def fil_x(words, ent_form, ent_number):
inds = [i for i,x in enumerate(ent_form) if x.isupper()]
if inds:
ent_form2 = ent_form.lower()
else:
ent_form2 = ent_form
if '[X]' in words:
i = words.index('[X]')
words[i] = ent_form
ent_case = "NOM"
elif "[X.Loc]" in words:
i = words.index('[X.Loc]')
temp = inflect(ent_form2, f"N;LOC;{ent_number}", language='tur')[0]
if inds:
words[i] = fix_up(temp, inds)
else:
words[i] = temp
ent_case = "LOC"
elif "[X.Gen]" in words:
i = words.index('[X.Gen]')
temp = inflect(ent_form2, f"N;GEN;{ent_number}", language='tur')[0]
if inds:
words[i] = fix_up(temp, inds)
else:
words[i] = temp
ent_case = "GEN"
elif "[X.Acc]" in words:
i = words.index('[X.Acc]')
temp = inflect(ent_form2, f"N;ACC;{ent_number}", language='tur')[0]
if inds:
words[i] = fix_up(temp, inds)
else:
words[i] = temp
ent_case = "ACC"
elif "[X.Dat]" in words:
i = words.index('[X.Dat]')
temp = inflect(ent_form2, f"N;DAT;{ent_number}", language='tur')[0]
if inds:
words[i] = fix_up(temp, inds)
else:
words[i] = temp
ent_case = "DAT"
elif "[X.Abl]" in words:
i = words.index('[X.Abl]')
temp = inflect(ent_form2, f"N;ABL;{ent_number}", language='tur')[0]
if inds:
words[i] = fix_up(temp, inds)
else:
words[i] = temp
ent_case = "ABL"
if '[X;be]' in words:
i = words.index('[X;be]')
words[i] = add_be(ent_form, ent_number)
return words
def fil_x(words, ent_form, ent_number, is_human):
if some_roman_chars(ent_form) or ent_form.isupper():
do_not_inflect = True
else:
do_not_inflect = False
if '[X]' in words:
i = words.index('[X]')
words[i] = ent_form
ent_case = "NOM"
elif "[X.LOC]" in words:
i = words.index('[X.LOC]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form, f"N;LOC;{ent_number}",
language='ben')[0]
#words[i] = ent_form+'त'
ent_case = "LOC"
elif "[X.GEN]" in words:
i = words.index('[X.GEN]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form, f"N;GEN;{ent_number}",
language='ben')[0]
#words[i] = ent_form+'त'
ent_case = "GEN"
# Now check for the ones that we have a fixed suffix:
for i, w in enumerate(words):
if w[:3] == '[X]' and len(w) > 3:
words[i] = ent_form + w[3:]
# Now also check the corresponfing verbs, if they exist.
# Needed for subject-verb agreement
for i, w in enumerate(words):
if w[0] == '[' and 'human' in w:
if '|' in w:
options = w.strip()[1:-1].split('|')
if is_human:
form = options[0].strip().split(':')[1]
words[i] = form
else:
form = options[1].strip().split(':')[1]
words[i] = form
return words
def fil_x(words, ent_form, ent_number, is_human):
if some_roman_chars(ent_form) or ent_form.isupper():
do_not_inflect = True
else:
do_not_inflect = False
if '[X]' in words:
i = words.index('[X]')
words[i] = ent_form
ent_case = "NOM"
elif "[X.ACC]" in words:
i = words.index('[X.ACC]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form, f"N;ACC;{ent_number}",
language='hun')[0]
#words[i] = ent_form+'त'
ent_case = "ACC"
elif "[X.DAT]" in words:
i = words.index('[X.DAT]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form, f"N;DAT;{ent_number}",
language='hun')[0]
#words[i] = ent_form+'त'
ent_case = "DAT"
elif "[X.ON+ESS]" in words:
i = words.index('[X.ON+ESS]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form,
f"N;ON+ESS;{ent_number}",
language='hun')[0]
ent_case = "ON+ESS"
return words
def fil_y(words, ent_form, ent_number, is_human):
if some_roman_chars(ent_form) or ent_form.isupper():
do_not_inflect = True
else:
do_not_inflect = False
if '[Y]' in words:
i = words.index('[Y]')
words[i] = ent_form
ent_case = "NOM"
elif "[Y.LOC]" in words:
i = words.index('[Y.LOC]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form, f"N;LOC;{ent_number}",
language='ben')[0]
#words[i] = ent_form+'त'
ent_case = "LOC"
elif "[Y.GEN]" in words:
i = words.index('[Y.GEN]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form, f"N;GEN;{ent_number}",
language='ben')[0]
#words[i] = ent_form+'त'
ent_case = "GEN"
# Now check for the ones that we have a fixed suffix:
for i, w in enumerate(words):
if w[:3] == '[Y]' and len(w) > 3:
words[i] = ent_form + w[3:]
return words
def fil_y(words, ent_form, ent_number, is_human):
if some_roman_chars(ent_form) or ent_form.isupper():
do_not_inflect = True
else:
do_not_inflect = False
if '[Y]' in words:
i = words.index('[Y]')
words[i] = ent_form
ent_case = "NOM"
elif "[Y.IN+ESS]" in words:
i = words.index('[Y.IN+ESS]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form,
f"N;IN+ESS;{ent_number}",
language='hun')[0]
ent_case = "IN+ESS"
elif "[Y.IN+ABL]" in words:
i = words.index('[Y.IN+ABL]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form,
f"N;IN+ABL;{ent_number}",
language='hun')[0]
ent_case = "IN+ABL"
elif "[Y.ON+ESS]" in words:
i = words.index('[Y.ON+ESS]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form,
f"N;ON+ESS;{ent_number}",
language='hun')[0]
ent_case = "ON+ESS"
elif "[Y.DAT]" in words:
i = words.index('[Y.DAT]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form, f"N;DAT;{ent_number}",
language='hun')[0]
ent_case = "DAT"
elif "[Y.ACC]" in words:
i = words.index('[Y.ACC]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form, f"N;ACC;{ent_number}",
language='hun')[0]
ent_case = "ACC"
elif "[Y.INST]" in words:
i = words.index('[Y.INST]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form,
f"N;INST;{ent_number}",
language='hun')[0]
ent_case = "INST"
# Now check for the ones that we have a fixed suffix:
for i, w in enumerate(words):
if w[:3] == '[Y]' and len(w) > 3:
words[i] = ent_form + w[3:]
return words
def fil_x(words, ent_form, ent_gender):
#ent_form = entities[ent_id][0]
#ent_gender = entities[ent_id][1].upper()
ent_number = "SG"
if some_roman_chars(ent_form) or ent_form.isupper():
do_not_inflect = True
else:
do_not_inflect = False
if '[X]' in words:
i = words.index('[X]')
words[i] = ent_form
ent_case = "NOM"
elif "[X.Nom]" in words:
# In Greek the default case is Nominative so we don't need to try to inflect it
i = words.index('[X.Nom]')
words[i] = ent_form
ent_case = "NOM"
elif "[X.Masc.Nom]" in words:
# In Greek the default case is Nominative so we don't need to try to inflect it
i = words.index('[X.Masc.Nom]')
words[i] = ent_form
ent_case = "NOM"
ent_gender = "Masc"
elif "[X.Gen]" in words:
i = words.index('[X.Gen]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form, f"N;GEN;{ent_number}", language='rus')[0]
ent_case = "GEN"
elif "[X.Ess]" in words:
i = words.index('[X.Ess]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form, f"N;ESS;{ent_number}", language='rus')[0]
ent_case = "ESS"
# Now also check the corresponfing verbs, if they exist
for i,w in enumerate(words):
if w[0] == '[' and 'X-Gender' in w:
if '|' in w:
options = w.strip()[1:-1].split('|')
if ent_gender == "MASC":
form = options[0].strip().split(';')[0]
words[i] = form
elif ent_gender == "FEM":
form = options[1].strip().split(';')[0]
words[i] = form
elif ent_gender == "NEUT":
form = options[2].strip().split(';')[0]
words[i] = form
else:
form = options[0].strip().split(';')[0]
words[i] = form
#else:
# lemma = w.strip()[1:-1].split('.')[0]
# if "Pst" in w:
# form2 = inflect(lemma, f"V;PST;SG;{ent_gender}", language='rus')[0]
# elif "Lgspec1" in w:
# form2 = inflect(lemma, f"ADJ;{ent_gender};SG;LGSPEC1", language='rus')[0]
# words[i] = form2
return words
def fil_y(words, ent_form, ent_gender):
#ent_form = entities[ent_id][0]
#ent_gender = entities[ent_id][1].upper()
ent_number = "SG"
if some_roman_chars(ent_form) or ent_form.isupper():
do_not_inflect = True
else:
do_not_inflect = False
if '[Y]' in words:
i = words.index('[Y]')
words[i] = ent_form
ent_case = "NOM"
elif "[Y.Nom]" in words:
# In Greek the default case is Nominative so we don't need to try to inflect it
i = words.index('[Y.Nom]')
words[i] = ent_form
ent_case = "NOM"
elif "[Y.Gen]" in words:
i = words.index('[Y.Gen]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form, f"N;GEN;{ent_number}", language='rus')[0]
ent_case = "GEN"
elif "[Y.Acc]" in words:
i = words.index('[Y.Acc]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form, f"N;ACC;{ent_number}", language='rus')[0]
ent_case = "ACC"
elif "[Y.Dat]" in words:
i = words.index('[Y.Dat]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form, f"N;DAT;{ent_number}", language='rus')[0]
ent_case = "DAT"
elif "[Y.Ess]" in words:
i = words.index('[Y.Ess]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form, f"N;ESS;{ent_number}", language='rus')[0]
ent_case = "ESS"
elif "[Y.Ins]" in words:
i = words.index('[Y.Ins]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form, f"N;INS;{ent_number}", language='rus')[0]
ent_case = "INS"
# Now also check the correponsing articles, if the exist
for i,w in enumerate(words):
if w[0] == '[' and 'Y-Gender' in w:
if '|' in w:
options = w.strip()[1:-1].split('|')
if ent_gender == "MASC":
form = options[0].strip().split(';')[0]
words[i] = form
elif ent_gender == "FEM":
form = options[1].strip().split(';')[0]
words[i] = form
elif ent_gender == "NEUT":
form = options[2].strip().split(';')[0]
words[i] = form
if "Pst" in w:
form2 = inflect(lemma, f"V;PST;SG;{ent_gender}", language='rus')[0]
elif "Lgspec1" in w:
form2 = inflect(lemma, f"ADJ;{ent_gender};SG;LGSPEC1", language='rus')[0]
words[i] = form2
return words
def fil_x(words, ent_form, ent_gender, ent_number, article):
#ent_form = entities[ent_id][0]
#ent_gender = entities[ent_id][1].upper()
#ent_number = "SG"
if ent_form[-2:] == "ες":
ent_number = "PL"
ent_gender = "FEM"
if some_roman_chars(ent_form) or ent_form.isupper() or ent_form[-1] in [
'β', 'γ', 'δ', 'ζ', 'κ', 'λ', 'μ', 'ν', 'ξ', 'π', 'ρ', 'τ', 'φ',
'χ', 'ψ'
]:
do_not_inflect = True
else:
do_not_inflect = False
if '[X]' in words:
i = words.index('[X]')
words[i] = ent_form
ent_case = "NOM"
elif "[X.Nom]" in words:
# In Greek the default case is Nominative so we don't need to try to inflect it
i = words.index('[X.Nom]')
words[i] = ent_form
ent_case = "NOM"
elif "[X.Gen]" in words:
i = words.index('[X.Gen]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form,
f"N;GEN;{ent_number}",
language='ell2')[0]
ent_case = "GEN"
elif "[X.Acc]" in words:
i = words.index('[X.Acc]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form,
f"N;ACC;{ent_number}",
language='ell2')[0]
ent_case = "ACC"
# Now also check the correponsing articles, if the exist
if "[DEF;X]" in words:
i = words.index('[DEF;X]')
words[i] = article[f"ART;DEF;{ent_gender};{ent_number};{ent_case}"]
if "[DEF.Gen;X]" in words:
i = words.index('[DEF.Gen;X]')
words[i] = article[f"ART;DEF;{ent_gender};{ent_number};GEN"]
if "[PREPDEF;X]" in words:
i = words.index('[PREPDEF;X]')
words[i] = article[f"ART;PREPDEF;{ent_gender};{ent_number};{ent_case}"]
# Now also check the corresponfing verbs, if they exist.
# Needed for subject-verb agreement
for i, w in enumerate(words):
if w[0] == '[' and 'X-Number' in w:
if '|' in w:
options = w.strip()[1:-1].split('|')
if ent_number == "SG":
form = options[0].strip().split(';')[0]
words[i] = form
else:
form = options[1].strip().split(';')[0]
words[i] = form
return words
def fil_y(words, ent_form, ent_gender, ent_number, article):
#ent_form = entities[ent_id][0]
#ent_gender = entities[ent_id][1].upper()
#ent_number = "SG"
if ent_form[-2:] == "ες" or ent_form[-2:] == "ές":
ent_number = "PL"
ent_gender = "FEM"
#elif ent_form[-1] == "ά":
# ent_number = "PL"
# ent_gender = "NEUT"
if some_roman_chars(ent_form) or ent_form.isupper() or ent_form[-1] in [
'β', 'γ', 'δ', 'ζ', 'κ', 'λ', 'μ', 'ν', 'ξ', 'π', 'ρ', 'τ', 'φ',
'χ', 'ψ'
]:
do_not_inflect = True
else:
do_not_inflect = False
if '[Y]' in words:
i = words.index('[Y]')
words[i] = ent_form
ent_case = "NOM"
elif "[Y.Nom]" in words:
# In Greek the default case is Nominative so we don't need to try to inflect it
i = words.index('[Y.Nom]')
words[i] = ent_form
ent_case = "NOM"
elif "[Y.Gen]" in words:
i = words.index('[Y.Gen]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form,
f"N;GEN;{ent_number}",
language='ell2')[0]
ent_case = "GEN"
elif "[Y.Acc]" in words:
i = words.index('[Y.Acc]')
if do_not_inflect:
words[i] = ent_form
else:
words[i] = inflect(ent_form,
f"N;ACC;{ent_number}",
language='ell2')[0]
ent_case = "ACC"
# Now also check the correponsing articles, if they exist
if "[DEF;Y]" in words:
i = words.index('[DEF;Y]')
words[i] = article[f"ART;DEF;{ent_gender};{ent_number};{ent_case}"]
if "[DEF.Gen;Y]" in words:
i = words.index('[DEF.Gen;Y]')
words[i] = article[f"ART;DEF;{ent_gender};{ent_number};GEN"]
if "[PREPDEF;Y]" in words:
i = words.index('[PREPDEF;Y]')
words[i] = article[f"ART;PREPDEF;{ent_gender};{ent_number};{ent_case}"]
if "[INDEF;Y]" in words:
i = words.index('[INDEF;Y]')
#print(f"ART;INDEF;{ent_gender};{ent_number};{ent_case}")
#print(article[f"ART;INDEF;{ent_gender};{ent_number};{ent_case}"])
words[i] = article[f"ART;INDEF;{ent_gender};{ent_number};{ent_case}"]
if "[DEF;Y.Fem]" in words:
i = words.index('[DEF;Y.Fem]')
words[i] = article[f"ART;DEF;FEM;{ent_number}"]
return words
help='output file path',
default=None)
# misc arguments
parser.add_argument('--force-download',
help='force download of models',
action='store_true')
args = parser.parse_args()
# set output file path
if args.output is None:
output_file_path = 'text_file.out'
else:
output_file_path = args.output
language = args.language
print(inflect("love", "V;3;SG", language=language))
print(inflect("love", "V;NFIN", language=language))
print(inflect("love", "V;V.PTCP;PRS", language=language))
print(inflect("drink", "V;3;SG", language=language))
print(inflect("drink", "V;NFIN", language=language))
print(inflect("drink", "V;V.PTCP;PRS", language=language))
print(inflect("drink", "N;3;SG", language=language))
print(inflect("αντίο", "V;3;SG", language=language))
language = 'ell'
print(inflect("Βέλγιο", "N;NOM;PL", language=language))
print(inflect("Βέλγιο", "N;NEUT;GEN;SG", language=language))
print(inflect("βέλγικη", "ADJ;FEM;GEN;SG", language=language))
print(inflect("ανταγωνιστικότητα", "N;ACC;PL", language=language))
print(
inflect(["βλέπω", "ακούω"], ["V;3;SG;IPFV;PRS", "V;PFV;PST;3;PL"],
parser.add_argument(
'-t',
'--testfile',
help='file to test on (in Unimorph format) | default: <empty>',
default='')
parser.add_argument('-o',
'--output',
help='output file path',
default=None)
# misc arguments
parser.add_argument('--force-download',
help='force download of models',
action='store_true')
args = parser.parse_args()
# set output file path
if args.output is None:
output_file_path = 'text_file.out'
else:
output_file_path = args.output
language = args.language
try:
inputs, outputs, tags = simple_read_data(args.testfile)
curr_out = inflect(inputs, tags, language=language)
correct = [o == c for o, c in zip(outputs, curr_out)]
print(f"Accuracy: {float(sum(correct))/len(correct)}")
except:
print("dangit")
