def __init__(self, seqs):
self._seqs = list(seqs)
self._seq_lens = np.array([len(seq) for seq in self._seqs])
assert self._seq_lens.max() == self._seq_lens.min()
self.seq_len = self._seq_lens[0]
self.one_hot_coded_seqs = one_hot_encode_sequences(self._seqs)
self.freq_one_hot_coded_seqs = self._init_freq_one_hot_coded_seqs()
def __init__(self, seqs, include_shape=True):
self._seqs = list(seqs)
self._seq_lens = np.array([len(seq) for seq in self._seqs])
assert self._seq_lens.max() == self._seq_lens.min()
self.seq_len = self._seq_lens[0]
self.fwd_one_hot_coded_seqs = one_hot_encode_sequences(self._seqs)
self.rc_one_hot_coded_seqs = self.fwd_one_hot_coded_seqs[:,::-1,::-1]
if include_shape:
(self.fwd_shape_features, self.rc_shape_features
) = code_seqs_shape_features(
self._seqs, self.seq_len, len(self._seqs))
self.fwd_coded_seqs = np.dstack(
(self.fwd_one_hot_coded_seqs, self.fwd_shape_features))
self.rc_coded_seqs = np.dstack(
(self.rc_one_hot_coded_seqs, self.rc_shape_features))
else:
self.fwd_shape_features, self.rc_shape_features = None, None
self.fwd_coded_seqs = self.fwd_one_hot_coded_seqs
self.rc_coded_seqs = self.rc_one_hot_coded_seqs
