def main(dirn, fname):
(xs, ysPerSolver, ydevsPerSolver) = CommonViz.parseData(dirn, fname)
CommonConf.setupMPPDefaults()
fmts = CommonConf.getLineFormats()
mrkrs = CommonConf.getLineMarkers()
fig = pp.figure()
ax = fig.add_subplot(111)
ax.set_xscale("log", basex=2)
#ax.set_yscale("symlog" )
index = 0
for (solver, ys), (solver, ydevs) in zip(ysPerSolver.iteritems(),
ydevsPerSolver.iteritems()):
ax.errorbar(xs,
ys,
yerr=ydevs,
label=solver,
marker=mrkrs[index],
linestyle=fmts[index])
index = index + 1
ax.set_xlabel(X_LABEL)
ax.set_ylabel(Y_LABEL)
ax.legend(loc='best', fancybox=True)
pp.savefig(dirn + "/" + fname + ".pdf")
pp.show()
def main(dirn, fname):
(xs, ysPerSolver, ydevsPerSolver) = CommonViz.parseData(dirn, fname)
CommonConf.setupMPPDefaults()
fmts = CommonConf.getLineFormats()
mrkrs = CommonConf.getLineMarkers()
fig = pp.figure()
ax = fig.add_subplot(111)
# ax.set_xscale("log", basex=2)
index = 0
for (solver, ys), (solver, ydevs) in zip(ysPerSolver.iteritems(),
ydevsPerSolver.iteritems()):
ax.errorbar(xs,
ys,
yerr=ydevs,
label=solver,
marker=mrkrs[index],
linestyle=fmts[index])
index = index + 1
ax.set_xlabel(X_LABEL)
ax.set_ylabel(Y_LABEL)
ax.legend(loc='best', fancybox=True)
# ax.get_yaxis().set_major_formatter(
# matplotlib.ticker.FuncFormatter(lambda x, p: format(int(x/1000000), ',')))
ax.get_xaxis().set_major_formatter(
matplotlib.ticker.FuncFormatter(
lambda x, p: format(int(x * 100), ',')))
pp.savefig(dirn + "/" + fname + ".pdf")
pp.show()
def main(dirn, fname):
(xs, ysPerSolver, ydevsPerSolver) = CommonViz.parseData(dirn, fname)
CommonConf.setupMPPDefaults()
fmts = CommonConf.getLineFormats()
mrkrs = CommonConf.getLineMarkers()
fig = pp.figure()
ax = fig.add_subplot(111)
# ax.set_xscale("log", basex=2)
index = 0
for (solver, ys), (solver, ydevs) in zip(ysPerSolver.iteritems(),ydevsPerSolver.iteritems()) :
ax.errorbar(xs, ys, yerr=ydevs, label=solver, marker=mrkrs[index], linestyle=fmts[index])
index = index + 1
ax.set_xlabel(X_LABEL);
ax.set_ylabel(Y_LABEL);
ax.legend(loc='best', fancybox=True)
ax.get_yaxis().set_major_formatter(
matplotlib.ticker.FuncFormatter(lambda x, p: format(int(x/1000000), ',')))
ax.get_xaxis().set_major_formatter(
matplotlib.ticker.FuncFormatter(lambda x, p: format(int(x * 100), ',')))
pp.savefig(dirn+"/"+fname+".pdf")
pp.show()
def main(dirn, fname):
blockSizes = []
localitiesPerApproach = OrderedDict()
with open(dirn + "/" + fname + ".dat") as fin:
lines = fin.readlines()
for line in lines:
if line.startswith("#"):
continue
(blockSize, approach, locality, nline) = re.split("[\t]", line)
blockSize = float(blockSize)
locality = float(locality)
if approach in localitiesPerApproach:
localitiesPerApproach[approach].append(locality)
else:
localitiesPerApproach[approach] = [locality]
if len(blockSizes) == 0 or blockSize > blockSizes[-1]:
blockSizes.append(blockSize)
CommonConf.setupMPPDefaults()
fmts = CommonConf.getLineFormats()
mrkrs = CommonConf.getLineMarkers()
fig = pp.figure()
ax = fig.add_subplot(111)
#ax.set_xscale("log", basex=2)
index = 0
for approach, localities in localitiesPerApproach.iteritems():
ax.plot(blockSizes,
localities,
label=approach,
marker=mrkrs[index],
linestyle=fmts[index])
index = index + 1
ax.set_xlabel('block size (bytes)')
ax.set_ylabel('locality')
# ax.set_xlim(0, 2100)
ax.legend(loc='best', fancybox=True)
pp.savefig(dirn + "/" + fname + ".pdf")
pp.show()
def main(dirn, fname, solvers):
(xs, ysPerSolver, ydevsPerSolver) = CommonViz.parseData(dirn, fname, solvers)
CommonConf.setupMPPDefaults()
fmts = CommonConf.getLineFormats()
mrkrs = CommonConf.getLineMarkers()
fig = pp.figure()
ax = fig.add_subplot(111)
# ax.set_xscale("log", basex=2)
index = 0
for (solver, ys), (solver, ydevs) in zip(ysPerSolver.iteritems(),ydevsPerSolver.iteritems()) :
ax.errorbar(xs, ys, yerr=ydevs, label=solver, marker=mrkrs[index], linestyle=fmts[index])
index = index + 1
ax.set_xlabel(X_LABEL);
ax.set_ylabel(Y_LABEL);
ax.legend(loc='best', fancybox=True)
pp.savefig(dirn+"/"+fname+"-".join(solvers)+".svg")
def main(dirn, fname):
blockSizes = []
localitiesPerApproach = OrderedDict()
with open(dirn+"/"+fname+".dat") as fin:
lines = fin.readlines()
for line in lines:
if line.startswith("#"):
continue
(blockSize, approach, locality, nline) = re.split("[\t]", line)
blockSize = float(blockSize)
locality = float(locality)
if approach in localitiesPerApproach:
localitiesPerApproach[approach].append(locality)
else:
localitiesPerApproach[approach] = [locality]
if len(blockSizes) == 0 or blockSize > blockSizes[-1]:
blockSizes.append(blockSize)
CommonConf.setupMPPDefaults()
fmts = CommonConf.getLineFormats()
mrkrs = CommonConf.getLineMarkers()
fig = pp.figure()
ax = fig.add_subplot(111)
#ax.set_xscale("log", basex=2)
index = 0
for approach, localities in localitiesPerApproach.iteritems():
ax.plot(blockSizes, localities, label=approach,
marker=mrkrs[index], linestyle=fmts[index])
index = index + 1
ax.set_xlabel('block size (bytes)');
ax.set_ylabel('locality');
# ax.set_xlim(0, 2100)
ax.legend(loc='best', fancybox=True)
pp.savefig(dirn+"/"+fname+".pdf")
pp.show()
def plotfigure(metric_name, dotfile, mode):
curf = "expData"
todofolder = [
f for f in listdir(curf)
if op.isdir(op.join(curf, f)) & ("burst_" + mode + "_" + dotfile in f)
]
mydict = {}
curname = 0
totiter = 0
xx = []
for folder in todofolder:
print(folder)
xx.append(100 * float(folder[folder.rfind('_') + 1:]))
curname += 1
picked_file = [
f for f in listdir(op.join(curf, folder))
if op.isfile(op.join(curf, folder, f))
& (metric_name + '.dat' in f) & (not 'swp' in f)
]
if (len(picked_file) > 1):
assert (False)
picked_file = picked_file[0]
f = open(op.join(curf, folder, picked_file))
for line in f.readlines():
if "solver" in line:
continue
words = line.split("\t")
if (len(words) < 4):
continue
if (not words[0] in mydict):
mydict[words[0]] = [[] for i in range(len(todofolder))]
if (int(words[1]) > totiter):
totiter = int(words[1])
mydict[words[0]][curname - 1].append(float(words[2]))
mean_list = {}
var_list = {}
for alg in mydict:
mean_list[alg] = OrderedDict()
var_list[alg] = OrderedDict()
i = 0
for all_num in mydict[alg]:
#mean_list[alg][xx[i]]=numpy.mean(numpy.array(all_num), axis=0)
#var_list [alg][xx[i]]=numpy.std(numpy.array(all_num), axis=0)
mean_list[alg][xx[i]] = all_num[flash_time]
var_list[alg][xx[i]] = 0
i += 1
totiter += 1
fig = pp.figure(figsize=(18, 9))
dotted = 0
m = CommonConf.getLineMarkers()
linest = CommonConf.getLineFormats()
colors = CommonConf.getLineColors()
for alg in sorted(mydict):
print(alg, ':')
for key in sorted(mean_list[alg]):
print(key, mean_list[alg][key])
ml = []
kl = []
vl = []
for key in sorted(mean_list[alg]):
kl.append(key)
ml.append(mean_list[alg][key])
vl.append(var_list[alg][key])
xs = np.asarray(kl) + random.random() * 3
pp.errorbar(xs,
ml,
vl,
label=alg,
linestyle=linest[dotted],
marker=m[dotted],
color=colors[dotted])
dotted += 1
x1, x2, y1, y2 = pp.axis()
pp.axis((x1, x2 + 0.1, y1, y2 + 0.05))
pp.legend(loc='best', prop={'size': 14}, bbox_to_anchor=(1.02, 1))
pp.tight_layout(pad=20)
pp.xlabel('Burst amount')
pp.ylabel(metric_name)
pp.savefig("expData/burst_" + dotfile + "_" + metric_name + ".svg")
pp.close(fig)
def plotfigure(metric_name,dotfile):
curf="expData";
todofolder=[f for f in listdir(curf) if op.isdir(op.join(curf, f)) & ("prediction_"+dotfile in f) ]
mydict={}
curname=0;
totiter=0
xx=[]
for folder in todofolder:
xx.append(100*float(folder[len("prediction_"+dotfile):]))
curname += 1
picked_file=[f for f in listdir(op.join(curf,folder)) if op.isfile(op.join(curf,folder, f)) & (metric_name+'.dat' in f) & (not 'swp' in f)]
if (len(picked_file)>1):
assert(False)
picked_file=picked_file[0]
f=open(op.join(curf,folder,picked_file))
for line in f.readlines():
if "solver" in line:
continue
words=line.split("\t")
if (len(words)<4):
continue
if (not words[0] in mydict):
mydict[words[0]]=[[] for i in range(len(todofolder))]
if (int(words[1])>totiter):
totiter=int(words[1])
mydict[words[0]][curname-1].append(float(words[2]))
mean_list={}
var_list={}
for alg in mydict:
mean_list[alg]=OrderedDict()
var_list[alg]=OrderedDict()
i=0
for all_num in mydict[alg]:
mean_list[alg][xx[i]]=numpy.mean(numpy.array(all_num), axis=0)
var_list [alg][xx[i]]=numpy.std(numpy.array(all_num), axis=0)
i+=1
totiter+= 1
fig=pp.figure(figsize=(18,9))
dotted=0
m=CommonConf.getLineMarkers()
linest=CommonConf.getLineFormats()
colors=CommonConf.getLineColors()
for alg in sorted(mydict):
print(alg,':')
for key in sorted(mean_list[alg]):
print(key,mean_list[alg][key])
ml=[]
kl=[]
vl=[]
for key in sorted(mean_list[alg]):
kl.append(key)
ml.append(mean_list[alg][key])
vl.append(var_list[alg][key])
xs=np.asarray(kl)+random.random()*5
pp.errorbar(xs,ml,vl,label=alg, linestyle=linest[dotted], marker=m[dotted],color=colors[dotted])
dotted+=1
x1,x2,y1,y2 = pp.axis()
pp.axis((x1,x2+0.1,y1,y2+0.05))
pp.legend(loc='best',prop={'size':14},bbox_to_anchor=(1.02, 1))
pp.tight_layout(pad=20)
pp.xlabel('Random perturbation %')
pp.ylabel(metric_name)
pp.savefig("expData/prediction_"+dotfile+"_"+metric_name+".svg")
#pp.show()
pp.close(fig)