PV-Client/predict_client.py at main · BiB1Labu/PV-Client · 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
import os
import argparse
import random
import pandas as pd
import joblib
import torch
from sklearn.preprocessing import StandardScaler
import warnings
warnings.simplefilter('ignore')

from models.Client import Model

def getRes(args, data_input, scaler_,k, device):
    configs = argparse.ArgumentParser()
    configs.pred_len = k
    configs.output_attention = False
    configs.seq_len = args.seq_len
    configs.n_heads = 4
    configs.dropout = 0.1
    configs.d_ff = 128
    configs.activation = 'gelu'
    configs.factor = 3
    configs.e_layers = 2
    configs.w_lin = 1.0
    configs.w_trans = 1.0
    # configs.enc_in = 25
    configs.enc_in = 6
    model = Model(configs)
    model.load_state_dict(torch.load('model/model.pt'))
    model.to(device)
    model.eval()

    # mean_ = scaler_.mean_[-1]
    # scale_ = scaler_.scale_[-1]

    # data
    data = data_input.copy()

    features2 = ['if30', 'if90', 'if100', 'if70', 'if110','ir10']
    data_s = data[features2].values
    data_scaled = scaler_.fit_transform(data_s)

    df_input = pd.DataFrame(data = data_scaled, columns = ['if30', 'if90', 'if100', 'if70', 'if110','ir10'])
    df_input['time'] = data['time']

    df_input['ir10_k'] = df_input['ir10'].shift(k)

    df_input.dropna(inplace = True)

    df_input_s = df_input[['if30', 'if90', 'if100', 'if70', 'if110', 'ir10_k', 'ir10']]

    df_x = df_input_s.values
    mean_ = scaler_.mean_[-1]
    scale_ = scaler_.scale_[-1]
    y_predList = []

    with torch.no_grad():
        for i in range(0, len(df_x) - args.seq_len + 1-k, k):
            x_i = df_x[i: i + args.seq_len, 0:6]
            y_i = df_x[i + args.seq_len: i + args.seq_len+k, 6]

            x_i = torch.tensor(x_i)
            x_i = x_i.float().to(device)
            x_i = x_i.unsqueeze(0)
            # print(x_i.shape)
            # print(y_i.shape)
            # print(model)
            output_ = model(x_i, None, None, None)[0,:,-1] + model(x_i, None, None, None)[0,:,0]
            y_pred = output_*scale_+mean_
            y_predList.extend(list(y_pred.cpu().numpy()))
    return y_predList

def main():
    parser = argparse.ArgumentParser()
    parser.add_argument('--input_dim', type = int, default = 2)
    parser.add_argument('--hid_dim', type = int, default = 6)

    parser.add_argument('--seq_len', type = int, default = 192)

    parser.add_argument('--seed', type = int, default = 0)

    parser.add_argument('--dataPath', type = str, default = './input')

    parser.add_argument('--fileNameR', type = str, default = 'input_real.xlsx')
    parser.add_argument('--fileNameF', type = str, default = 'input_forecast.xlsx')
    parser.add_argument('--fileNameS', type = str, default = 'input_station.xlsx')

    parser.add_argument('--modelPath', type = str, default = './model')

    parser.add_argument('--outPath', type = str, default = './output')
    parser.add_argument('--outfile', type = str, default = 'output_result.xlsx')

    parser.add_argument('--scalerPath', type = str, default = './scaler')
    parser.add_argument('--scalerName', type = str, default = 'scaler.save')
    parser.add_argument('--cap', type=int, default=375)

    args = parser.parse_args()

    if not os.path.exists(args.outPath):
        os.makedirs(args.outPath)

    random.seed(args.seed)
    torch.manual_seed(args.seed)
    torch.cuda.manual_seed(args.seed)

    device = torch.device('cpu')
    if torch.cuda.is_available():
        device = torch.device('cuda')

    scaler_ = StandardScaler()

    # scaler_ = joblib.load(os.path.join(args.scalerPath, args.scalerName))
    #print(data_real.shape)

    #data loading

    filePathF = os.path.join(args.dataPath, args.fileNameF)
    df_f = pd.read_excel(filePathF)
    featuresF = ['time', 'if30']
    df_fs = df_f[featuresF]

    filePathR = os.path.join(args.dataPath, args.fileNameR)
    df_r = pd.read_excel(filePathR)
    featuresR = ['time', 'ir10']
    df_rs = df_r[featuresR]

    df_m = pd.merge(df_f, df_r)

    #station information

    filePathS = os.path.join(args.dataPath, args.fileNameS)
    df_s = pd.read_excel(filePathS)
    featuresS = ['time', 'is10', 'is20']
    df_ss = df_s[featuresS]

    df_ss['coef'] = df_ss['is20']/df_ss['is10']
    df_ss_  = df_ss[args.seq_len+96:]
    y_res = getRes(args, df_m, scaler_,96, device)
    # print(len(y_res))

    df_o = pd.DataFrame(data=y_res, columns=['power'])
    # print(len(df_o))

    df_f['time'] = pd.to_datetime(df_f['time'], format='%Y-%m-%d %H:%M:%S')
    df_f = df_f[args.seq_len+96:]
    df_f.set_index(['time'], inplace=True)
    df_o['time'] = df_f.index

    df_o['if30'] = df_f['if30'].values

    df_o.loc[df_o['if30'] <= 0, 'power'] = 0
    df_o.loc[df_o['power'] <= 0, 'power'] = 0
    df_o.loc[df_o['power'] >= args.cap, 'power'] = args.cap

    df_o['or10'] = df_o['power'] * df_ss_['coef'].to_numpy()

    df_o['or20'] = df_o['or10'] * 1.2
    df_o['or30'] = df_o['or10'] * 0.8

    features = ['time', 'or10', 'or20', 'or30']
    df_oo = df_o[features]
    df_oo.to_excel(os.path.join(args.outPath, args.outfile), float_format = '%.3f', index = False)


if __name__ == '__main__':
    main()