def cum_regret_plot(experiment_name, data_path=_DEFAULT_DATA_PATH):
"""Simple plot of average instantaneous regret by agent, per timestep.
Args:
experiment_name: string = name of experiment config.
data_path: string = where to look for the files.
Returns:
https://web.stanford.edu/~bvr/pubs/TS_Tutorial.pdf
"""
df = load_data(experiment_name, data_path)
plt_df = (df.groupby(['t', 'agent']).agg({
'cum_regret': [np.mean, lower_interval, upper_interval]
}).reset_index())
plt_df.columns = ['_'.join(i) for i in plt_df.columns.values]
p = (gg.ggplot(plt_df) + gg.aes('t_', 'cum_regret_mean', colour='agent_') +
gg.geom_line(size=1.25, alpha=0.75) +
gg.geom_ribbon(gg.aes(ymin='cum_regret_lower_interval',
ymax='cum_regret_upper_interval',
fill='agent_'),
alpha=0.1) + gg.xlab('time period (t)') +
gg.ylab('cumulative regret') +
gg.scale_colour_brewer(name='agent_', type='qual', palette='Set1'))
plot_dict = {experiment_name + '_cum_regret': p}
return plot_dict
def misspecified_plot(experiment_name='finite_misspecified',
data_path=_DEFAULT_DATA_PATH):
"""Specialized plotting script for TS tutorial paper misspecified TS."""
df = load_data(experiment_name, data_path)
def _parse_np_array(np_string):
return np.array(np_string.replace('[', '')
.replace(']', '')
.strip()
.split())
df['posterior_mean'] = df.posterior_mean.apply(_parse_np_array)
# Action means
new_col_list = ['mean_0', 'mean_1', 'mean_2']
for n, col in enumerate(new_col_list):
df[col] = df['posterior_mean'].apply(lambda x: float(x[n]))
plt_df = (df.groupby(['agent', 't'])
.agg({'instant_regret': np.mean,
'mean_0': np.mean,
'mean_1': np.mean,
'mean_2': np.mean})
.reset_index())
regret_plot = (gg.ggplot(plt_df)
+ gg.aes('t', 'instant_regret', colour='agent')
+ gg.geom_line(size=1.25, alpha=0.75)
+ gg.xlab('Timestep (t)')
+ gg.ylab('Average instantaneous regret')
+ gg.scale_colour_brewer(name='Agent', type='qual', palette='Set1')
+ gg.coord_cartesian(ylim=(0, 0.02)))
melt_df = pd.melt(plt_df, id_vars=['agent', 't'], value_vars=new_col_list)
melt_df['group_id'] = melt_df.agent + melt_df.variable
action_plot = (gg.ggplot(melt_df)
+ gg.aes('t', 'value', colour='agent', group='group_id')
+ gg.geom_line(size=1.25, alpha=0.75)
+ gg.coord_cartesian(ylim=(0, 0.05))
+ gg.xlab('Timestep (t)')
+ gg.ylab('Expected mean reward')
+ gg.scale_colour_brewer(name='Agent', type='qual', palette='Set1'))
plot_dict = {}
plot_dict['misspecified_regret'] = regret_plot
plot_dict['misspecified_action'] = action_plot
return plot_dict
def plot_action_proportion(df_agent):
"""Plot the action proportion for the sub-dataframe for a single agent."""
n_action = np.max(df_agent.action) + 1
plt_data = []
for i in range(n_action):
probs = (df_agent.groupby('t').agg({
'action': lambda x: np.mean(x == i)
}).rename(columns={'action': 'action_' + str(i)}))
plt_data.append(probs)
plt_df = pd.concat(plt_data, axis=1).reset_index()
p = (gg.ggplot(pd.melt(plt_df, id_vars='t')) +
gg.aes('t', 'value', colour='variable', group='variable') +
gg.geom_line(size=1.25, alpha=0.75) + gg.xlab('Timestep (t)') +
gg.ylab('Action probability') + gg.ylim(0, 1) +
gg.scale_colour_brewer(name='Variable', type='qual', palette='Set1'))
return p
def plot_qq(df, color_var, facet_var=None, title=''):
"""
Inspired by https://www.cureffi.org/2012/08/15/qq-plots-with-matplotlib/
"""
# retrive pmin, the most significant (i.e. min) p value (for defining
# the axes)
axis_max = max(df['pval_neglog10'])
if facet_var is None:
pvals = df.groupby(
by=color_var).apply(calculate_expected_pval).reset_index(
level=color_var, drop=True)
else:
pvals = df.groupby(by=[color_var, facet_var]).apply(
calculate_expected_pval).reset_index(level=[color_var, facet_var],
drop=True)
# now plot these two arrays against each other
n_colors = pvals[color_var].nunique()
qqplot = plt9.ggplot(
pvals,
plt9.aes(x='expected_pval_neglog10',
y='pval_neglog10',
color=color_var))
qqplot = qqplot + plt9.geom_point(size=0.1, alpha=0.25)
qqplot = qqplot + plt9.geom_abline(
slope=1, intercept=0, color='black', linetype='dashed')
qqplot = qqplot + plt9.theme_bw()
if n_colors < 9:
qqplot = qqplot + plt9.scale_colour_brewer(palette='Dark2',
type='qual')
qqplot = qqplot + plt9.labs(x='Expected (-log10 p-value)',
y='Observed (-log10 p-value)',
title=title,
color='')
qqplot = qqplot + plt9.lims(x=(0, axis_max), y=(0, axis_max))
if facet_var is not None:
qqplot = qqplot + plt9.facet_wrap('~ {}'.format(facet_var), ncol=5)
qqplot = qqplot + plt9.theme(strip_text=plt9.element_text(size=5),
axis_text_x=plt9.element_text(angle=-45,
hjust=0))
# set guide legend alpha to 1
qqplot = qqplot + plt9.guides(color=plt9.guide_legend(override_aes={
'size': 2.0,
'alpha': 1.0
}))
return (qqplot)
def simple_algorithm_plot(experiment_name, data_path=_DEFAULT_DATA_PATH):
"""Simple plot of average instantaneous regret by agent, per timestep.
Args:
experiment_name: string = name of experiment config.
data_path: string = where to look for the files.
Returns:
plot_dict: {experiment_name: ggplot plot}
"""
df = load_data(experiment_name, data_path)
plt_df = (df.groupby(['t', 'agent']).agg({
'instant_regret': np.mean
}).reset_index())
p = (gg.ggplot(plt_df) + gg.aes('t', 'instant_regret', colour='agent') +
gg.geom_line(size=1.25, alpha=0.75) + gg.xlab('Timestep (t)') +
gg.ylab('Average instantaneous regret') +
gg.scale_colour_brewer(name='Agent', type='qual', palette='Set1'))
return {experiment_name: p}
def simple_algorithm_plot(experiment_name, data_path=_DEFAULT_DATA_PATH):
"""Simple plot of average instantaneous regret by agent, per timestep.
Args:
experiment_name: string = name of experiment config.
data_path: string = where to look for the files.
Returns:
https://web.stanford.edu/~bvr/pubs/TS_Tutorial.pdf
"""
df = load_data(experiment_name, data_path)
plt_df = (df.groupby(['t', 'agent']).agg({
'instant_regret': np.mean
}).reset_index())
p = (gg.ggplot(plt_df) + gg.aes('t', 'instant_regret', colour='agent') +
gg.geom_line(size=1.25, alpha=0.75) + gg.xlab('time period (t)') +
gg.ylab('per-period regret') +
gg.scale_colour_brewer(name='agent', type='qual', palette='Set1'))
plot_dict = {experiment_name + '_simple': p}
return plot_dict
def cumulative_travel_time_plot(experiment_name, data_path=_DEFAULT_DATA_PATH):
"""Plot cumulative ratio total travel time relative to optimal shortest path.
Args:
experiment_name: string = name of experiment config.
data_path: string = where to look for the files.
Returns:
plot_dict: {experiment_name: ggplot plot}
"""
df = load_data(experiment_name, data_path)
df['cum_ratio'] = (df.cum_optimal - df.cum_regret) / df.cum_optimal
plt_df = (df.groupby(['t', 'agent']).agg({
'cum_ratio': np.mean
}).reset_index())
p = (gg.ggplot(plt_df) + gg.aes('t', 'cum_ratio', colour='agent') +
gg.geom_line(size=1.25, alpha=0.75) + gg.xlab('Timestep (t)') +
gg.ylab('Total distance / optimal') +
gg.scale_colour_brewer(name='Agent', type='qual', palette='Set1') +
gg.aes(ymin=1) +
gg.geom_hline(yintercept=1, linetype='dashed', size=2, alpha=0.5))
return {experiment_name + '_cum': p}
def main():
"""Run CLI."""
parser = argparse.ArgumentParser(description="""
Read anndata object. Read chrX and chrY genes, plot scatterplot
of mean expression of those signature across samples. Anndata
should have experiment_id and sex columns.
""")
parser.add_argument('-h5',
'--h5_anndata',
dest='h5',
required=True,
help='H5 AnnData file.')
parser.add_argument(
'-Y',
'--chrY_genes',
default='',
dest='Y',
required=False,
help='TSV file of Y genes. If none, uses all genes on Y chr.')
parser.add_argument('-X',
'--chrX_genes',
default='',
dest='X',
required=False,
help='TSV file of X genes. If none, uses XIST.')
parser.add_argument('-o',
'--output_file',
default='scatterplot-sex_sample_swap_check',
dest='o',
help='Basename for output files.')
options = parser.parse_args()
# Load the AnnData file
adata = sc.read_h5ad(filename=options.h5)
# If we have a flag for cells that pass QC then filter down to them
if 'cell_passes_qc' in adata.obs:
adata = adata[adata.obs['cell_passes_qc'], :]
del adata.obs['cell_passes_qc']
# Read Chr X and Chr Y genes
if options.X != '':
X = pd.read_csv(options.X, sep="\t")
X = X['ensembl_gene_id']
X_lab = "Mean X chr gene expression (counts)"
else:
X = ['ENSG00000229807']
# X = ['XIST']
X_lab = "Mean XIST gene expression (counts)"
if options.Y != '':
Y = pd.read_csv(options.Y, sep="\t")
Y = Y['ensembl_gene_id']
else:
Y = [
"ENSG00000184895", "ENSG00000129824", "ENSG00000067646",
"ENSG00000176679", "ENSG00000099715", "ENSG00000168757",
"ENSG00000099721", "ENSG00000092377", "ENSG00000099725",
"ENSG00000233803", "ENSG00000229549", "ENSG00000228927",
"ENSG00000258992", "ENSG00000238074", "ENSG00000236424",
"ENSG00000114374", "ENSG00000067048", "ENSG00000183878",
"ENSG00000154620", "ENSG00000129864", "ENSG00000129862",
"ENSG00000165246", "ENSG00000129873", "ENSG00000182415",
"ENSG00000172468", "ENSG00000169953", "ENSG00000286265",
"ENSG00000012817", "ENSG00000198692", "ENSG00000280969",
"ENSG00000242875", "ENSG00000234414", "ENSG00000244395",
"ENSG00000242389", "ENSG00000169807", "ENSG00000169800",
"ENSG00000226941", "ENSG00000169789", "ENSG00000183753",
"ENSG00000188120", "ENSG00000205944", "ENSG00000169763",
"ENSG00000172352", "ENSG00000183795", "ENSG00000187191",
"ENSG00000205916", "ENSG00000185894", "ENSG00000172288"
]
# Same as above, but hugo names.
# Y = [
# "SRY",
# "RPS4Y1",
# "ZFY",
# "TGIF2LY",
# "PCDH11Y",
# "TSPY2",
# "AMELY",
# "TBL1Y",
# "PRKY",
# "TSPY4",
# "TSPY8",
# "TSPY3",
# "TSPY1",
# "TSPY9P",
# "TSPY10",
# "USP9Y",
# "DDX3Y",
# "UTY",
# "TMSB4Y",
# "VCY",
# "VCY1B",
# "NLGN4Y",
# "CDY2B",
# "CDY2A",
# "HSFY1",
# "HSFY2",
# "AC007244.1",
# "KDM5D",
# "EIF1AY",
# "RPS4Y2",
# "RBMY1B",
# "RBMY1A1",
# "RBMY1D",
# "RBMY1E",
# "PRY2",
# "RBMY1F",
# "RBMY1J",
# "PRY",
# "BPY2",
# "DAZ1",
# "DAZ2",
# "PRYP3",
# "CDY1B",
# "BPY2B",
# "DAZ3",
# "DAZ4",
# "BPY2C",
# "CDY1"
# ]
# Make the plot
adata.var['X_chr-gene'] = np.in1d(adata.var.index, X)
adata.var['Y_chr-gene'] = np.in1d(adata.var.index, Y)
adata.obs['X_chr-sum'] = adata[:, adata.var['X_chr-gene']].X.todense().sum(
axis=1)
adata.obs['Y_chr-sum'] = adata[:, adata.var['Y_chr-gene']].X.todense().sum(
axis=1)
if 'sex' not in adata.obs.columns:
adata.obs['sex'] = 'not reported'
df = adata.obs[['experiment_id', 'sex', 'Y_chr-sum', 'X_chr-sum']]
df = df.groupby(['experiment_id', 'sex']).mean().dropna().reset_index()
# Save scatterplot with mean expression per sample
plt = plt9.ggplot(df) + plt9.aes(x='X_chr-sum', y='Y_chr-sum', color='sex')
plt = plt + plt9.theme_bw()
plt = plt + plt9.scale_colour_brewer(type='qual', palette='Dark2')
plt = plt + plt9.geom_point(alpha=0.45)
plt = plt + plt9.ylab("Mean Y chr gene expression (counts)")
plt = plt + plt9.xlab(X_lab)
plt.save('{}.png'.format(options.o), dpi=300, width=4, height=4)
def concurrent_agents_plot(experiment_name='graph_indep_concurrent',
data_path=_DEFAULT_DATA_PATH,
paper_version=True):
'''Passing paper_version=True should be used to reproduce Fig. 14 of the paper
for K = 1,10,20,50,100. In this case, the labels in the legend are manually
ordered by the values of K. Otherwise, the labels are ordered alphabetically.'''
df = load_data(data_path, experiment_name)
plt_df_per_action = (df.groupby(['agent', 't', 'agent_id',
'action_id']).agg({
'instant_regret':
np.mean
}).reset_index())
plt_df_per_period = (df.groupby(['agent', 't']).agg({
'instant_regret':
np.mean
}).reset_index())
if not paper_version:
p_per_action = (
gg.ggplot(plt_df_per_action) +
gg.aes('action_id', 'instant_regret', colour='agent') +
gg.geom_line() + gg.geom_line(size=1.25, alpha=0.75) +
gg.xlim(0, 2.5 * len(plt_df_per_period.groupby('t'))) +
gg.scale_colour_brewer(name='agent', type='qual', palette='Set1') +
gg.labels.xlab('number of actions') +
gg.labels.ylab('per-period regret'))
p_per_period = (
gg.ggplot(plt_df_per_period) +
gg.aes('t', 'instant_regret', colour='agent') + gg.geom_line() +
gg.geom_line(size=1.25, alpha=0.75) +
gg.scale_colour_brewer(name='agent', type='qual', palette='Set1') +
gg.labels.xlab('time period (t)') +
gg.labels.ylab('per-period regret'))
else:
plt_df_per_action['agent_id'] = plt_df_per_action.agent.apply(
get_agent_id)
plt_df_per_period['agent_id'] = plt_df_per_period.agent.apply(
get_agent_id)
custom_labels = ['K = 1', 'K = 10', 'K = 20', 'K = 50', 'K = 100']
custom_colors = ["#E41A1C", "#377EB8", "#4DAF4A", "#984EA3", "#FF7F00"]
p_per_action = (
gg.ggplot(plt_df_per_action) +
gg.aes('action_id', 'instant_regret', colour='agent_id') +
gg.geom_line() + gg.geom_line(size=1.25, alpha=0.75) +
gg.xlim(0, 2.5 * len(plt_df_per_period.groupby('t'))) +
gg.scale_color_manual(
name='agent', labels=custom_labels, values=custom_colors) +
gg.labels.xlab('number of actions') +
gg.labels.ylab('per-action regret'))
p_per_period = (
gg.ggplot(plt_df_per_period) +
gg.aes('t', 'instant_regret', colour='agent_id') + gg.geom_line() +
gg.geom_line(size=1.25, alpha=0.75) + gg.scale_color_manual(
name='agent', labels=custom_labels, values=custom_colors) +
gg.labels.xlab('time period (t)') +
gg.labels.ylab('per-period regret'))
plot_dict = {}
plot_dict['per_action_plot'] = p_per_action
plot_dict['per_period_plot'] = p_per_period
return plot_dict
def plot_ecdf(df_plot,
variable_column,
color_column='none',
output_file='plot_distribution',
facet_column='none',
x_log10=False):
"""Plot plot_distribution to png.
Parameters
----------
df_plot : pandas.DataFrame
DataFrame with <variable_column> as a column.
variable_column : string
String of variable_column column to plot.
color_column : string
String of color column to plot.
output_file : string
Basename of output file.
facet_column : string
Column to facet the plot by.
Returns
-------
NULL
"""
n_colors = 0
if color_column != 'none':
gplt = plt9.ggplot(df_plot,
plt9.aes(x=variable_column, color=color_column))
n_colors = df_plot[color_column].nunique()
else:
gplt = plt9.ggplot(df_plot, plt9.aes(x=variable_column))
gplt = gplt + plt9.theme_bw()
gplt = gplt + plt9.stat_ecdf(alpha=0.8)
if x_log10:
gplt = gplt + plt9.scale_x_continuous(
trans='log10',
# labels=comma_labels,
minor_breaks=0)
else:
gplt = gplt + plt9.scale_x_continuous(
# trans='log10',
# labels=comma_labels,
minor_breaks=0)
gplt = gplt + plt9.scale_y_continuous(
# trans='log10',
# labels=comma_labels,
minor_breaks=0)
gplt = gplt + plt9.labs(y='Cumulative density', title='')
if n_colors != 0 and n_colors > 20:
gplt = gplt + plt9.theme(legend_position='none')
elif n_colors != 0 and n_colors < 9:
gplt = gplt + plt9.scale_colour_brewer(palette='Dark2', type='qual')
if facet_column != 'none':
gplt = gplt + plt9.facet_wrap('~ {}'.format(facet_column), ncol=5)
n_facets = df_plot[facet_column].nunique()
gplt.save('{}.png'.format(output_file),
dpi=300,
width=6 * (n_facets / 4),
height=4 * (n_facets / 4),
limitsize=False)
else:
gplt.save('{}.png'.format(output_file), dpi=300, width=4, height=4)
return 0
def plot_umi_mt_density(df_plot,
output_file='plot_umi_mt_density',
facet_column='none',
color_var='density',
density_contour=False):
"""Plot plot_umi_mt_density to png.
Parameters
----------
df_plot : pandas.DataFrame
DataFrame with the followig keys 'total_counts', 'pct_counts_gene_group__mito_transcript'.
output_file : string
Basename of output file.
facet_column : string
Column to facet the output by.
Returns
-------
NULL
"""
if color_var == 'density':
color_title = 'Density\n'
# Also calculate density using a gaussian 2d kernal -- use random
# name for plot column
color_var = "1251234_density"
df_plot[color_var] = calculate_density(df_plot, facet_column)
elif color_var == 'pct_counts_gene_group__mito_transcript':
color_title = '% MT\n'
elif color_var == 'cell_passes_qc':
color_title = 'Cell passed QC\n'
else:
color_title = color_var
gplt = plt9.ggplot(
df_plot,
plt9.aes(x='total_counts',
y='pct_counts_gene_group__mito_transcript',
color=color_var))
gplt = gplt + plt9.theme_bw()
gplt = gplt + plt9.geom_point(alpha=0.5, size=0.8)
gplt = gplt + plt9.scale_x_continuous(
trans='log10', labels=comma_labels, minor_breaks=0)
if color_var == 'pct_counts_gene_group__mito_transcript':
gplt = gplt + plt9.scale_color_gradient2(low='#3B9AB2',
mid='#EBCC2A',
high='#F21A00',
midpoint=50,
limits=[0, 100])
gplt = gplt + plt9.guides(color=plt9.guide_colorbar(ticks=False))
elif color_var == 'cell_passes_qc':
gplt = gplt + plt9.scale_colour_brewer(type='qual', palette='Dark2')
elif color_var == '1251234_density':
gplt = gplt + plt9.scale_color_cmap(cmap_name='viridis')
if density_contour:
gplt = gplt + plt9.geom_density_2d(alpha=0.5)
gplt = gplt + plt9.labs(x='Number of molecules',
y='Percent of molecules from MT genes',
title='',
color=color_title)
if facet_column != 'none':
gplt = gplt + plt9.facet_wrap('~ {}'.format(facet_column), ncol=5)
n_samples = df_plot[facet_column].nunique()
gplt.save('{}.png'.format(output_file),
dpi=300,
width=4 * (n_samples / 2),
height=4 * (n_samples / 4),
limitsize=False)
else:
gplt.save('{}.png'.format(output_file), dpi=300, width=4, height=4)
