracy(self):
return (
ggplot(self.combined_df) + facet_wrap('seen')
+ aes(x='n_train', fill='Outcome')
+ geom_histogram(binwidth=1)
)
def test_facet_wrap_direction_v():
p = g + facet_wrap('~var1', dir='v')
assert p == 'facet_wrap_direction_v'
def test_facet_wrap_one_var():
p = g + facet_wrap('~var1')
p2 = g + facet_wrap('~class') # python keyword in formula
assert p == 'facet_wrap_one_var'
assert p2 == 'facet_wrap_one_var'
def test_facet_wrap_not_as_table_direction_v():
p = g + facet_wrap('~var1', as_table=False, dir='v')
assert p == 'facet_wrap_not_as_table_direction_v'
def test_facet_wrap_two_vars():
p = g + facet_wrap('~var1+var2')
p2 = g + facet_wrap('~class+var2') # python keyword in formula
assert p == 'facet_wrap_two_vars'
assert p2 == 'facet_wrap_two_vars'
dfwords['Date'] = X_train_all.Date.values
dfwords.Speaker = [to_speaker_dict[spkr] for spkr in dfwords.Speaker]
pres_Nelson = (dfwords.query(f"Speaker in {first_presidency}").groupby(
[dfwords['Date'].map(lambda x: x.year),
'Speaker']).mean()[interesting_words].reset_index())
pres_Nelson['combined'] = [str(dt)
for dt in pres_Nelson.Date] + pres_Nelson.Speaker
pres_Nelson = (pres_Nelson.drop(
columns=['Date', 'Speaker']).set_index('combined').unstack().reset_index())
pres_Nelson['Date'] = [int(comb[:4]) for comb in pres_Nelson.combined]
pres_Nelson['Speaker'] = [comb[4:] for comb in pres_Nelson.combined]
pres_Nelson = pres_Nelson.drop(columns='combined')
pres_Nelson.columns = ['Word', 'Mean TF-IDF Score', 'Date', 'Speaker']
(ggplot(pres_Nelson, aes(x='Date', y='Mean TF-IDF Score', color='Word')) +
geom_line() + facet_wrap('Speaker', scales='free', nrow=3) +
labs(title='Word Usage Change over Time in First Presidency'))
int_words = (dfwords.groupby(dfwords['Date'].map(lambda x: x.year)).mean()
[interesting_words].unstack().reset_index())
int_words.columns = ['Word', 'Date', 'Mean TF-IDF Score']
(ggplot(int_words, aes(x='Date', y='Mean TF-IDF Score', color='Word')) +
geom_line() +
labs(title='Word Usage Change over Time in First Presidency and the 12'))
missionary_temple = (dfwords.groupby(
dfwords['Date'].map(lambda x: x.year)).mean()[[
'missionary work', 'family history'
]].unstack().reset_index())
missionary_temple.columns = ['Word', 'Date', 'Mean TF-IDF Score']
(ggplot(missionary_temple, aes(x='Date', y='Mean TF-IDF Score',
def test_facet_wrap_not_as_table():
p = g + facet_wrap('~var1', as_table=False)
assert p == 'facet_wrap_not_as_table'
def plot_char_percent_vs_accuracy_smooth(self, expo=False):
if expo:
p = (ggplot(self.char_plot_df) +
facet_wrap('Guessing_Model', nrow=1) +
aes(x='char_percent', y='correct', color='Dataset') +
stat_smooth(
method='mavg', se=False, method_args={'window': 200}) +
scale_y_continuous(breaks=np.linspace(0, 1, 11)) +
scale_x_continuous(breaks=[0, .5, 1]) +
xlab('Percent of Question Revealed') + ylab('Accuracy') +
theme(legend_position='top'))
if os.path.exists('data/external/human_gameplay.json'):
with open('data/external/human_gameplay.json') as f:
gameplay = json.load(f)
control_correct_positions = gameplay[
'control_correct_positions']
control_wrong_positions = gameplay[
'control_wrong_positions']
control_positions = control_correct_positions + control_wrong_positions
control_positions = np.array(control_positions)
control_result = np.array(
len(control_correct_positions) * [1] +
len(control_wrong_positions) * [0])
argsort_control = np.argsort(control_positions)
control_x = control_positions[argsort_control]
control_sorted_result = control_result[argsort_control]
control_y = control_sorted_result.cumsum(
) / control_sorted_result.shape[0]
control_df = pd.DataFrame({
'correct': control_y,
'char_percent': control_x
})
control_df['Dataset'] = 'Test Questions'
control_df['Guessing_Model'] = ' Human'
adv_correct_positions = gameplay['adv_correct_positions']
adv_wrong_positions = gameplay['adv_wrong_positions']
adv_positions = adv_correct_positions + adv_wrong_positions
adv_positions = np.array(control_positions)
adv_result = np.array(
len(adv_correct_positions) * [1] +
len(adv_wrong_positions) * [0])
argsort_adv = np.argsort(adv_positions)
adv_x = adv_positions[argsort_adv]
adv_sorted_result = adv_result[argsort_adv]
adv_y = adv_sorted_result.cumsum(
) / adv_sorted_result.shape[0]
adv_df = pd.DataFrame({
'correct': adv_y,
'char_percent': adv_x
})
adv_df['Dataset'] = 'Challenge Questions'
adv_df['Guessing_Model'] = ' Human'
human_df = pd.concat([control_df, adv_df])
p = p + (geom_line(data=human_df))
return p
else:
return (
ggplot(self.char_plot_df) +
aes(x='char_percent', y='correct', color='Guessing_Model') +
stat_smooth(
method='mavg', se=False, method_args={'window': 500}) +
scale_y_continuous(breaks=np.linspace(0, 1, 21)))
def test_facet_wrap_not_as_table_direction_v():
p = g + facet_wrap('~var1', as_table=False, dir='v')
assert p == 'facet_wrap_not_as_table_direction_v'
def plot_n_train_vs_accuracy(self):
return (ggplot(self.combined_df) + facet_wrap('seen') +
aes(x='n_train', fill='Outcome') + geom_histogram(binwidth=1))
#-----------------------------
#libraries https://pythonplot.com/#bar-count
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
#pip install plotnine #similar to ggplots
#https://plotnine.readthedocs.io/en/stable/index.html
import plotnine #ggplot type
from plotnine import ggplot, geom_point, aes, stat_smooth, facet_wrap
help(plotnine.facet_wrap)
from plotnine.data import mtcars
(ggplot(mtcars, aes('wt', 'mpg', color='factor(gear)')) + geom_point() +
stat_smooth(method='lm') + facet_wrap('~gear'))
from plotnine import *
(ggplot(mtcars, aes('factor(cyl)', fill='factor(am)')) +
geom_bar(position='fill'))
(ggplot(mtcars, aes('factor(cyl)', fill='factor(am)')) +
geom_bar(position='fill') +
geom_text(aes(label='stat(count)'), stat='count', position='fill'))
(ggplot(mpg) + aes(x='manufacturer') + geom_bar(size=20) + coord_flip() +
labs(y='Count', x='Manufacturer', title='Number of Cars by Make'))
#https://plotnine.readthedocs.io/en/stable/tutorials/miscellaneous-order-plot-series.html
from pydataset import data
data()
def control_list(in_file=None,
out_dir=None,
reference_gene_file=None,
log2=False,
page_width=None,
page_height=None,
user_img_file=None,
page_format=None,
pseudo_count=1,
set_colors=None,
dpi=300,
rug=False,
jitter=False,
skip_first=False):
# -------------------------------------------------------------------------
#
# Check in_file content
#
# -------------------------------------------------------------------------
for p, line in enumerate(in_file):
line = chomp(line)
line = line.split("\t")
if len(line) > 2:
message("Need a two columns file.",
type="ERROR")
if skip_first:
if p == 0:
continue
try:
fl = float(line[1])
except ValueError:
msg = "It seems that column 2 of input file"
msg += " contains non numeric values. "
msg += "Check that no header is present and that "
msg += "columns are ordered properly. "
msg += "Or use '--skip-first'. "
message(msg, type="ERROR")
if log2:
fl = fl + pseudo_count
if fl <= 0:
message("Can not log transform negative/zero values. Add a pseudo-count.",
type="ERROR")
# -------------------------------------------------------------------------
#
# Check colors
#
# -------------------------------------------------------------------------
set_colors = set_colors.split(",")
if len(set_colors) != 2:
message("Need two colors. Please fix.", type="ERROR")
mcolors_name = mcolors.cnames
for i in set_colors:
if i not in mcolors_name:
if not is_hex_color(i):
message(i + " is not a valid color. Please fix.", type="ERROR")
# -------------------------------------------------------------------------
#
# Preparing output files
#
# -------------------------------------------------------------------------
# Preparing pdf file name
file_out_list = make_outdir_and_file(out_dir, ["control_list.txt",
"reference_list.txt",
"diagnostic_diagrams." + page_format],
force=True)
control_file, reference_file_out, img_file = file_out_list
if user_img_file is not None:
os.unlink(img_file.name)
img_file = user_img_file
if not img_file.name.endswith(page_format):
msg = "Image format should be: {f}. Please fix.".format(f=page_format)
message(msg, type="ERROR")
test_path = os.path.abspath(img_file.name)
test_path = os.path.dirname(test_path)
if not os.path.exists(test_path):
os.makedirs(test_path)
# -------------------------------------------------------------------------
#
# Read the reference list
#
# -------------------------------------------------------------------------
try:
reference_genes = pd.read_csv(reference_gene_file.name, sep="\t", header=None)
except pd.errors.EmptyDataError:
message("No genes in --reference-gene-file.", type="ERROR")
reference_genes.rename(columns={reference_genes.columns.values[0]: 'gene'}, inplace=True)
# -------------------------------------------------------------------------
#
# Delete duplicates
#
# -------------------------------------------------------------------------
before = len(reference_genes)
reference_genes = reference_genes.drop_duplicates(['gene'])
after = len(reference_genes)
msg = "%d duplicate lines have been deleted in reference file."
message(msg % (before - after))
# -------------------------------------------------------------------------
#
# Read expression data and add the pseudo_count
#
# -------------------------------------------------------------------------
if skip_first:
exp_data = pd.read_csv(in_file.name, sep="\t",
header=None, index_col=None,
skiprows=[0], names=['exprs'])
else:
exp_data = pd.read_csv(in_file.name, sep="\t", names=['exprs'], index_col=0)
exp_data.exprs = exp_data.exprs.values + pseudo_count
# -------------------------------------------------------------------------
#
# log transformation
#
# -------------------------------------------------------------------------
ylabel = 'Expression'
if log2:
if len(exp_data.exprs.values[exp_data.exprs.values == 0]):
message("Can't use log transformation on zero or negative values. Use -p.",
type="ERROR")
else:
exp_data.exprs = np.log2(exp_data.exprs.values)
ylabel = 'log2(Expression)'
# -------------------------------------------------------------------------
#
# Are reference gene found in control list
#
# -------------------------------------------------------------------------
# Sort in increasing order
exp_data = exp_data.sort_values('exprs')
# Vector with positions indicating which in the
# expression data list are found in reference_gene
reference_genes_found = [x for x in reference_genes['gene'] if x in exp_data.index]
msg = "Found %d genes of the reference in the provided signal file" % len(reference_genes_found)
message(msg)
not_found = [x for x in reference_genes['gene'] if x not in exp_data.index]
if len(not_found):
if len(not_found) == len(reference_genes):
message("Genes from reference file where not found in signal file (n=%d)." % len(not_found), type="ERROR")
else:
message("List of reference genes not found :%s" % not_found)
else:
message("All reference genes were found.")
# -------------------------------------------------------------------------
#
# Search for genes with matched signal
#
# -------------------------------------------------------------------------
exp_data_save = exp_data.copy()
control_list = list()
nb_candidate_left = exp_data.shape[0] - len(reference_genes_found)
message("Searching for genes with matched signal.")
if nb_candidate_left < len(reference_genes_found):
message("Not enough element to perform selection. Exiting", type="ERROR")
for i in reference_genes_found:
not_candidates = reference_genes_found + control_list
not_candidates = list(set(not_candidates))
diff = abs(exp_data.loc[i] - exp_data)
control_list.extend(diff.loc[np.setdiff1d(diff.index, not_candidates)].idxmin(axis=0, skipna=True).tolist())
# -------------------------------------------------------------------------
#
# Prepare a dataframe for plotting
#
# -------------------------------------------------------------------------
message("Preparing a dataframe for plotting.")
reference = exp_data_save.loc[reference_genes_found].sort_values('exprs')
reference = reference.assign(genesets=['Reference'] * reference.shape[0])
control = exp_data_save.loc[control_list].sort_values('exprs')
control = control.assign(genesets=['Control'] * control.shape[0])
data = pd.concat([reference, control])
data['sets'] = pd.Series(['sets' for x in data.index.tolist()], index=data.index)
data['genesets'] = Categorical(data['genesets'])
# -------------------------------------------------------------------------
#
# Diagnostic plots
#
# -------------------------------------------------------------------------
p = ggplot(data, aes(x='sets', y='exprs', fill='genesets'))
p += scale_fill_manual(values=dict(zip(['Reference', 'Control'], set_colors)))
p += geom_violin(color=None)
p += xlab('Gene sets') + ylab(ylabel)
p += facet_wrap('~genesets')
if rug:
p += geom_rug()
if jitter:
p += geom_jitter()
p += theme_bw()
p += theme(axis_text_x=element_blank())
# -------------------------------------------------------------------------
# Turn warning off. Both pandas and plotnine use warnings for deprecated
# functions. I need to turn they off although I'm not really satisfied with
# this solution...
# -------------------------------------------------------------------------
def fxn():
warnings.warn("deprecated", DeprecationWarning)
# -------------------------------------------------------------------------
#
# Saving
#
# -------------------------------------------------------------------------
with warnings.catch_warnings():
warnings.simplefilter("ignore")
fxn()
message("Saving diagram to file : " + img_file.name)
message("Be patient. This may be long for large datasets.")
try:
p.save(filename=img_file.name, width=page_width, height=page_height, dpi=dpi, limitsize=False)
except PlotnineError as err:
message("Plotnine message: " + err.message)
message("Plotnine encountered an error.", type="ERROR")
# -------------------------------------------------------------------------
#
# write results
#
# -------------------------------------------------------------------------
exp_data_save.loc[reference_genes_found].sort_values('exprs').to_csv(reference_file_out.name, sep="\t")
exp_data_save.loc[control_list].sort_values('exprs').to_csv(control_file.name, sep="\t")
('probability', pipeline.predict_proba(X)[:, 1])
])
predict_df = predict_df.append(df)
predict_df['probability_str'] = predict_df['probability'].apply('{:.1%}'.format)
# In[27]:
# Top predictions amongst negatives (potential hidden responders to a targeted cancer therapy)
(predict_df
.sort_values('decision_function', ascending=False)
.query("status == 0 and feature_set == 'full'")
.head(10)
)
# In[28]:
predict_df['status_'] = predict_df['status'].map(
lambda x: 'negative' if x == 0 else 'positive')
(gg.ggplot(predict_df, gg.aes(x='probability',
fill='status_'))
+ gg.geom_density(alpha=0.6)
+ gg.facet_wrap('~feature_set', ncol=1)
+ gg.labs(x='probability', y='density')
+ gg.guides(fill=gg.guide_legend(title=""))
+ theme_cognoma())
def plot(df: 'DataFrame',
group_colname: str = None,
time_colname: str = None,
max_num_groups: int = 1,
split_dt: Optional[np.datetime64] = None,
**kwargs) -> 'DataFrame':
"""
:param df: The output of `.to_dataframe()`.
:param group_colname: The name of the group-column.
:param time_colname: The name of the time-column.
:param max_num_groups: Max. number of groups to plot; if the number of groups in the dataframe is greater than
this, a random subset will be taken.
:param split_dt: If supplied, will draw a vertical line at this date (useful for showing pre/post validation).
:param kwargs: Further keyword arguments to pass to `plotnine.theme` (e.g. `figure_size=(x,y)`)
:return: A plot of the predicted and actual values.
"""
from plotnine import (
ggplot, aes, geom_line, geom_ribbon, facet_grid, facet_wrap, theme_bw, theme, ylab, geom_vline
)
is_components = ('process' in df.columns and 'state_element' in df.columns)
if group_colname is None:
group_colname = 'group'
if group_colname not in df.columns:
raise TypeError("Please specify group_colname")
if time_colname is None:
time_colname = 'time'
if 'time' not in df.columns:
raise TypeError("Please specify time_colname")
df = df.copy()
if df[group_colname].nunique() > max_num_groups:
subset_groups = df[group_colname].drop_duplicates().sample(max_num_groups).tolist()
if len(subset_groups) < df[group_colname].nunique():
print("Subsetting to groups: {}".format(subset_groups))
df = df.loc[df[group_colname].isin(subset_groups), :]
num_groups = df[group_colname].nunique()
aes_kwargs = {'x': time_colname}
if is_components:
aes_kwargs['group'] = 'state_element'
plot = (
ggplot(df, aes(**aes_kwargs)) +
geom_line(aes(y='mean'), color='#4C6FE7', size=1.5, alpha=.75) +
geom_ribbon(aes(ymin='lower', ymax='upper'), color=None, alpha=.25) +
ylab("")
)
if is_components:
num_processes = df['process'].nunique()
if num_groups > 1 and num_processes > 1:
raise ValueError("Cannot plot components for > 1 group and > 1 processes.")
elif num_groups == 1:
plot = plot + facet_wrap(f"~ measure + process", scales='free_y', labeller='label_both')
if 'figure_size' not in kwargs:
from plotnine.facets.facet_wrap import n2mfrow
nrow, _ = n2mfrow(len(df[['process', 'measure']].drop_duplicates().index))
kwargs['figure_size'] = (12, nrow * 2.5)
else:
plot = plot + facet_grid(f"{group_colname} ~ measure", scales='free_y', labeller='label_both')
if 'figure_size' not in kwargs:
kwargs['figure_size'] = (12, num_groups * 2.5)
if (df.groupby('measure')['process'].nunique() <= 1).all():
plot = plot + geom_line(aes(y='mean', color='state_element'), size=1.5)
else:
if 'actual' in df.columns:
plot = plot + geom_line(aes(y='actual'))
if num_groups > 1:
plot = plot + facet_grid(f"{group_colname} ~ measure", scales='free_y', labeller='label_both')
else:
plot = plot + facet_wrap("~measure", scales='free_y', labeller='label_both')
if 'figure_size' not in kwargs:
kwargs['figure_size'] = (12, 5)
if split_dt:
plot = plot + geom_vline(xintercept=np.datetime64(split_dt), linetype='dashed')
return plot + theme_bw() + theme(**kwargs)
x['k'],
x['resubAccuracy'],
x['testAccuracy'])
for x in repeatedKnnResults],
columns = ['p',
'k',
'resubAccuracy',
'testAccuracy'])
ggdata = pd.concat(
[DataFrame({'p' : knnResultsSimplified.p,
'k' : knnResultsSimplified.k.apply(int),
'type' : 'resub',
'Accuracy' : knnResultsSimplified.resubAccuracy}),
DataFrame({'p' : knnResultsSimplified.p,
'k' : knnResultsSimplified.k.apply(int),
'type' : 'test',
'Accuracy' : knnResultsSimplified.testAccuracy})],
axis = 0
)
plt.close()
ggo = gg.ggplot(ggdata, gg.aes(x='p', y='Accuracy',
color='type', group='type', linetype='type'))
ggo += gg.facet_wrap('~ k')
ggo += gg.scale_x_log10()
ggo += gg.geom_point(alpha=0.6)
ggo += gg.stat_smooth()
ggo += gg.theme_bw()
print(ggo)
p9.ggplot(data=df_p4k, mapping=p9.aes(x="score")) + p9.geom_density()
# In[6]:
df_p4k_sum = df_p4k.groupby("genre").mean()
df_p4k_sum
# In[25]:
df_p4k_best = df_p4k[df_p4k['best'] == 1]
p9.ggplot(data=df_p4k_best, mapping=p9.aes(x="score")) + p9.geom_density()
# In[26]:
p9.ggplot(data=df_p4k, mapping=p9.aes(
x="score")) + p9.facet_wrap("~genre") + p9.geom_density()
# Word Clouds of the review
# In[23]:
wc_text = " ".join(df_p4k['review'].head(10).as_matrix().astype('str'))
wc_text = " ".join(stripNonAlphaNum(wc_text))
p4k_wordcloud = WordCloud().generate(wc_text)
wordcloud = WordCloud(max_font_size=40).generate(wc_text)
plt.figure()
plt.imshow(wordcloud, interpolation="bilinear")
# Split word clouds by genre
# In[8]:
