label_n=LABEL_N):
pattern = f'dp_{DP_PROBES_NAME}_{METRIC}.csv'
for p in param_path.rglob(pattern):
s = pd.read_csv(p, index_col=0, squeeze=True)
n = 1
y_mean = s.values
h = np.zeros((len(s))) # margin of error
# collect for comparison figure
summary = (s.index.values, y_mean, h, label, n)
summaries.append(summary)
if not summaries:
raise RuntimeError(f'Did not find csv files matching {pattern}')
# plot comparison
fig = make_summary_fig(
summaries,
ylabel=Y_LABEL,
title=TITLE,
log_y=LOG_Y,
ylims=Y_LIMS,
xlims=X_LIMS,
figsize=FIG_SIZE,
legend_loc='best',
# vline=200_000,
# legend_labels=['reverse age-ordered', 'age-ordered'],
# palette_ids=[0, 1], # re-assign colors to each line
)
fig.show()
print(f'--------------------- End section {p.name}')
print()
# sort data
summaries = sorted(summaries, key=lambda s: s[1][-1], reverse=True)
if not summaries:
raise SystemExit('No data found')
# print to console
for s in summaries:
_, y_mean, y_std, label, n = s
print(label)
print(y_mean)
print(y_std)
print()
# plot
fig = make_summary_fig(
summaries,
Y_LABEL,
title=TITLE,
palette_ids=PALETTE_IDS,
figsize=FIG_SIZE,
ylims=Y_LIMS,
legend_labels=LABELS,
vlines=V_LINES,
plot_max_lines=PLOT_MAX_LINES,
plot_max_line=PLOT_MAX_LINE,
legend_loc='best',
)
fig.show()
concatenated_df = pd.concat(series_list, axis=1)
y_mean = concatenated_df.mean(axis=1).values.flatten()
y_sem = sem(concatenated_df.values, axis=1)
h = y_sem * t.ppf((1 + CONFIDENCE) / 2, n - 1) # margin of error
return concatenated_df.index.values, y_mean, h, lb, n
summaries = []
project_name = __name__
for param_path, label in gen_param_paths(project_name,
param2requests,
param2default,
runs_path=RUNS_PATH,
ludwig_data_path=LUDWIG_DATA_PATH,
label_n=LABEL_N):
summary = make_summary(param_path, label, f'*_{PROBES_NAME}_js.csv')
summaries.append(summary)
fig = make_summary_fig(
summaries,
ylabel='Noun ' + Y_LABEL,
title='',
log_y=LOG_Y,
ylims=Y_LIMS,
figsize=FIG_SIZE,
legend_loc='best',
vline=200_000,
palette_ids=[0, 1], # re-assign colors to each line
)
fig.show()