def __eq__(self, other):
if self.tag != other.tag:
return False
if self.tag == "script":
if self.src != 'None' and self.src == other.src:
#print "DEBUG: euqal external scripts"
return True
elif self.src != 'None':
#print "DEBUG: NOT equal external scripts"
return False
elif self.src == 'None' and self.src != other.src:
#print "DEBUG: one external one internal scripts"
return False
else: # both src are None
#print "DEBUG: compare two internal scripts"
if 'json' in self.tp and 'json' in other.tp:
try:
#print "JSON: %s " %self.val
obj1 = json.loads(self.val)
keys1 = set(obj1.keys())
obj2 = json.loads(other.val)
keys2 = set(obj2.keys())
intersect = keys1.intersection(keys2)
if len(intersect) >= 2 and \
len(intersect) > len(keys1)/2:
#print "RESULT: TRUE"
return True
else:
#print "RESULT FALSE %d %d %d" %(len(intersect), len(keys1), len(keys2))
return False
except Exception as e:
print "failed parse JSON: %s " % str(e)
return compare_two_string(self.val, other.val)
if self.tag == "link":
if self.href != 'None' and self.href == other.src:
return True
elif self.href != 'None':
return False
elif self.href == 'None' and self.href != other.href:
return False
else:
self_data = "%s-%s" % (self.tp, self.data_reactid)
other_data = "%s-%s" % (other.tp, other.data_reactid)
return self_data == other_data
self_data = "%s-%s" % (self.cls, self.val)
other_data = "%s-%s" % (other.cls, other.val)
return self_data == other_data
def __eq__(self, other): if self.tag != other.tag: return False if self.tag == "script": if self.src != 'None' and self.src == other.src: #print "DEBUG: euqal external scripts" return True elif self.src != 'None': #print "DEBUG: NOT equal external scripts" return False elif self.src == 'None' and self.src != other.src: #print "DEBUG: one external one internal scripts" return False else: # both src are None #print "DEBUG: compare two internal scripts" if 'json' in self.tp and 'json' in other.tp: try: #print "JSON: %s " %self.val obj1 = json.loads(self.val) keys1 = set(obj1.keys()) obj2 = json.loads(other.val) keys2 = set(obj2.keys()) intersect = keys1.intersection(keys2) if len(intersect) >= 2 and \ len(intersect) > len(keys1)/2: #print "RESULT: TRUE" return True else: #print "RESULT FALSE %d %d %d" %(len(intersect), len(keys1), len(keys2)) return False except Exception as e: print "failed parse JSON: %s " %str(e) return compare_two_string(self.val, other.val) if self.tag == "link": if self.href != 'None' and self.href == other.src: return True elif self.href != 'None': return False elif self.href == 'None' and self.href != other.href: return False else: self_data = "%s-%s" % (self.tp, self.data_reactid) other_data = "%s-%s" % (other.tp, other.data_reactid) return self_data == other_data self_data = "%s-%s" % (self.cls, self.val) other_data = "%s-%s" % (other.cls,other.val) return self_data == other_data
def mmdiffR(src_ld, dst_ld, D, \
src_script_hosts, src_script_contents, \
dst_script_hosts, dst_script_contents):
M = len(src_ld) #[1 ... M-1]
N = len(dst_ld) #[1 ... N-1]
i = M - 1
j = N - 1
ins_scripts_hosts = set()
del_scripts_hosts = set()
ins_scripts_contents = set()
del_scripts_contents = set()
updated_scripts_count = 0
while i > 0 and j > 0:
if (D[i][j] == D[i-1][j] + costDelete(src_ld[i])) and \
(j == N-1 or dst_ld[j+1].level <= src_ld[i].level ):
print "DEL:", src_ld[i].tag
if src_ld[i].tag == "script":
if src_ld[i].src != 'None':
del_scripts_hosts.add(src_ld[i].src)
elif src_ld[i].val != "":
del_scripts_contents.add(src_ld[i].val)
try:
print " DEL:%s" % src_ld[i].toString().replace('\n', '\t')
except Exception as e:
print " Error displaying contents: ", str(e)
i = i - 1
elif (D[i][j] == D[i][j-1] + costInsert(dst_ld[j])) and \
(i == M-1 or src_ld[i+1].level <= dst_ld[j].level):
print "INS:", dst_ld[j].tag
if dst_ld[j].tag == "script":
if dst_ld[j].src != 'None':
ins_scripts_hosts.add(dst_ld[j].src)
elif dst_ld[j].val != "":
ins_scripts_contents.add(dst_ld[j].val)
try:
print " INS:%s" % dst_ld[j].toString().replace('\n', '\t')
except Exception as e:
print " Error displaying contents: ", str(e)
j = j - 1
elif not src_ld[i] == dst_ld[j]:
print "UPD: %s => %s " % (src_ld[i].tag, dst_ld[j].tag)
if src_ld[i].tag == "script" and dst_ld[j].tag == "script":
updated_scripts_count += 1
try:
print " REPLACE SRC:", src_ld[i].toString().replace(
'\n', '\t')
print " DST:", dst_ld[j].toString().replace(
'\n', '\t')
print " RS:", compare_two_string(
src_ld[i].val, dst_ld[j].val), "END"
except Exception as e:
print " Error displaying contents: ", str(e)
i = i - 1
j = j - 1
else:
i = i - 1
j = j - 1
while i > 0:
print "DEL:", src_ld[i].tag
i = i - 1
while j > 0:
print "INS:", dst_ld[j].tag
j = j - 1
#print "insert hosts:"
#for item in ins_scripts_hosts:
# print item
#print "delete hosts"
cost = 0
for item in del_scripts_hosts:
if item in dst_script_hosts:
cost += (SCRIPT_COST - 1)
#print item
for item in del_scripts_contents:
for other in dst_script_contents:
if compare_two_string(item, other):
cost += (SCRIPT_COST - 1)
break
for item in ins_scripts_hosts:
if item in src_script_hosts:
cost += (SCRIPT_COST - 1)
#print item
for item in ins_scripts_contents:
for other in src_script_contents:
if compare_two_string(item, other):
cost += (SCRIPT_COST - 1)
break
print "inserted hosts scripts: %d" % len(ins_scripts_hosts)
print "deleted hosts scripts: %d" % len(del_scripts_hosts)
print "inserted hosts contents: %d" % len(ins_scripts_contents)
print "deleted hosts contents: %d" % len(del_scripts_contents)
print "updated scripts: %d" % updated_scripts_count
'''
print "insert contents:"
for item in ins_scripts_contents:
try:
print item
except Exception as e:
print "Error: ",str(e)
print "delete contentes:"
for item in del_scripts_contents:
try:
print item
except Exception as e:
print "Error: ",str(e)
'''
rs = float(D[M - 1][N - 1] - cost)
norm = float(M + N)
print "Final cost: %f[%f]" % (rs / norm, D[M - 1][N - 1] / norm)
return rs / norm
def mmdiffR(src_ld, dst_ld, D, \
src_script_hosts, src_script_contents, \
dst_script_hosts, dst_script_contents):
M = len(src_ld) #[1 ... M-1]
N = len(dst_ld) #[1 ... N-1]
i = M - 1
j = N - 1
ins_scripts_hosts = set()
del_scripts_hosts = set()
ins_scripts_contents = set()
del_scripts_contents = set()
updated_scripts_count = 0
while i >0 and j > 0:
if (D[i][j] == D[i-1][j] + costDelete(src_ld[i])) and \
(j == N-1 or dst_ld[j+1].level <= src_ld[i].level ):
print "DEL:",src_ld[i].tag
if src_ld[i].tag == "script":
if src_ld[i].src != 'None':
del_scripts_hosts.add(src_ld[i].src)
elif src_ld[i].val != "":
del_scripts_contents.add(src_ld[i].val)
try:
print " DEL:%s" % src_ld[i].toString().replace('\n','\t')
except Exception as e:
print " Error displaying contents: ",str(e)
i = i - 1
elif (D[i][j] == D[i][j-1] + costInsert(dst_ld[j])) and \
(i == M-1 or src_ld[i+1].level <= dst_ld[j].level):
print "INS:",dst_ld[j].tag
if dst_ld[j].tag == "script":
if dst_ld[j].src != 'None':
ins_scripts_hosts.add(dst_ld[j].src)
elif dst_ld[j].val != "":
ins_scripts_contents.add(dst_ld[j].val)
try:
print " INS:%s" % dst_ld[j].toString().replace('\n','\t')
except Exception as e:
print " Error displaying contents: ",str(e)
j = j - 1
elif not src_ld[i] == dst_ld[j]:
print "UPD: %s => %s " % (src_ld[i].tag, dst_ld[j].tag)
if src_ld[i].tag == "script" and dst_ld[j].tag == "script":
updated_scripts_count += 1
try:
print " REPLACE SRC:",src_ld[i].toString().replace('\n','\t')
print " DST:",dst_ld[j].toString().replace('\n','\t')
print " RS:",compare_two_string(src_ld[i].val, dst_ld[j].val), "END"
except Exception as e:
print " Error displaying contents: ",str(e)
i = i - 1
j = j - 1
else:
i = i - 1
j = j - 1
while i > 0:
print "DEL:",src_ld[i].tag
i = i - 1
while j > 0:
print "INS:",dst_ld[j].tag
j = j - 1
#print "insert hosts:"
#for item in ins_scripts_hosts:
# print item
#print "delete hosts"
cost = 0
for item in del_scripts_hosts:
if item in dst_script_hosts:
cost += (SCRIPT_COST-1)
#print item
for item in del_scripts_contents:
for other in dst_script_contents:
if compare_two_string(item, other):
cost += (SCRIPT_COST-1)
break
for item in ins_scripts_hosts:
if item in src_script_hosts:
cost += (SCRIPT_COST-1)
#print item
for item in ins_scripts_contents:
for other in src_script_contents:
if compare_two_string(item, other):
cost += (SCRIPT_COST-1)
break
print "inserted hosts scripts: %d" % len(ins_scripts_hosts)
print "deleted hosts scripts: %d" % len(del_scripts_hosts)
print "inserted hosts contents: %d" % len(ins_scripts_contents)
print "deleted hosts contents: %d" % len(del_scripts_contents)
print "updated scripts: %d" % updated_scripts_count
'''
print "insert contents:"
for item in ins_scripts_contents:
try:
print item
except Exception as e:
print "Error: ",str(e)
print "delete contentes:"
for item in del_scripts_contents:
try:
print item
except Exception as e:
print "Error: ",str(e)
'''
rs = float(D[M-1][N-1] - cost)
norm = float(M + N)
print "Final cost: %f[%f]" %(rs/norm, D[M-1][N-1]/norm)
return rs/norm
