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