seq_load.py

from Bio import SeqIO
from Bio.Seq import Seq
import numpy as np
import pandas as pd
import sys


import torch
from torch.autograd import Variable
from torch import optim
from sklearn.model_selection import train_test_split
import itertools



def convert_seq_to_code(seq):
    seq=seq.upper()
    feat_bicoding=[]
    bicoding_dict={'A':[1,0,0,0,1,1,1,0.1260],'C':[0,1,0,0,0,1,0,0.1340],'G':[0,0,1,0,1,0,0,0.0806],'T':[0,0,0,1,0,0,1,0.1335],'N':[0,0,0,0,0,0,0,0]}
    for each_nt in seq:
        feat_bicoding+=bicoding_dict[each_nt]
    return feat_bicoding



def load_data_code(in_fa):
    data=[]
    for record in SeqIO.parse(in_fa, "fasta"):
        seq=str(record.seq)
        bicoding=convert_seq_to_code(seq)
        data.append(bicoding)

    return data



def header_and_seqload(in_fa):
    data=[]
    fa_header=[]
    for record in SeqIO.parse(in_fa, "fasta"):
        seq=str(record.seq).upper()
        bicoding=convert_seq_to_code(seq)
        data.append(bicoding)
        fa_header.append(str(record.description))

    return data, fa_header



def load_train_val_code(pos_train_fa,neg_train_fa):
    data_pos_train = []
    data_neg_train = []

    data_pos_train = load_data_code(pos_train_fa)
    data_neg_train = load_data_code(neg_train_fa)


    data_train = np.array([_ + [1] for _ in data_pos_train] + [_ + [0] for _ in data_neg_train])
    np.random.seed(33)
    np.random.shuffle(data_train)

    X = np.array([_[:-1] for _ in data_train])
    y = np.array([_[-1] for _ in data_train])

    # X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=1.0/8, random_state=42)

    # return X_train,y_train,X_test,y_test
    return X, y


def to_torch_fmt(X_train, y_train, X_test, y_test):
    X_train = X_train.reshape(X_train.shape[0], 51, 8)
    X_test = X_test.reshape(X_test.shape[0], 51, 8)


    X_train = torch.from_numpy(X_train).float()
    y_train = torch.from_numpy(y_train).long()
    #y_train = torch.from_numpy(y_train).float()
    X_test = torch.from_numpy(X_test).float()

    return X_train, y_train, X_test, y_test