preprocessing.py 文件源码

def one_hot_encoding_sequences(seqs, sequenceLength):
    """
    input: genome sequences
    output: one_hot encoded sequence array
    """

    le = preprocessing.LabelEncoder()

    one_hot_sequences = []

    le.fit_transform(['a','t','g','c'])
    for si, seq in enumerate(seqs):
        seqlen = len(seq)
        arr = np.chararray((seqlen,), buffer=seq)
        a = le.transform(arr)

        a = np.array(a)

        b = np.zeros((len(a), 4))
        b[np.arange(len(a)), a] = 1

        #b = np.array(b)
        b = b.transpose()

        if b.shape[1] == sequenceLength:
            one_hot_sequences.append(b)

    return one_hot_sequences

#print one_hot_encoding_sequences(['atgctgc','gctatgc'])

#print numpy.arange(8).reshape((4,8/4), order = 'F')