-
Notifications
You must be signed in to change notification settings - Fork 0
/
preprocess.py
executable file
·43 lines (31 loc) · 1.32 KB
/
preprocess.py
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
#!/bin/python
from argparse import ArgumentParser
from multiprocessing import Pool
import pandas as pd
import numpy as np
import psutil
def process(data: pd.DataFrame) -> pd.DataFrame:
"""
Process an equally-split chunk of the entire dataset to ensure maximum utility of multicore-CPUs.
The results of the functions running parallel will in the end be concatenated.
"""
return data
if __name__ == "__main__":
parser = ArgumentParser()
parser.add_argument('infile')
parser.add_argument('outfile')
args = parser.parse_args()
all_items = pd.read_csv(args.infile)
original_size = len(all_items)
print(f"Original dataset-length: {original_size}")
# Get amount of physical CPU-cores for maximum utility usage.
num_cpus = psutil.cpu_count(logical=False)
# Divide DataFrame into equal chunks to be processed in parallel.
sub_dfs = np.array_split(all_items, num_cpus)
del all_items
process_pool = Pool(processes=num_cpus)
resulting_dfs = process_pool.map(process, sub_dfs)
dataset = pd.concat(resulting_dfs, ignore_index=True)
preprocessed_size = len(dataset)
print(f'The preprocessed dataset consists of {preprocessed_size} samples. This equates to {100*preprocessed_size/original_size}% of the original dataset.', flush=True)
dataset.to_csv(args.outfile)