def median(self, collection):
try:
self.allTimers = collection
self.stat = median(collection)
except StatisticsError:
self.stat = 0
return self.stat
def apply(log, parameters=None):
"""
Measure performance between couples of attributes in the DFG graph
Parameters
----------
log
Log
parameters
Possible parameters passed to the algorithms:
aggregationMeasure -> performance aggregation measure (min, max, mean, median)
activity_key -> Attribute to use as activity
timestamp_key -> Attribute to use as timestamp
Returns
-------
dfg
DFG graph
"""
if parameters is None:
parameters = {}
activity_key = parameters[
PARAMETER_CONSTANT_ACTIVITY_KEY] if PARAMETER_CONSTANT_ACTIVITY_KEY in parameters else DEFAULT_NAME_KEY
timestamp_key = parameters[
PARAMETER_CONSTANT_TIMESTAMP_KEY] if PARAMETER_CONSTANT_TIMESTAMP_KEY in parameters else DEFAULT_TIMESTAMP_KEY
aggregation_measure = parameters[
"aggregationMeasure"] if "aggregationMeasure" in parameters else "mean"
dfgs0 = map((lambda t: [(
(t[i - 1][activity_key], t[i][activity_key]),
(t[i][timestamp_key] - t[i - 1][timestamp_key]).total_seconds())
for i in range(1, len(t))]), log)
ret0 = {}
for el in dfgs0:
for couple in el:
if not couple[0] in ret0:
ret0[couple[0]] = []
ret0[couple[0]].append(couple[1])
ret = Counter()
for key in ret0:
if aggregation_measure == "median":
ret[key] = median(ret0[key])
elif aggregation_measure == "min":
ret[key] = min(ret0[key])
elif aggregation_measure == "max":
ret[key] = max(ret0[key])
elif aggregation_measure == "stdev":
ret[key] = stdev(ret0[key])
elif aggregation_measure == "sum":
ret[key] = sum(ret0[key])
else:
ret[key] = mean(ret0[key])
return ret
def get_median_case_duration(list_cases, timestamp_key=xes.DEFAULT_TIMESTAMP_KEY):
"""
Gets the median case duration of a list of cases
Parameters
-------------
list_cases
List of cases
timestamp_key
Attribute of the event to use as timestamp
Returns
-------------
median_case_duration
Median case duration
"""
durations = []
for trace in list_cases:
durations.append(get_case_duration(trace, timestamp_key=timestamp_key))
return median(durations)