def accuracyPlot(topo_types, task_types, trace_types, pair_seq=0):
accuracys, errors, x = [], [], None
cherns, chern_errors, chern_x = [], [], None
acc_boxplot_data = []
for trace_type in trace_types:
topo_type = topo_types[1]
task_type = task_types[0]
if topo_type == topo_types[0]:
pair_num = 4
elif topo_type == topo_types[1]:
pair_num = 8
elif topo_type == topo_types[2]:
pair_num = 14
infile_name1, infile_name2 = init_files(trace_type,
task_type,
topo_type,
x_var='PVALUE')
group_fields = ['#flows', 'is_correlated', 'key']
agg_fields = ['count_chern', 'byte_chern']
chern_x, case_num, acc_dict = retieveData(infile_name2,
group_fields,
agg_fields,
accuracys,
errors,
div=False,
pair_num=pair_num)
acc_boxplot_data.append(acc_dict.values())
acc_legend = ['Database', 'Web', 'Hadoop']
xlabel, ylabel = 'The threshold for p-value', 'Average accuracy'
print accuracys, chern_x
pl.plot(accuracys,
x=chern_x,
k=2,
errors=errors,
xlabel=xlabel,
ylabel=ylabel,
title=infile_name1,
xlog=False,
ylog=False,
acc_legend=acc_legend,
legend_y=0.85)
print acc_boxplot_data
pl.box_plot([acc_boxplot_data[0]],
x=chern_x,
k=2,
errors=errors,
xlabel=xlabel,
ylabel=ylabel,
title=infile_name1,
xlog=False,
ylog=False,
acc_legend=acc_legend,
legend_y=0.85,
xticks=[str(i) for i in chern_x])
def accuracyPlot(topo_types, task_types, trace_types, pair_seq=0):
accuracys, errors, x = [], [], None
cherns, chern_errors, chern_x = [], [], None
acc_boxplot_data = []
q_75ss, xss = [], []
for topo_type in topo_types[0:3]:
trace_type = trace_types[0]
task_type = task_types[0]
if topo_type == topo_types[0]:
pair_num = 4
elif topo_type == topo_types[1]:
pair_num = 8
elif topo_type == topo_types[2]:
pair_num = 14
infile_name1, infile_name2 = init_files(
trace_type, task_type, topo_type,
x_var = 'CHERNPVALUEBYTE' #CHERNPVALUE
)
infile_name3, infile_name4 = init_files(
trace_type, task_type, topo_type,
x_var = 'TTESTPVALUEBYTE'
)
key_field = 'epsilon'
group_fields = [key_field,'is_correlated', 'key']
agg_fields = ['count_chern', 'byte_chern']
chern_x, case_num, acc_dict = retieveData(
infile_name2, group_fields, agg_fields,
accuracys, errors, div=False, pair_num=pair_num, q_75ss=q_75ss,
key_field=key_field, topo_type=topo_type, xss=xss
)
chern_x, case_num, acc_dict = retieveData(
infile_name4, group_fields, agg_fields,
accuracys, errors, div=False, pair_num=pair_num, q_75ss=q_75ss,
key_field=key_field, topo_type=topo_type, xss=xss
)
acc_boxplot_data.append(acc_dict.values())
acc_legend = ['B4-Chernoff', 'B4-t-test','Tree-Chernoff', 'Tree-t-test', 'Jupiter-Chernoff', 'Jupiter-t-test']
xlabel, ylabel = '$P_b$', 'Average accuracy' #'Relative elephant flow size'
print 'accuracys, chern_x', len(accuracys[0]), len(chern_x)
print accuracys, xss
pl.plot(accuracys, x=xss, k=2, errors=[],
xlabel=xlabel, ylabel=ylabel, title=infile_name1.split('/outputs')[1].split('.csv')[0].replace('.', '_'),
xlog=True, ylog=False, acc_legend=acc_legend, legend_x=0.0,
legend_y=0.3
)
print acc_boxplot_data
pl.box_plot([acc_boxplot_data[0]], x=chern_x, k=2, errors=errors,
xlabel=xlabel, ylabel=ylabel, title=infile_name1.split('/outputs')[1].split('.csv')[0].replace('.', '_'),
xlog=False, ylog=False, acc_legend=acc_legend,
legend_y=0.85, xticks=[str(i) for i in chern_x])
def accuracyPlot(topo_types, task_types, trace_types, pair_seq=0):
accuracys, errors, times, x = [], [], [], None
cherns, chern_errors, time_chern, chern_x = [], [], [], None
for trace_type in trace_types:
topo_type = topo_types[0]
task_type = task_types[1]
infile_name1, infile_name2 = init_files(trace_type,
task_type,
topo_type,
x_var='TTESTFLOWNUM',
p=0.01)
group_fields = ['#flows', 'is_correlated']
agg_fields = ['#correlated_detect', '#non-uniform_detect']
x, case_num = retieveData(infile_name1,
group_fields,
agg_fields,
accuracys,
errors,
times,
div=True)
print x, case_num
group_fields = ['#flows', 'is_correlated', 'key']
agg_fields = ['count_chern', 'byte_chern', 'time']
chern_x = retieveData(infile_name2,
group_fields,
agg_fields,
cherns,
chern_errors,
time_chern,
div=False)
acc_legend = [
'Database', 'Web', 'Hadoop', 'e1', 'e5', 'e10', 'e50', 'e100'
]
xlabel, ylabel = 'The number of flows', 'Average accuracy'
pl.plot(accuracys,
x=x,
k=2,
errors=errors,
xlabel=xlabel,
ylabel=ylabel,
title=infile_name1.split('.')[0],
xlog=False,
ylog=False,
acc_legend=acc_legend[0:3])
"""xlabel, ylabel = 'The number of flows', 'Average p-value'
def rankingPlot(topo_types, task_types, trace_types, pair_seq=0):
global ACAP
accuracys, errors, x = [], [], None
cherns, chern_errors, chern_x = [], [], None
for trace_type in trace_types:
topo_type = topo_types[1]
task_type = task_types[1]
cut = 5
if topo_type == topo_types[2]:
ACAP = 256
cut = 10
infile_name1, infile_name2 = init_files(trace_type, task_type,
topo_type)
group_fields = ['#flows', 'is_correlated']
agg_fields = ['#correlated_detect', '#correlated_detect']
x, case_num = retieveData(infile_name1,
group_fields,
agg_fields,
accuracys,
errors,
div=True,
task_type='RANKING')
"""print(len(x))
group_fields = ['#flows','is_correlated', 'key']
agg_fields = ['a'+str(i) for i in range(ACAP)]
#print(agg_fields)
chern_x = rankingPair(
infile_name2, cherns, 6, pair_seq,
file_type=1, topo_type=topo_type, cut=cut
)"""
acc_legend = [
'Database', 'Web', 'Hadoop', 'e1', 'e5', 'e10', 'e50', 'e100'
]
xlabel, ylabel = 'The number of flows', 'Average accuracy'
pl.plot(accuracys,
x=x,
k=2,
errors=errors,
xlabel=xlabel,
ylabel=ylabel,
title=infile_name1.split('.')[0],
xlog=False,
ylog=False,
acc_legend=acc_legend[3:7])
"""xlabel, ylabel = 'Previous hop', 'Average p-value'
def rankingPlot(topo_types, task_types, trace_types, pair_seq=0):
global ACAP
accuracys, errors, x, q_75s, q_25s, times = [], [], None, [], [], []
cherns, chern_errors, chern_x, time_chern = [], [], None, []
xss = []
xss1 = []
for topo_type in topo_types[0:3]:
trace_type = trace_types[0]
task_type = task_types[1]
cut = 5
if topo_type == topo_types[2]:
ACAP = 256
cut = 10
if topo_type == topo_types[0]:
pair_num = 4
elif topo_type == topo_types[1]:
pair_num = 8
elif topo_type == topo_types[2]:
pair_num = 14
infile_name1, infile_name2 = init_files(trace_type,
task_type,
topo_type,
x_var='CHERNFLOWNUM',
p=0.01)
infile_name3, infile_name4 = init_files(trace_type,
task_type,
topo_type,
x_var='TTESTFLOWNUM',
p=0.01)
key_field = '#flows'
group_fields = ['#flows', 'is_correlated']
agg_fields = ['#correlated_detect', '#correlated_detect']
x, case_num = retieveData(infile_name1,
group_fields,
agg_fields,
accuracys,
errors,
times,
div=True,
task_type='RANKING',
q_75s=q_75s,
q_25s=q_25s,
xss=xss)
x, case_num = retieveData(infile_name3,
group_fields,
agg_fields,
accuracys,
errors,
time_chern,
div=True,
task_type='RANKING',
q_75s=q_75s,
q_25s=q_25s,
xss=xss)
getTime(infile_name2,
times,
group_fields=group_fields,
key_field=key_field,
file_type=1,
topo_type=topo_type,
case_num=pair_num,
xss=xss1)
getTime(infile_name4,
times,
group_fields=group_fields,
key_field=key_field,
file_type=1,
topo_type=topo_type,
case_num=pair_num,
xss=xss1)
print(len(x))
group_fields = ['#flows', 'is_correlated', 'key']
agg_fields = ['a' + str(i) for i in range(ACAP)]
#print(agg_fields)
"""chern_x = rankingPair(
infile_name2, cherns, 2, pair_seq,
file_type=1, topo_type=topo_type, cut=cut
)"""
acc_legend = [
'B4-Chernoff', 'B4-t-test', 'Tree-Chernoff', 'Tree-t-test',
'Jupiter-Chernoff', 'Jupiter-t-test', 'e1', 'e5', 'e10', 'e50', 'e100'
]
xlabel, ylabel = 'The number of flows', 'Average accuracy'
pl.plot(accuracys,
x=xss,
k=2,
errors=[],
xlabel=xlabel,
ylabel=ylabel,
title=infile_name1.split('/outputs')[1].split('.csv')[0].replace(
'.', '_'),
xlog=False,
ylog=False,
acc_legend=acc_legend[:len(accuracys)],
legend_x=0.45)
x = [20, 30, 40, 50, 60, 70]
xss = [x] * len(times)
print times, len(times[0])
pl.plot(times,
x=xss1,
k=2,
errors=[],
xlabel=xlabel,
ylabel='Execution time (seconds)',
title=infile_name1.split('/outputs')[1].split('.csv')[0].replace(
'.', '_') + '_time',
xlog=False,
ylog=False,
acc_legend=acc_legend,
legend_y=0.85,
legend_x=0.0)
xlabel, ylabel = 'Previous hop', 'Average p-value'
print chern_x
xticks = []
if chern_x != None:
xticks = chern_x
"""pl.plot(cherns, x=[], k=len(cherns), errors=[],
def rankingPlot(topo_types, task_types, trace_types, pair_seq=0):
global ACAP
accuracys, errors, x, q_75s, q_25s = [], [], None, [], []
cherns, chern_errors, chern_x = [], [], None
for trace_type in trace_types:
topo_type = topo_types[1]
task_type = task_types[1]
cut = 5
if topo_type == topo_types[2]:
ACAP = 256
cut = 10
infile_name1, infile_name2 = init_files(trace_type, task_type,
topo_type)
group_fields = ['#flows', 'is_correlated']
agg_fields = ['#correlated_detect', '#correlated_detect']
x, case_num = retieveData(infile_name1,
group_fields,
agg_fields,
accuracys,
errors,
div=True,
task_type='RANKING',
q_75s=q_75s,
q_25s=q_25s)
print(len(x))
group_fields = ['#flows', 'is_correlated', 'key']
agg_fields = ['a' + str(i) for i in range(ACAP)]
#print(agg_fields)
chern_x = rankingPair(infile_name2,
cherns,
0,
pair_seq,
file_type=1,
topo_type=topo_type,
cut=cut)
acc_legend = [
'Database', 'Web', 'Hadoop', 'e1', 'e5', 'e10', 'e50', 'e100'
]
xlabel, ylabel = 'The number of flows', 'Average accuracy'
pl.plot(accuracys,
x=x,
k=2,
errors=q_75s,
xlabel=xlabel,
ylabel=ylabel,
title=infile_name1.split('.')[0],
xlog=False,
ylog=False,
acc_legend=acc_legend[3:8])
xlabel, ylabel = 'Previous hop', 'Average p-value'
print chern_x
xticks = []
if chern_x != None:
xticks = chern_x
pl.plot(cherns,
x=None,
k=len(cherns),
errors=[],
xlabel=xlabel,
ylabel=ylabel,
title=infile_name1.split('.')[0] + 'pair-key' + str(pair_seq),
xlog=False,
ylog=True,
acc_legend=acc_legend,
xticks=chern_x,
figure_width=4.3307 * 1.25)
def rankingPlot(topo_types, task_types, trace_types, pair_seq=0):
global ACAP
accuracys, errors, x, q_75s, q_25s = [], [], None, [], []
cherns, chern_errors, chern_x = [], [], None
for topo_type in topo_types[1:2]:
trace_type = trace_types[0]
task_type = task_types[1]
cut = 4
if topo_type == topo_types[2]:
ACAP = 256
cut = 10
infile_name1, infile_name2 = init_files(
trace_type,
task_type,
topo_type,
x_var='CHERNRANKINGPVALUEFLOWNUM')
infile_name3, infile_name4 = init_files(
trace_type,
task_type,
topo_type,
x_var='TTESTRANKINGPVALUEFLOWNUM')
group_fields = ['#flows', 'is_correlated', 'key']
agg_fields = ['a' + str(i) for i in range(ACAP)]
chern_x = rankingPair(infile_name2,
cherns,
0, [0, 1],
file_type=1,
topo_type=topo_type,
cut=cut,
q_25s=q_25s,
q_75s=q_75s)
chern_x = rankingPair(infile_name4,
cherns,
0, [0, 1],
file_type=1,
topo_type=topo_type,
cut=cut,
q_25s=q_25s,
q_75s=q_75s)
xlabel, ylabel = 'Rank', 'Average p-value'
xticks = []
if chern_x != None:
xticks = chern_x
print chern_x, cherns,
print 'q****', q_25s, q_75s
pl.plot(cherns,
x=[],
k=2,
errors=[],
xlabel=xlabel,
ylabel=ylabel,
title=infile_name1.split('/outputs')[1].split('.csv')[0] +
'pair-key' + str(pair_seq),
xlog=False,
ylog=True,
acc_legend=[
'0-hop-Chernoff', '1-hop-Chernoff', '0-hop-t-test',
'1-hop-t-test'
],
xticks=np.arange(1, cut + 1),
figure_width=4.3307,
figure_height=3.346,
x_shift=0.06,
y_shift=0.05,
legend_x=0.3) #[q_25s, q_75s]
def accuracyPlot(topo_types, task_types, trace_types, pair_seq=0):
accuracys, errors, x = [], [], None
cherns, chern_errors, chern_x = [], [], None
acc_boxplot_data = []
q_75ss, xss = [], []
for topo_type in topo_types[2:3]:
trace_type = 'ABC'
task_type = task_types[2]
if topo_type == topo_types[0]:
pair_num = 4
elif topo_type == topo_types[1]:
pair_num = 8
elif topo_type == topo_types[2]:
pair_num = 16
infile_name1, infile_name2 = init_files(
trace_type,
task_type,
topo_type,
x_var='TOPKBIASONEFIVE' #CHERNPVALUE
)
infile_name3, infile_name4 = init_files(
trace_type,
task_type,
topo_type,
x_var='TOPKBIASSMALL' #CHERNPVALUE
)
infile_name5, infile_name6 = init_files(
trace_type,
task_type,
topo_type,
x_var='TOPK' #CHERNPVALUE
)
key_field = 'topk'
group_fields = [key_field, 'is_correlated', 'key']
agg_fields = ['count_chern', 'byte_chern']
chern_x, case_num, acc_dict = retieveData(infile_name2,
group_fields,
agg_fields,
accuracys,
errors,
div=False,
pair_num=pair_num,
q_75ss=q_75ss,
key_field=key_field,
topo_type=topo_type,
xss=xss)
acc_boxplot_data.append(acc_dict.values())
agg_fields = ['count_chern', 'byte_chern']
chern_x, case_num, acc_dict = retieveData(infile_name4,
group_fields,
agg_fields,
accuracys,
errors,
div=False,
pair_num=pair_num,
q_75ss=q_75ss,
key_field=key_field,
topo_type=topo_type,
xss=xss)
acc_boxplot_data.append(acc_dict.values())
agg_fields = ['count_chern', 'byte_chern']
chern_x, case_num, acc_dict = retieveData(infile_name6,
group_fields,
agg_fields,
accuracys,
errors,
div=False,
pair_num=pair_num,
q_75ss=q_75ss,
key_field=key_field,
topo_type=topo_type,
xss=xss)
acc_boxplot_data.append(acc_dict.values())
acc_legend = [
'dropping rate = 0.1', 'dropping rate = 0.15', 'dropping rate = 0.2'
]
xlabel, ylabel = 'The number of links', 'Average accuracy' #'Relative elephant flow size'
print 'accuracys, chern_x', len(accuracys[0]), len(chern_x)
print accuracys, xss
pl.plot(accuracys,
x=xss,
k=2,
errors=q_75ss,
xlabel=xlabel,
ylabel=ylabel,
title=infile_name1.split('/outputs')[1].split('.csv')[0].replace(
'.', '_'),
xlog=False,
ylog=False,
acc_legend=acc_legend,
legend_x=0.3,
legend_y=0.2)
print len(acc_boxplot_data)
chern_x = [2 * x for x in chern_x]
pl.box_plot(
acc_boxplot_data,
x=chern_x,
k=2,
errors=errors,
xlabel=xlabel,
ylabel=ylabel,
title=infile_name1.split('/outputs')[1].split('.csv')[0].replace(
'.', '_'),
xlog=False,
ylog=False,
acc_legend=[],
legend_y=0.85,
xticks=[str(i) for i in chern_x])
def accuracyPlot(topo_types, task_types, c, pair_seq=0):
accuracys, errors, times, x = [], [], [], None
cherns, chern_errors, chern_x = [], [], None
acc_boxplot_data = []
q_75ss, xss = [], []
for topo_type in topo_types[0:3]:
trace_type = trace_types[0]
task_type = task_types[0]
if topo_type == topo_types[0]:
pair_num = 4
elif topo_type == topo_types[1]:
pair_num = 8
elif topo_type == topo_types[2]:
pair_num = 14
infile_name1, infile_name2 = init_files(trace_type,
task_type,
topo_type,
x_var='XFLOWNUM')
infile_name3, infile_name4 = init_files(trace_type,
task_type,
topo_type,
x_var='TTESTFLOWNUM')
infile_name5, infile_name6 = init_files(trace_type,
task_type,
topo_type,
x_var='CHERNTTFLOWNUM')
group_fields = ['#flows', 'is_correlated', 'key']
agg_fields = ['count_chern', 'byte_chern']
chern_x, case_num, acc_dict = retieveData(infile_name2,
group_fields,
agg_fields,
accuracys,
errors,
times,
div=False,
pair_num=pair_num,
q_75ss=q_75ss,
xss=xss)
chern_x, case_num, acc_dict = retieveData(infile_name4,
group_fields,
agg_fields,
accuracys,
errors,
times,
div=False,
pair_num=pair_num,
q_75ss=q_75ss,
xss=xss)
"""chern_x, case_num, acc_dict = retieveData(
infile_name6, group_fields, agg_fields,
accuracys, errors, times, div=False, pair_num=pair_num, q_75ss=q_75ss
)"""
acc_boxplot_data.append(acc_dict.values())
acc_legend = [
'B4-Chernoff', 'B4-t-test', 'Tree-Chernoff', 'Tree-t-test',
'Jupiter-Chernoff', 'Jupiter-t-test'
]
xlabel, ylabel = 'The number of flows', 'Average accuracy'
len_x = min([len(acc) for acc in accuracys])
#accuracys = [acc[:len_x] for acc in accuracys]
#chern_x = chern_x[:len_x]
plot_cut = 20
accuracys = [acc[:plot_cut] for acc in accuracys]
q_75ss = [[q_75[0][:plot_cut], q_75[1][:plot_cut]] for q_75 in q_75ss]
print len(chern_x), len(accuracys[0]), q_75ss[0]
print xss
pl.plot(accuracys,
x=xss,
k=2,
errors=[],
xlabel=xlabel,
ylabel=ylabel,
title=infile_name1.split('/outputs')[1].split('.csv')[0].replace(
'.', '_'),
xlog=False,
ylog=False,
acc_legend=acc_legend,
legend_y=0.3,
legend_x=0.45)
times = [time[:plot_cut] for time in times]
pl.plot(times,
x=xss,
k=2,
errors=[],
xlabel=xlabel,
ylabel='Execution time (seconds)',
title=infile_name1.split('/outputs')[1].split('.csv')[0].replace(
'.', '_') + '_time',
xlog=False,
ylog=False,
acc_legend=acc_legend,
legend_y=0.8,
legend_x=0.0)
pl.box_plot(
[acc_boxplot_data[0]],
x=chern_x,
k=2,
errors=errors,
xlabel=xlabel,
ylabel=ylabel,
title=infile_name1.split('/outputs')[1].split('.csv')[0].replace(
'.', '_'),
xlog=False,
ylog=False,
acc_legend=acc_legend,
legend_y=0.85,
xticks=[str(i) for i in chern_x])
def accuracyPlot(topo_types, task_types, trace_types, pair_seq=0):
accuracys, errors, x = [], [], None
cherns, chern_errors, chern_x = [], [], None
acc_boxplot_data = []
q_75ss = []
for topo_type in topo_types[0:2]:
trace_type = trace_types[0]
task_type = task_types[0]
if topo_type == topo_types[0]:
pair_num = 4
elif topo_type == topo_types[1]:
pair_num = 8
elif topo_type == topo_types[2]:
pair_num = 14
infile_name1, infile_name2 = init_files(trace_type,
task_type,
topo_type,
x_var='TTESTRTHRESHOLD')
group_fields = ['r_threshold', 'is_correlated', 'key']
agg_fields = ['count_chern', 'byte_chern']
chern_x, case_num, acc_dict = retieveData(infile_name2,
group_fields,
agg_fields,
accuracys,
errors,
div=False,
pair_num=pair_num,
q_75ss=q_75ss)
acc_boxplot_data.append(acc_dict.values())
acc_legend = ['Database', 'Web', 'Hadoop']
xlabel, ylabel = '$P_b$', 'Average accuracy'
print 'accuracys, chern_x', len(accuracys[0]), len(chern_x)
pl.plot(accuracys,
x=chern_x,
k=2,
errors=q_75ss,
xlabel=xlabel,
ylabel=ylabel,
title=infile_name1.split('/outputs')[1].split('.csv')[0].replace(
'.', '_'),
xlog=False,
ylog=False,
acc_legend=acc_legend,
legend_x=0.2,
legend_y=0.2)
print acc_boxplot_data
pl.box_plot(
[acc_boxplot_data[0]],
x=chern_x,
k=2,
errors=errors,
xlabel=xlabel,
ylabel=ylabel,
title=infile_name1.split('/outputs')[1].split('.csv')[0].replace(
'.', '_'),
xlog=False,
ylog=False,
acc_legend=acc_legend,
legend_y=0.85,
xticks=[str(i) for i in chern_x])