I want to randomly sample a file in Python

things to do

1. Create 100 files
2. Get the file path with the specified extension
3. Randomly select 10 and copy to another directory

Directory structure

.
├── input
│   └── _sampled
└── src
    └── random_sampling.py

code

• All are standard libraries, so no installation is required.

random_smpling.py

from pprint import pprint
import os
from pathlib import Path
import random
import shutil


class FileControler(object):
    def make_file(self, output_dir, file_name, content='') -> None:
        """File creation"""
        #Create if there is no output directory
        os.makedirs(output_dir, exist_ok=True)
        #Combine output directory and file name
        file_path = os.path.join(output_dir, file_name)
        #File creation
        with open(file_path, mode='w') as f:
            #Create an empty file by default
            f.write(content)

    def get_files_path(self, input_dir, pattern):
        """Get file path"""
        #Create a path object by specifying a directory
        path_obj = Path(input_dir)
        #Match files in glob format
        files_path = path_obj.glob(pattern)
        #Posix conversion to treat as a character string
        files_path_posix = [file_path.as_posix() for file_path in files_path]
        return files_path_posix

    def random_sampling(self, files_path, sampl_num, output_dir, fix_seed=True) -> None:
        """Random sampling"""
        #If you sample the same file every time, fix the seed value
        if fix_seed is True:
            random.seed(0)
        #Specify the file group path and the number of samples
        files_path_sampled = random.sample(files_path, sampl_num)
        #Create if there is no output directory
        os.makedirs(output_dir, exist_ok=True)
        #copy
        for file_path in files_path_sampled:
            shutil.copy(file_path, output_dir)

Run

Instance creation

file_controler = FileControler()

directory

Create 100 files in ʻinput / Copy the sampled file into ʻinput / _sampled

all_files_dir = '../input'
sampled_dir = '../input/_sampled'

First, make 100 files

50 files each of .py and .txt

for i in range(1, 51):
    file_controler.make_file(all_files_dir, f'file{i}.py')
    file_controler.make_file(all_files_dir, f'file{i}.txt')

Get only the path of the .py file

from pprint import pprint


pattern = '*.py'
files_path = file_controler.get_files_path(all_files_dir, pattern)

pprint(files_path)
# ['../input/file8.py',
#  '../input/file28.py',
#  '../input/file38.py',
#  .
#  .
#  .
#  '../input/file25.py',
#  '../input/file35.py',
#  '../input/file50.py']
#

print(len(files_path))
# 50

Random sampling for 10 files

sample_num = 10
file_controler.random_sampling(files_path, sample_num, sampled_dir)

Verification

Terminal

ls input/_sampled

file10.py file23.py file3.py  file35.py file36.py file37.py file38.py file4.py  file41.py file43.py