The Container abstractionΒΆ
All analysis in scared
start with a container. A Container
is an abstraction which is responsible to provide traces and metadata to leakage or analysis objects.
Basically, a container is defined by:
a
TraceHeaderSet
instancea
frame
optionala list of
preprocesses
Once initialized, it can provides batches of traces samples and metadatas on-demand, by applying the preprocess chain just-in-time.
A container is created like this:
import scared
ths = scared.traces.read_ths_from_ets_file('traceset.ets')
container = scared.Container(
ths,
frame=slice(20, 30),
preprocesses=[scared.preprocesses.standardize]
)