def token_emotion_mat(vocab: Vocab):
"""pass"""
emotion_mat = np.zeros(shape=(vocab.size()))
emotion_mat[vocab.get_group(vocab.postive_name)] = 1
emotion_mat[vocab.get_group(vocab.negtive_name)] = -1
return emotion_mat
def doc_onehot_mat(doc_tokens: List[List[Text]], vocab: Vocab):
"""pass"""
tk2idx = vocab.tk2idx
all_tks = list(tk2idx.keys())
onehot_mat = np.zeros(shape=(vocab.size() + 1, len(doc_tokens)),
dtype=np.int8)
for id, doc in enumerate(doc_tokens):
tks = list(map(lambda tk: tk2idx[tk] if tk in doc else -1, all_tks))
onehot_mat[tks, id] = 1
return onehot_mat
