def get_similarity(self, a, b):
#return self.lcs(a, b)
a = np.array(list(a), dtype='U1').view(np.uint32)
b = np.array(list(b), dtype='U1').view(np.uint32)
length, path = mlpy.lcs_std(a, b)
return length
def get_similarity(self, a, b):
#return self.lcs(a, b)
a = np.array(list(a), dtype='U1').view(np.uint32)
b = np.array(list(b), dtype='U1').view(np.uint32)
length, path = mlpy.lcs_std(a, b)
return length
def longest(ovv,cand):
try:
ovv_int = [char_map[x] for x in ovv.encode('ascii',"ignore").lower()]
cand_int = [char_map[y] for y in cand.encode('ascii',"ignore").lower()]
lcs = mlpy.lcs_std(ovv_int,cand_int)[0]
except Exception as e:
print(ovv,cand,e)
lcs = difflib.SequenceMatcher(None, ovv,cand).find_longest_match(0, len(ovv), 0, len(cand))[2]
return lcs
def caculLCS():
labelPath = GetLabelFile()
for i in range(0, len(labelPath) - 1):
for j in range(i + 1, len(labelPath)):
seqi = label2number(labelPath[i])
namei = labelPath[i].split('\\')[4]
seqj = label2number(labelPath[j])
namej = labelPath[j].split('\\')[4]
length, path = mlpy.lcs_std(seqi, seqj)
Write_LCS_Ans([namei, namej, length])
def main():
with open(sys.argv[1]) as handle:
file = handle.read()
n = int(file.split('\n')[0])
input_list = map(int, file.split('\n')[1].split())
print n # , input_list
x_incre = range(1, n + 1)
x_decre = x_incre[::-1]
y = input_list
incre_path = mlpy.lcs_std(x_incre, y)[1][1]
decre_path = mlpy.lcs_std(x_decre, y)[1][1]
incre_sub_list = [input_list[i] for i in incre_path]
decre_sub_list = [input_list[i] for i in decre_path]
for i in incre_sub_list:
print i,
print
for j in decre_sub_list:
print j,
'''
def caculLCS():
labelPath=GetLabelFile()
for i in range(0,len(labelPath)-1):
for j in range(i+1,len(labelPath)):
seqi=label2number(labelPath[i])
namei=labelPath[i].split('\\')[4]
seqj=label2number(labelPath[j])
namej=labelPath[j].split('\\')[4]
length, path = mlpy.lcs_std(seqi,seqj)
Write_LCS_Ans([namei,namej,length])
def commonSubseq(x, y, table): str1 = [] str2 = [] res = [] for l in x: str1.append(table[l]) for l in y: str2.append(table[l]) length, path = mlpy.lcs_std(str1, str2) for i in path[0]: h = str1[i] for k, v in table.items(): if v == h: res.append(k) return ''.join(res)
def LCS_dist(
self, xx, X
): # xx is the testing instance, X is the training instance (passed one by one)
import mlpy
i, j = int(xx[0]), int(
X[0]) # extract indices, i for testing, j for training
#print "i = " + str(i) + ", j = " +str(j)
length, path = mlpy.lcs_std(self.__Xtest[i], self.__Xtrain[j])
dist_lcs = float(length) / np.sqrt(
len(self.__Xtest[i]) * len(self.__Xtrain[j])
) ## Formula taken from section 4.1.2 in paper: Anomaly Detection for Discrete Sequences: A Survey
if dist_lcs != 0:
dist_lcs_inv = float(1 / float(dist_lcs))
else:
dist_lcs_inv = 9999.0
return dist_lcs_inv
def string_lcs(self, x, y):
int_x = [ord(l) for l in x]
int_y = [ord(l) for l in y]
return mlpy.lcs_std(int_x, int_y)
for index, word in enumerate(allwords):
if word.isupper() and len(word) > 2:
acronym_list.append(''.join(e for e in word if e.isalnum()))
acronym_indices.append(index)
x = PrettyTable(["Acronym", "Definition"])
for index, acronym in enumerate(acronym_list):
acronym_split = []
for c in acronym:
acronym_split.append(ord(c.upper()))
word_window = [
word for word in
allwords[minString(acronym_indices[index] -
10):maxString(acronym_indices[index] +
10, len(allwords))]
]
letter_window = [ord(word[0]) for word in word_window]
length, path = mlpy.lcs_std(letter_window, acronym_split)
printacronym = ''.join(chr(c).upper() for c in acronym_split)
printfull = []
word_path = path[0]
print(word_path)
printfull = [word for word in word_window[word_path[0]:word_path[-1] + 1]]
printfull = ' '.join(printfull)
x.add_row([printacronym, printfull])
print(x)
#print acronym_list
#print acronym_indices
