def compute_metric(self):
bad_hosts=tspl_utils.lost_data(self.ts)
if len(bad_hosts) > 0:
print(self.ts.j.id, ': Detected hosts with bad data: ', bad_hosts)
return
vals=[]
for i in [x + 2 for x in range(self.ts.size-4)]:
vals.append(self.compute_fit_params(i))
vals2=[]
for v in vals:
vals2.append([ b/a for (a,b) in v])
arr=numpy.array(vals2)
brr=numpy.transpose(arr)
(m,n)=numpy.shape(brr)
r=[]
for i in range(m):
jnd=numpy.argmin(brr[i,:])
r.append((jnd,brr[i,jnd]))
for (ind,ratio) in r:
self.metric = min(ratio,self.metric)
return
def compute_metric(self):
if len(tspl_utils.lost_data(self.ts)) > 0:
print(self.ts.j.id, ': Detected hosts with bad data')
return
self.tmid=(self.ts.t[:-1]+self.ts.t[1:])/2.0
self.dt = numpy.diff(self.ts.t)
#skip first and last two time slices
vals=[]
for i in [x + 2 for x in range(self.ts.size-4)]:
vals.append(self.compute_fit_params(i))
#times hosts ---->
# |
# |
# |
# V
self.metric = numpy.array(vals).min()
return
def compute_metric(self):
if len(tspl_utils.lost_data(self.ts)) > 0:
print(self.ts.j.id, ': Detected hosts with bad data')
return
self.tmid = (self.ts.t[:-1] + self.ts.t[1:]) / 2.0
self.dt = numpy.diff(self.ts.t)
#skip first and last two time slices
vals = []
for i in [x + 2 for x in range(self.ts.size - 4)]:
vals.append(self.compute_fit_params(i))
#times hosts ---->
# |
# |
# |
# V
try:
self.metric = numpy.array(vals).min()
except:
pass
return