def parse_fsm_file(fsm_file,prefix4states=''):
with open(fsm_file,'r') as reader:
lines=[l.strip() for l in reader]
index = 0
num_startings = int(lines[index])
index+=1
startings = []
for _ in range(num_startings):
startings+=[prefix4states+lines[index]]
index+=1
num_endings = int(lines[index])
index+=1
endings = []
for _ in range(num_endings):
endings+=[prefix4states+lines[index]]
index+=1
num_edges = int(lines[index])
index+=1
edges=[]
for _ in range(num_edges):
source,dest,label = lines[index].split()
index+=1
edges+=[(prefix4states+source,prefix4states+dest,label)]
return StandardAutomata(startings,edges,endings)
def extending_ending_states(fsm, ending_methods):
startings = set(fsm.startings)
endings = set(fsm.endings)
edges = set(fsm.transitions)
for (source, dest, label) in fsm.transitions:
if label in ending_methods:
for end in endings:
edges.add((dest, end, lib.ending_char()))
return StandardAutomata(startings, edges, endings)
def make_fsm(startings, endings, edges, id2cluster, generated_traces):
log = {}
for one_trace in generated_traces:
previous_cluster = None
previous_label = None
previous_prob_id = None
for (prob_id, word) in one_trace:
if prob_id not in id2cluster or id2cluster[prob_id] == -1:
print(
"Skipping due to missing prob_id in id-to-cluster mapping")
continue
cluster_name = id2cluster[prob_id]
if is_constructor(word):
cluster_name = 'CSTART'
# if word == ending_char():
# cluster_name = ending_cluster()
if previous_cluster is not None:
edges.add((previous_cluster, cluster_name, previous_label))
the_edge = (previous_cluster, cluster_name, previous_label)
if the_edge not in log:
log[the_edge] = set()
log[the_edge].add((str(previous_prob_id) if previous_prob_id
is not None else 'None', str(prob_id)))
else:
startings.add(cluster_name)
previous_cluster = cluster_name
previous_label = word
previous_prob_id = prob_id
###
if previous_label == ending_char():
edges.add((previous_cluster, ending_cluster(), previous_label))
return StandardAutomata(startings, edges, endings), log
def create_fsm(id2cluster, generated_traces):
startings = set()
endings = set()
edges = set()
endings.add(ending_cluster())
log = {}
for one_trace in generated_traces:
previous_cluster = None
previous_label = None
previous_prob_id = None
for (prob_id, word) in one_trace:
cluster_name = id2cluster[prob_id]
if is_constructor(word):
cluster_name = 'CSTART'
# if word == ending_char():
# cluster_name = ending_cluster()
if previous_cluster is not None:
edges.add((previous_cluster, cluster_name, previous_label))
the_edge = (previous_cluster, cluster_name, previous_label)
if the_edge not in log:
log[the_edge] = set()
log[the_edge].add((str(previous_prob_id) if previous_prob_id
is not None else 'None', str(prob_id)))
else:
startings.add(cluster_name)
previous_cluster = cluster_name
previous_label = word
previous_prob_id = prob_id
###
if previous_label == ending_char():
edges.add((previous_cluster, ending_cluster(), previous_label))
return StandardAutomata(startings, edges, endings), log