Informer2020/sim_graph_test_sc.sh at main · Isaacwilliam4/Informer2020 · 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
#!/bin/bash --login

#SBATCH --time=48:00:00   # walltime
#SBATCH --ntasks=4   # number of processor cores (i.e. tasks)
#SBATCH --nodes=1   # number of nodes
#SBATCH --gpus=4
#SBATCH --mem=128G   # memory per CPU core
#SBATCH -J "sim_graph"   # job name
#SBATCH --mail-user=isaacwilliam4@gmail.com   # email address
#SBATCH --mail-type=BEGIN
#SBATCH --mail-type=END
#SBATCH --mail-type=FAIL
#SBATCH -C pascal

# Execute Python script with command-line arguments

mamba activate informer

if [ "$1" == "simulation" ]; then

    if [ "$#" -lt 6 ]; then
        echo -e "Args required for graph simulation: \n
        number of nodes: the number of nodes in the original graph G \n
        timesteps: how many timesteps of simulated data you want to create \n
        line graph embedding: true or false for line graph embedding on the data \n
        alpha value: value (0,1], that represents the percent of edges that will be assigned values \n
	    permute: Whether or not to permute the data for line graph"

        # sequence len: The length of the sequence of timesteps the model will train on \n
        # label len: The length of the context that the model is given in order to make prediction \n
        # pred len: The number of timesteps the model will predict into the future \n
        # batch size: The batch size for data processing
        exit 1
    fi

    echo "Args: Type=$1, NumNodes=$2 TimeSteps=$3, LineGraphPartitioning=$4, Alpha=$5, Permute=$6" #SeqLen=$6, LabelLen=$7, PredLen=$8, BatchSize=$9


    line_graph=$4
    permute=$6

    if [ "$line_graph" == "true" ]; then
        echo 'Running training with line graph partitions'
        # Calculate enc_in and c_out
        enc_in=$(( $2 * $2 * 4))
        c_out=$(( $2 * $2 ))
        name=""
        if [ "$permute" == "true" ]; then
            name="lg_p_n$2_t$3"
            if [ -f ./data/$name.csv ]; then
                echo "Simulated, permuted, line graph file exists, skipping file generation..."
            else
                echo "Simulated, permuted, line graph file doesn't exist, generating file..."
                python ./line_graph.py --type "simulation" --line_graph --num_nodes $2 --timesteps $3 --alpha $5 --permute  --random_seed 42 --name $name
            fi


            # Load modules, insert code, and run your programs here
            python -u ./main_informer.py --model informer --target 'none' --data 'sim_graph' --m_true_len $c_out --data_path $name.csv --root_path "./data/" --features M --freq d --enc_in $enc_in --dec_in $enc_in --c_out $c_out --num_workers 0 --des lg_p_n$2_t$3_test --use_multi_gpu

        else
            name="lg_n$2_t$3"
            if [ -f ./data/$name.csv ]; then
                echo "Simulated line graph file exists, skipping file generation..."
            else
                echo "Simulated line graph file doesn't exist, generating file..."
                python ./line_graph.py --type "simulation" --line_graph --num_nodes $2 --timesteps $3 --alpha $5 --random_seed 42
            fi

            python -u ./main_informer.py --model informer --target 'none' --data 'sim_graph' --m_true_len $c_out --data_path $name.csv --root_path "./data/" --features M --freq d --enc_in $enc_in --dec_in $enc_in --c_out $c_out --num_workers 0 --des lg_p_n$2_t$3_test --use_multi_gpu

        fi


    elif [ "$line_graph" == "false" ]; then
        if [ "$permute" == "true" ]; then
            echo "Cannot permute unless using line graph data, exiting..."
            exit 1
        fi
        echo 'Running training with original data'

        if [ -f ./data/g_n$2_t$3.csv ]; then
            echo "Simulated graph file exists, skipping file generation..."
        else
            echo "Simulated graph file doesn't exist, generating file..."
            python ./line_graph.py --type "simulation" --num_nodes $2 --timesteps $3 --alpha $5 --random_seed 42
        fi
        # Calculate enc_in and c_out
        enc_in=$(( $2 * $2 ))
        c_out=$(( $2 * $2 ))

        # Load modules, insert code, and run your programs here
        python -u ./main_informer.py --model informer --target 'none' --data 'sim_graph' --m_true_len $c_out --data_path g_n$2_t$3.csv --root_path "./data/" --features M --freq d --enc_in $enc_in --dec_in $enc_in --c_out $c_out --num_workers 0 --des g_n$2_t$3_test --use_multi_gpu

    else

        echo "third argument should be 'true' or 'false'"
        exit 1

    fi

elif [ "$1" == "custom" ]; then
    echo "Preparing custom data... "

    if [ "$#" -lt 4 ]; then
        echo -e "Args required for custom data: \n
        data path: the path to the custom data should be of shape numnodes x timesteps \n
        line graph partitioning: true or false for including the line graph partition of the data \n
        number of nodes: number of nodes in the graph \n
        permute: Whether or not to permute the data for line graph"

    fi

    line_graph=$3
    num_edges=$(( $4 * $4 ))
    filename="${2##*/}"
    name="${filename%.*}"
    permute=$5
    echo "Args: Type=$1, Datapath=$2, LineGraphPartitioning=$3, NumNodes=$4, Permute=$5"

    if [ "$line_graph" == "true" ]; then

        ext="_lg_prepared"


        echo "Running custom data with line graph partitioning"

        python ./line_graph.py --type 'custom' --line_graph --data_path "$2" --name "$name$ext"

        tot_in=$(( 4 * $num_edges ))

        if [ -f ./data/$name$ext.csv ]; then

            python -u ./main_informer.py --pred_len 6 --batch_size 1 --seq_len 6 --label_len 3 --model informer --target 'none' --data 'custom' --m_true_len $num_edges --data_path $name$ext.csv --root_path "./data/" --features M --freq d --enc_in $tot_in --dec_in $tot_in --c_out $num_edges --num_workers 0 --des $name --use_multi_gpu

        else

            echo "File generation failed, exiting..."
            exit 1

        fi

    else

        ext="_prepared"


        echo "Running custom data without line graph partitioning"


        python ./line_graph.py --type 'custom' --data_path "$2" --name "$name$ext"

        if [ -f ./data/$name$ext.csv ]; then

            python -u ./main_informer.py --pred_len 6 --batch_size 1 --seq_len 6 --label_len 3 --model informer --target 'none' --data 'custom' --m_true_len $num_edges --data_path $name$ext.csv --root_path "./data/" --features M --freq d --enc_in $num_edges --dec_in $num_edges --c_out $num_edges --num_workers 0 --des $name --use_multi_gpu

        else

            echo "File generation failed, exiting..."
            exit 1

        fi

    fi

else

    echo "first argument must be either custom or simulation, run without further arguments for more information"

fi