Move data cleaning scripts to clean module
This commit is contained in:
parent
87be49f30f
commit
3b78a84782
7 changed files with 5 additions and 5 deletions
41
code/clean/files.py
Normal file
41
code/clean/files.py
Normal file
|
|
@ -0,0 +1,41 @@
|
|||
import csv
|
||||
from pathlib import Path
|
||||
|
||||
|
||||
def filesize_csv(input_dir: Path, output_dir: Path) -> None:
|
||||
output_file = output_dir / "files.csv"
|
||||
with output_file.open("w") as fw:
|
||||
writer = csv.writer(fw)
|
||||
writer.writerow(["date", "filename", "mtime", "filesize"])
|
||||
|
||||
for j in input_dir.glob("*.json"):
|
||||
p_date = j.stem
|
||||
p_fname = j.name
|
||||
stat = j.stat()
|
||||
p_mtime = stat.st_mtime
|
||||
p_size = stat.st_size
|
||||
writer.writerow([p_date, p_fname, p_mtime, p_size])
|
||||
|
||||
|
||||
def ensure_dirs(input_dir: Path, output_dir: Path):
|
||||
if not input_dir.is_dir():
|
||||
raise ValueError
|
||||
output_dir.mkdir(exist_ok=True, parents=True)
|
||||
|
||||
|
||||
def main(input: str, output: str) -> None:
|
||||
input_dir = Path(input)
|
||||
output_dir = Path(output)
|
||||
ensure_dirs(input_dir, output_dir)
|
||||
filesize_csv(input_dir, output_dir)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
import sys
|
||||
|
||||
if not len(sys.argv) == 3:
|
||||
print("Please provide exactly one input directory and one output directory.")
|
||||
sys.exit(1)
|
||||
inp = sys.argv[1]
|
||||
out = sys.argv[2]
|
||||
main(inp, out)
|
||||
Loading…
Add table
Add a link
Reference in a new issue