def predict_score(test_sequence):
max_sequence_length = 200
model = load_model('./my_model.h5')
with open('./word_integer_map.json') as f:
word_to_id = json.load(f)
temp = []
test_sequence = test_sequence.lower()
save = []
temp = []
ids = []
for word in test_sequence.split(" "):
try:
if (check(word) != -1):
continue
i = []
i.append(word_to_id[word])
ids.append(i)
save.append(word)
#print(word)
except:
x = 1
pads = []
for i in ids:
pads.append(sequence.pad_sequences([i], maxlen=max_sequence_length))
scores = []
for p in pads:
scores.append(model.predict(np.array([p][0]))[0][0])
s_dict = {}
i = -1
for s in scores:
i+=1
print(i)
if (check(save[i]) != -1):
continue
s_dict[s] = save[i]
ans = sorted(s_dict.items(),reverse=True)
return ans
def stop_response(self, word):
global user_score
if not (dict.check(word)) or len(word) < 4:
print("Nope. Real words with length >= 4 only.")
user_score -= 1
return
self.vocab_tree.add(word)
self.score -= 1
def stop_response(self, word):
global user_score, words_played
if not(dict.check(word)) or len(word) < 4 or word in words_played:
print("Nope. Real words with length >= 4 only.")
user_score -= 1
return False
self.vocab_tree.add(word)
self.score -= 1
return True
def get_show(self, word):
global done, user_score
print("Show")
word_ = input("~> ")
if not word in word_ or len(word_) < 4 :
print(f"What? We're talking about something that starts with {word.upper()}, of length greater or equal to 4.")
self.get_show(word)
return
elif dict.check(word_):
print("You're right; didn't think of that one :) ")
self.score -= 1
self.vocab_tree.add(word_)
else:
print("I knew you were bluffing ;) ")
user_score -= 1
done = True
def isWord(word):
for ch in word:
if not ('a' <= ch.lower() <= 'z'):
return False
return dict.check(word)
def isWord(word):
for ch in word:
if not ("a" <= ch.lower() <= "z"):
return False
return dict.check(word)
pads = []
for i in ids:
pads.append(sequence.pad_sequences([i], maxlen=max_sequence_length))
scores = []
for p in pads:
scores.append(model.predict(np.array([p][0]))[0][0])
temp_padded = sequence.pad_sequences([temp], maxlen=max_sequence_length)
pred_score = model.predict(np.array([temp_padded][0]))[0][0]
#print([temp][0][1])
temp_padded = sequence.pad_sequences([temp], maxlen=max_sequence_length)
pred_score = model.predict(np.array([temp_padded][0]))[0][0]
print("the score for the comment is %s" % (pred_score * 100))
s_dict = {}
i = -1
for s in scores:
i += 1
if (check(save[i]) != -1):
continue
s_dict[s] = save[i]
ans = sorted(s_dict.items(), reverse=True)
print(ans)
#print(scores)
# gives score percentage