def create_batches(self, contexts):
batch_data = []
label_data = []
for target, t_context in contexts:
target_index = Vocabulary.getIndex(target)
context_index = Vocabulary.getIndex(t_context)
if target_index is not None and context_index is not None:
batch_data.append(target_index)
label_data.append(context_index)
return batch_data, label_data
def createDataFrameOutput(self):
vector_column = []
embedding_lookup_table = self.embeddings.eval()
for text in self.texts:
word_tokens = text.tokens
vector_tokens = np.zeros([self.embedding_size])
for word in word_tokens:
word_index = Vocabulary.getIndex(word)
if word_index is not None:
vector_tokens += embedding_lookup_table[word_index, :]
vector_column.append(vector_tokens)
pd_column = pd.DataFrame({'VECTOR': vector_column})
self.dataframe['VECTOR'] = pd_column['VECTOR']
return self.dataframe
