#check if pattern invalid. if yes there is no point in calculation / calculate anyway for to be safe. but output file invalid.info
if os.path.exists(os.path.join(args.pattern_path,"invalid.info")):
print "Pattern is invalid... problem with sampling approaches"
with open(os.path.join(output_path,"invalid.info"),"w") as f:
f.write("invalid pattern "+args.pattern_path)
#writing input gml into output
if(not os.path.exists(output_path)):
os.makedirs(output_path)
nx.write_gml(pattern, output_path+'/input_pattern.gml')
#choose or load root node and nr observations
hist=analyzer.get_sorted_labels_by_occurence_frequency_in_graph(args.data_graph_path)
if not os.path.exists(os.path.join(args.output_path,'root_node.dec')):
hist=analyzer.get_sorted_labels_by_occurence_frequency_in_graph(args.data_graph_path)
root_node,root_predicate_name=u.choose_root_node(pattern,None,hist)
with open(os.path.join(args.output_path,'root_node.dec'),'w') as f:
f.write(str(root_node)+" "+root_predicate_name.rstrip().lstrip()+"\n")
f.write("Chosen by furer during the selection ...")
else: #else, it is not selection so root node was already decided by someone
with open(os.path.join(args.output_path,'root_node.dec'),'r') as f:
for line in f.readlines():
root_node=int(line.split(" ")[0])
root_node_predicate_name=str(line.split(" ")[1].rstrip().lstrip())
break
all_randnode_times = []
rndicts = []
exhaustive_approach_results_file=os.path.join(args.exhaustive_approach_results_path,'results_'+pattern_file_name+'.res')
if os.path.exists(exhaustive_approach_results_file):
all_randnode_times = []
def get_nr_emb_within_time(data_graph_path, pattern_path, output_path,
time_seconds):
print "Exhaustive checkup ...."
nr_emb = None
monitoring_marks = utils.generate_monitoring_marks(time_seconds,
time_seconds)
data_graph = None
try:
data_graph = nx.read_gpickle(data_graph_path)
except:
data_graph = nx.read_gml(data_graph_path)
number_of_nodes_in_data = len(data_graph)
pattern = nx.read_gml(pattern_path)
#vis.visualize_graph(pattern, "sat")
#analyzer.add_values_in_pattern_for_graph_if_missing(pattern)
output_path = os.path.join(output_path)
if not os.path.exists(output_path):
os.makedirs(output_path)
root_node_predicate_name = None #well, not predefined. Let the algorithm find it and denote it (its id in the pattern)
pattern_name = os.path.basename(pattern_path)[:-4]
# logging.basicConfig(
# level=logging.DEBUG,
# filename=os.path.join(output_path,'error_exhaustive.log'),
# filemode='w')
# sys.excepthook = my_excepthook
root_node = None
#first check if the root node is determined by some other algorithm
if not os.path.exists(os.path.join(output_path, 'root_node.dec')):
hist = analyzer.get_sorted_labels_by_occurence_frequency_in_graph(
data_graph_path)
root_node, root_node_predicate_name = ut.choose_root_node(
pattern, root_node_predicate_name, hist)
with open(os.path.join(output_path, 'root_node.dec'), 'w') as f:
f.write(str(root_node) + " ")
f.write(str(root_node_predicate_name) + "\n")
f.write("Determined by exhaustive approach")
else:
#read root node from the file
with open(os.path.join(output_path, 'root_node.dec'), 'r') as f:
for line in f.readlines():
root_node = int(line.split(" ")[0])
root_node_predicate_name = str(
line.split(" ")[1].rstrip().lstrip())
break
print "root node predicate name: ", root_node_predicate_name
#get root nodes
root_nodes = [
x for x in data_graph.nodes()
if data_graph.node[x]['predicate'] == root_node_predicate_name
]
print "Number of root nodes: ", len(root_nodes)
#get OBD
print "Root node,", pattern.node[root_node]
OBdecomp = OBDsearch.get_heuristic4_OBD(pattern, startNode=root_node)
if OBdecomp == None:
print "No ombdecomp!"
no_obd_decomp = True
with open(os.path.join(output_path, 'no_obdecomp.info'), 'w') as f:
f.write("No OBDecomp!")
OBdecomp = OBDsearch.get_flatList(pattern, startNode=root_node)
#get ordered list from OBD
Plist = [item for sublist in OBdecomp for item in sublist]
print "Using OBD: %s" % str(OBdecomp)
print "and Plist: %s" % str(Plist)
print "monitoring marks: ", monitoring_marks
start = timeit.default_timer()
try:
lock = threading.Lock()
print "starting scheduler"
s = sched.scheduler(time.time, time.sleep)
e1 = s.enter(
0, 4, exhaustive.find_nr_emb,
(data_graph, pattern, Plist, root_nodes, output_path, lock))
t = threading.Thread(target=s.run)
t.daemon = True
t.start()
time.sleep(time_seconds)
end = timeit.default_timer()
print "Main finished after ", end - start, "seconds"
freq_dict = experiments.globals.fdict_exhaustive_limited
if len(freq_dict) == 0:
nr_emb = None
else:
nr_emb = 0
for k in freq_dict.keys():
nr_emb = nr_emb + freq_dict[k]
except Wrong_root_node as e:
print "Exception for the node occurred!"
return nr_emb
#check if pattern invalid. if yes there is no point in calculation / calculate anyway for to be safe. but output file invalid.info
if os.path.exists(os.path.join(args.pattern_path,"invalid.info")):
print "Pattern is invalid... problem with sampling approaches"
with open(os.path.join(output_path,"invalid.info"),"w") as f:
f.write("invalid pattern "+args.pattern_path)
#writing input gml into output
nx.write_gml(pattern, args.output_path+'/input_pattern.gml')
#choose or load root node and nr observations
desired_predicate=None
print "Specified root node? ",os.path.join(args.output_path,'root_node.dec')
if args.sample or not(os.path.exists(os.path.join(args.output_path,'root_node.dec'))):
hist=analyzer.get_sorted_labels_by_occurence_frequency_in_graph(args.data_graph_path)
root_node,root_predicate_name=u.choose_root_node(pattern,desired_predicate,hist)
with open(os.path.join(args.output_path,'root_node.dec'),'w') as f:
f.write(str(root_node)+" "+root_predicate_name.rstrip().lstrip()+"\n")
f.write("Chosen by furer during the selection ...")
else: #else, it is not selection so root node was already decided by someone
with open(os.path.join(args.output_path,'root_node.dec'),'r') as f:
for line in f.readlines():
root_node=int(line.split(" ")[0])
root_node_predicate_name=str(line.split(" ")[1].rstrip().lstrip())
break
#if it's not the selection phase it means the exhaustive approach is run, therefore read NLIMIT values from the file
exhaustive_approach_results_file=os.path.join(args.exhaustive_approach_results_path,'results_'+pattern_file_name+'.res')
if os.path.exists(exhaustive_approach_results_file):
all_randnode_times = []
if (not os.path.exists(output_path)):
os.makedirs(output_path)
if not (os.path.exists(args.output_path + '/input_pattern.gml')):
nx.write_gml(pattern, args.output_path + '/input_pattern.gml')
#DETERMINING ROOT NODE
root_node = None
print "Root node predicate name: ", root_node_predicate_name
print "Exists? ", os.path.join(args.output_path,
'root_node.dec'), os.path.exists(
os.path.join(args.output_path,
'root_node.dec'))
if root_node_predicate_name != None or not os.path.exists(
os.path.join(args.output_path, 'root_node.dec')):
hist = analyzer.get_sorted_labels_by_occurence_frequency_in_graph(
args.data_graph_path)
root_node, root_node_predicate_name = ut.choose_root_node(
pattern, root_node_predicate_name, hist)
with open(os.path.join(args.output_path, 'root_node.dec'), 'w') as f:
f.write(str(root_node) + " ")
f.write(str(root_node_predicate_name) + "\n")
f.write("Determined by exhaustive approach")
else:
#read root node from the file
with open(os.path.join(args.output_path, 'root_node.dec'), 'r') as f:
for line in f.readlines():
root_node = int(line.split(" ")[0])
root_node_predicate_name = str(
line.split(" ")[1].rstrip().lstrip())
break
print "root node predicate name: ", root_node_predicate_name
#get root nodes
root_nodes = [