def GetTestSents(request):
nbr = request.POST['nbrTestSents'].strip()
if nbr != "":
try:
tool = MyToolKit()
hdb = DBHandler('data/model.db')
result = {}
hdb.connect()
cond = ' '
for i in range(int(nbr)):
if i < int(nbr)-1 : cond += "id="+str(random.randint(1,36111))+" or "
else : cond += "id="+str(random.randint(1,36111))
DataSents = hdb.getFromTable('sents_test',attribute='sent',condition=cond)
#pprint(sents)
sents_diac = ''
for sent in DataSents:
sents_diac += ' '.join(tool.words(sent[0])[1:len(tool.words(sent[0]))-1])+'\n'
sents = tool.DeleteDiacritic(sents_diac)
result['sents_diac'] = sents_diac
result['sents_whitout_diac'] = sents
except ValueError: result = None
#r = random.randint(1,100)
else : result = None
return HttpResponse(json.dumps(result),content_type="application/json")
def LettersVocaliser(self,sents,smooth_const):
tool = MyToolKit()
hdb = DBHandler('data/model.db')
hdb.connect()
res = hdb.getFromTable('letters_dictionary',attribute='type,vocabularies')
dict = {}
for r in res: dict[r[0]] = r[1]
not_vocalised = self.getNotVocalised(sents)
#parcourir les phrases
for i in range(len(sents)):
#pprint(not_vocalised[i])
#parcourir les chaines non vocalisé dans une phrase
for j in range(len(not_vocalised[i])):
matrice = []
char = tool.LettersDiac(not_vocalised[i][j])
list_dict = []
list_dict.append([-1,tool.HideChar(char[0],expect=['#',' ']),1])
matrice.append(list_dict)
list_dict = []
list_dict.append([0,tool.HideChar(char[1],expect=['#',' ']),1])
matrice.append(list_dict)
list_dict = []
list_dict.append([0,"#",1])
matrice.append(list_dict)
string = tool.DeleteDiacritic(not_vocalised[i][j])
#pprint(string)
#parcourir les caractéres
#if(string[0] == 'ا' and string[1] == 'ل')
k = 3
while k < len(string):
list_dict = []
if string[k-1] == "#" and string[k] == 'ا' and string[k+1] == 'ل':
list_dict = []
list_dict.append([0,'_',1])
matrice.append(list_dict)
list_dict = []
list_dict.append([0,'_ْ',1])
matrice.append(list_dict)
k += 2
else :
#parcourir les possibilités
for possib in tool.words(dict[string[k]]):
if possib == "#":list_dict.append([-1,possib,1])
else : list_dict.append([-1,tool.HideChar(possib,expect=['#',' ']),0])
matrice.append(list_dict)
k += 1
v = self.ViterbiLetter(matrice,smooth_const)
string = self.alignLetter(v,not_vocalised[i][j])
#pprint(string)
#pprint(not_vocalised[i][j])
#pprint(sents[i])
#not_vocalised[i] = not_vocalised[i].replace(not_vocalised[i][j],string)
sents[i] = sents[i].replace(not_vocalised[i][j].replace('#',' ').strip(),string.replace('#',' ').strip())
"""
if v[:2] == "##" : n = 0
else: n = 1
string1 = re.sub('#+',' ',not_vocalised[i][j]).strip().split(' '))
string2 = re.sub('#+',' ',v).strip().split(' ')
pprint([0])
pprint(string)
pprint('------------')
"""
#print(not_vocalised[i][j]+' => '+v)
#exit()
#pprint('---------------------------------')
#pprint(matrice)
"""
for a in matrice:
#matrice[i-1][k][1]
pprint(a[0][0])
exit()
"""
return sents
def LettersVocaliser(self, sents, smooth_const):
tool = MyToolKit()
hdb = DBHandler('data/model.db')
hdb.connect()
res = hdb.getFromTable('letters_dictionary',
attribute='type,vocabularies')
dict = {}
for r in res:
dict[r[0]] = r[1]
not_vocalised = self.getNotVocalised(sents)
#parcourir les phrases
for i in range(len(sents)):
#pprint(not_vocalised[i])
#parcourir les chaines non vocalisé dans une phrase
for j in range(len(not_vocalised[i])):
matrice = []
char = tool.LettersDiac(not_vocalised[i][j])
list_dict = []
list_dict.append(
[-1, tool.HideChar(char[0], expect=['#', ' ']), 1])
matrice.append(list_dict)
list_dict = []
list_dict.append(
[0, tool.HideChar(char[1], expect=['#', ' ']), 1])
matrice.append(list_dict)
list_dict = []
list_dict.append([0, "#", 1])
matrice.append(list_dict)
string = tool.DeleteDiacritic(not_vocalised[i][j])
#pprint(string)
#parcourir les caractéres
#if(string[0] == 'ا' and string[1] == 'ل')
k = 3
while k < len(string):
list_dict = []
if string[k - 1] == "#" and string[k] == 'ا' and string[
k + 1] == 'ل':
list_dict = []
list_dict.append([0, '_', 1])
matrice.append(list_dict)
list_dict = []
list_dict.append([0, '_ْ', 1])
matrice.append(list_dict)
k += 2
else:
#parcourir les possibilités
for possib in tool.words(dict[string[k]]):
if possib == "#": list_dict.append([-1, possib, 1])
else:
list_dict.append([
-1,
tool.HideChar(possib, expect=['#', ' ']), 0
])
matrice.append(list_dict)
k += 1
v = self.ViterbiLetter(matrice, smooth_const)
string = self.alignLetter(v, not_vocalised[i][j])
#pprint(string)
#pprint(not_vocalised[i][j])
#pprint(sents[i])
#not_vocalised[i] = not_vocalised[i].replace(not_vocalised[i][j],string)
sents[i] = sents[i].replace(
not_vocalised[i][j].replace('#', ' ').strip(),
string.replace('#', ' ').strip())
"""
if v[:2] == "##" : n = 0
else: n = 1
string1 = re.sub('#+',' ',not_vocalised[i][j]).strip().split(' '))
string2 = re.sub('#+',' ',v).strip().split(' ')
pprint([0])
pprint(string)
pprint('------------')
"""
#print(not_vocalised[i][j]+' => '+v)
#exit()
#pprint('---------------------------------')
#pprint(matrice)
"""
for a in matrice:
#matrice[i-1][k][1]
pprint(a[0][0])
exit()
"""
return sents