def nucleus_find_cand(dict_aword, nucleus_threshold):
dict_occ_cand = {}
for shortshape, windows in dict_aword.items():
count_s1 = 0 #Meme mot schema et même CAND
count_s2 = 0 #Meme mot schema et CAND differents
count_s3 = 0 #Mot schema different et même CAND
count_s4 = 0 #Mot schema different et CAND different
for window in windows:
linkword = ana_useful.which_linkword(window)
cand = ana_useful.which_cand(window)
for window1 in windows:
if window1 != window:
linkword1 = ana_useful.which_linkword(window1)
cand1 = ana_useful.which_cand(window1)
if linkword[1] == linkword1[1] and cand[2] == cand1[2]:
count_s1 += 1
elif linkword[1] == linkword1[1] and cand[2] != cand1[2]:
count_s2 += 1
elif linkword[1] != linkword1[1] and cand[2] == cand1[2]:
count_s3 += 1
elif linkword[1] != linkword1[1] and cand[2] != cand1[2]:
count_s4 += 1
if count_s1 / 2 >= nucleus_threshold[
0] or count_s2 / 2 >= nucleus_threshold[
1] or count_s3 / 2 >= nucleus_threshold[
2] or count_s4 / 2 >= nucleus_threshold[3]:
for window in windows:
for occurrence in window:
if occurrence[2] == 't':
dict_occ_cand.setdefault(shortshape,
[]).append(occurrence)
return dict_occ_cand
def nucleus_find_cand(dict_aword, nucleus_threshold):
dict_occ_cand = {}
for shortshape, windows in dict_aword.items():
count_s1 = 0 #Meme mot schema et même CAND
count_s2 = 0 #Meme mot schema et CAND differents
count_s3 = 0 #Mot schema different et même CAND
count_s4 = 0 #Mot schema different et CAND different
for window in windows:
linkword = ana_useful.which_linkword(window)
cand = ana_useful.which_cand(window)
for window1 in windows:
if window1 != window:
linkword1 = ana_useful.which_linkword(window1)
cand1 = ana_useful.which_cand(window1)
if linkword[1] == linkword1[1] and cand[2] == cand1[2]:
count_s1 += 1
elif linkword[1] == linkword1[1] and cand[2] != cand1[2]:
count_s2 += 1
elif linkword[1] != linkword1[1] and cand[2] == cand1[2]:
count_s3 += 1
elif linkword[1] != linkword1[1] and cand[2] != cand1[2]:
count_s4 += 1
if count_s1/2 >= nucleus_threshold[0] or count_s2/2 >= nucleus_threshold[1] or count_s3/2 >= nucleus_threshold[2] or count_s4/2 >= nucleus_threshold[3]:
for window in windows:
for occurrence in window:
if occurrence[2] == 't':
dict_occ_cand.setdefault(shortshape, []).append(occurrence)
return dict_occ_cand
def expression_valid_windows(windows, candidate):
valid_window = []
valid_windows = []
buff = []
for window in windows:
if (ana_useful.exists_linkword(window) == True
and ana_useful.count_cand(window) == 2):
if ana_useful.is_cand(
window[-1]): #list[-1] returns last item of the list
cand2 = ana_useful.which_cand([window[-1]])
if cand2[2] not in candidate.split(
): #to avoid building expression like "bâtiment de cet ensemble de bâtiments" -> "batiment de bâtiment"
buff.append(
window
) #dans ce cas la fenetre valide est de type (CAND1 + "aword" + CAND2) avec un mot de schéma quelque part.
# in the buffer because we need to know if the aword in center is not the same 3 times or more (this case it would be better to build an expansion first, then an expression)
else:
short_window = ana_useful.cut_window(window, 2)
#Puisqu'on a 2 CAND et que la fenetre fait 3 mots et que le dernier mot n'est pas un CAND alors la fenetre était de type CAND + CAND + mot quelconque
if ana_useful.exists_linkword(short_window) == True:
valid_windows.append(
short_window
) #dans ce cas la fenetre valide est de type (CAND1 + CAND2) avec un mot de schéma entre eux .
# check if the aword in center is not the same 3 times or more (this case it would be better to build an expansion first, then an expression)
valid_windows.extend(not_expa_inside_expre(buff))
return valid_windows
def expression_valid_windows(windows, candidate):
valid_window = []
valid_windows = []
buff = []
for window in windows:
if (ana_useful.exists_linkword(window) == True and ana_useful.count_cand(window) == 2):
if ana_useful.is_cand(window[-1]): #list[-1] returns last item of the list
cand2 = ana_useful.which_cand([window[-1]])
if cand2[2] not in candidate.split(): #to avoid building expression like "bâtiment de cet ensemble de bâtiments" -> "batiment de bâtiment"
buff.append(window) #dans ce cas la fenetre valide est de type (CAND1 + "aword" + CAND2) avec un mot de schéma quelque part.
# in the buffer because we need to know if the aword in center is not the same 3 times or more (this case it would be better to build an expansion first, then an expression)
else:
short_window = ana_useful.cut_window(window, 2)
#Puisqu'on a 2 CAND et que la fenetre fait 3 mots et que le dernier mot n'est pas un CAND alors la fenetre était de type CAND + CAND + mot quelconque
if ana_useful.exists_linkword(short_window) == True:
valid_windows.append(short_window) #dans ce cas la fenetre valide est de type (CAND1 + CAND2) avec un mot de schéma entre eux .
# check if the aword in center is not the same 3 times or more (this case it would be better to build an expansion first, then an expression)
valid_windows.extend(not_expa_inside_expre(buff))
return valid_windows