AutoHPC-autoencoder/prepare_data.py at main · wdong5/AutoHPC-autoencoder · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
import tensorflow as tf
import pathlib
import numpy as np
import os
import os.path
import math
import random
import pdb

from configuration import args
from feature_reduction import Shannon_reduction, PCA_reduction
from scipy.sparse import csr_matrix
from scipy.sparse import coo_matrix

np.random.seed(1337)  # for reproducibility
#from parse_tfrecord import get_parsed_dataset

# This is for CG data fetch
def get_CG_matrix(data_list, row_list, col_list):
    # data_list = open(os.path.join(args.dataset_dir, x_data_dir), 'r')
    # row_list = open(os.path.join(args.dataset_dir, x_row_dir), 'r')
    # col_list = open(os.path.join(args.dataset_dir, x_col_dir), 'r')
    sparse_x = []
    x_size = 0
    for x_data_str in data_list.readlines():
        x_data = x_data_str.split()
        x_data = np.array([float(i) for i in x_data])
        x_indptr = row_list.readline().split()
        x_indices = col_list.readline().split()
        X = csr_matrix((x_data, x_indices, x_indptr), shape=(args.matrix_row, args.matrix_col))
        sparse_x.append(X)
        x_size = x_size+1
        if x_size == args.sample_size:
            break
    return sparse_x

# This is for AMG data fetch
def get_AMG_files(A_data_dir, x_data_dir):
    A_data_files= os.listdir(A_data_dir)
    x_data_files= os.listdir(x_data_dir)
    A_data = []
    x_data = []
    #read matrix A as NN inputs
    for file in A_data_files:
        A_matrix_data = []
        if not os.path.isdir(file):
          data_list = open(os.path.join(A_data_dir,file),'r')
          row_col = data_list.readline().split()
          for A_data_str in data_list.readlines():
              # pdb.set_trace()
              A_ele = A_data_str.split()
              float(A_ele[2])
              A_matrix_data.append(A_ele)
          A_matrix_data = np.array(A_matrix_data)
          A_matrix = coo_matrix((A_matrix_data[:,2], (A_matrix_data[:,0], A_matrix_data[:,1])),shape=(int(row_col[1])+1, int(row_col[3])+1), dtype=np.float32)
        A_data.append(A_matrix)
        # pdb.set_trace()
    #read array x as NN single_output
    for file in x_data_files:
        x_array_data = []
        if not os.path.isdir(file):
          data_list = open(os.path.join(x_data_dir,file),'r')
          row_col = data_list.readline().split()
          for x_data_str in data_list.readlines():
              x_ele = x_data_str.split()
              # pdb.set_trace()
              x_array_data.append(float(x_ele[1]))
        x_data.append(x_array_data)
    x_data = np.array(x_data)
    return  A_data, x_data

# This is for MG data fetch
def get_MG_matrix(r_int_list, r_sol_list):
    r_array = []
    r_sol_array = []
    for r_int_str in r_int_list.readlines():
        r_int = r_int_str.split()
        r_data = np.array([float(i) for i in r_int])
        r_array.append(r_data)

    for r_sol_str in r_sol_list.readlines():
        r_sol = r_sol_str.split()
        r_sol_data = np.array([float(i) for i in r_sol])
        r_sol_array.append(r_sol_data)

    return r_array, r_sol_array

# This is for Lagos data fetch
def get_Lagos_files(A_data_dir):
    A_data_files= os.listdir(A_data_dir)
    A_data = []
    for file in A_data_files:
        A_matrix_data = []
        if not os.path.isdir(file):
          data_list = open(os.path.join(A_data_dir,file),'r')
          for i, A_data_str in enumerate(data_list):
              if (i>6):
                  A_ele = float(A_data_str)
                  A_matrix_data.append(A_ele)
          A_matrix_data = np.array(A_matrix_data)
        A_data.append(A_matrix_data)
    return A_data

def get_the_length_of_dataset(dataset):
    count = 0
    for i in dataset:
        count += 1
    return count

def load_dataset(list_file):
    raw_im_list = np.loadtxt(list_file)
    return  raw_im_list

def normalize(matrix):
    norm =np.linalg.norm(matrix)
    if norm == 0:
        return matrix
    boolArr = (matrix == 0)
    min = matrix.min()
    if min == 0:
        matrix_nonzero = matrix[matrix>0]
        min = matrix_nonzero.min()
    matrix_x = (matrix - min) / norm
    matrix_x[boolArr] = 0.0
    return matrix_x

def generate_datasets():
    if (args.benchmark=='CG'):
        #CG dataset
        path = args.data_dir + "NPB3.3-SER-C/CG/CG_dataset"
        cg_csr = open(os.path.join(path, "cg_csr_value.txt"), 'r')
        cg_csr_row = open(os.path.join(path, "cg_csr_row_ind.txt"), 'r')
        cg_csr_col = open(os.path.join(path, "cg_csr_col_ind.txt"), 'r')
        cg_x = open(os.path.join(path, "cg_x.txt"), 'r')
        train_value = get_CG_matrix(cg_csr, cg_csr_row, cg_csr_col)
        train_x= load_dataset(cg_x)

    elif (args.benchmark=='AMG'):
        path = args.data_dir + "AMG/test"
        # A_data_dir = os.path.join(path, "IJ_A")
        # x_data_dir = os.path.join(path, "IJ_x")
        # train_value, train_x = get_AMG_files(A_data_dir, x_data_dir)
        r_int_list = open(os.path.join(path, "IJ_A_diagonal.txt"), 'r')
        r_sol_list = open(os.path.join(path, "IJ_x.txt"), 'r')
        train_value, train_x = get_MG_matrix(r_int_list, r_sol_list)

    elif (args.benchmark=='MG'):
        path = args.data_dir + "NPB3.3-SER-C/MG/MG_dataset"
        r_int_list = open(os.path.join(path, "mg_init_r.txt"), 'r')
        r_sol_list = open(os.path.join(path, "mg_sol_r.txt"), 'r')
        train_value, train_x = get_MG_matrix(r_int_list, r_sol_list)

    elif (args.benchmark=='Lagos_fine'):
        path = args.data_dir + "Laghos/fine_grained"
        input_e = np.array(get_Lagos_files(os.path.join(path, "input_e")))
        input_v = np.array(get_Lagos_files(os.path.join(path, "input_v")))
        input_x = np.array(get_Lagos_files(os.path.join(path, "input_x")))
        output_e = np.array(get_Lagos_files(os.path.join(path, "output_e")))
        output_v = np.array(get_Lagos_files(os.path.join(path, "output_v")))
        output_x = np.array(get_Lagos_files(os.path.join(path, "output_x")))
        train_value = np.concatenate((input_e, input_v, input_x), axis=1)
        train_x = np.concatenate((output_e, output_v, output_x), axis=1)

    elif (args.benchmark=='Lagos_coarse'):
        path = args.data_dir + "Laghos/coarse_grained"
        input_e = np.array(get_Lagos_files(os.path.join(path, "input_e")))
        input_v = np.array(get_Lagos_files(os.path.join(path, "input_v")))
        input_x = np.array(get_Lagos_files(os.path.join(path, "input_x")))
        output_e = np.array(get_Lagos_files(os.path.join(path, "output_e")))
        output_v = np.array(get_Lagos_files(os.path.join(path, "output_v")))
        output_x = np.array(get_Lagos_files(os.path.join(path, "output_x")))
        train_value = np.concatenate((input_e, input_v, input_x), axis=1)
        train_x = np.concatenate((output_e, output_v, output_x), axis=1)

    return train_value,  train_x