Fetches all unique file types in all subfolders

README.md

Search unique file types

This script returns all unique file type in a given directory, including subdirectories.

Optional parameter: ROOT_DIR (default: ./)

py search-unique-file-types.py ./your-dir

search-unique-file-types-optimized.py

import os
import time
import sys
 
IGNORED_DIRECTORIES = ('node_modules', '.git', '.vs', 'obj', 'bin')
FILE_TYPES = []
ROOT_DIR = './'

def listDirectories(path):
    if not (os.access(path, os.F_OK) and os.access(path, os.R_OK) and os.access(path, os.X_OK)):
        return
        
    for obj in os.scandir(path):
        if (obj.is_file()):
            # TODO: optimize filename build and stack push
            fileName = obj.name if obj.name.startswith('.') else '*.' +obj.name.split('.')[-1]
            if fileName not in FILE_TYPES:
                FILE_TYPES.append(fileName)
                
        if (obj.is_dir()) and (not obj.path.endswith(IGNORED_DIRECTORIES)):
            listDirectories(obj.path)

ROOT_DIR = sys.argv[1] if os.access(sys.argv[1], os.F_OK) else ROOT_DIR

listDirectories(ROOT_DIR)

for t in FILE_TYPES:
    print(t)

search-unique-file-types.py

import os
import time
import sys
 
IGNORED_DIRECTORIES = ('node_modules', '.git', '.vs', 'obj', 'bin')
FILE_TYPES = []
FILE_COUNTER = 0
DIRECOTRY_COUNTER = 0
ROOT_DIR = './'

def listDirectories(path):
    if not (os.access(path, os.F_OK) and os.access(path, os.R_OK) and os.access(path, os.X_OK)):
        print('WARN: Cant access {}'.format(path))
        return
        
    for obj in os.scandir(path):
        if (obj.is_file()):
            global FILE_COUNTER
            FILE_COUNTER += 1

            # TODO: optimize filename build and stack push
            fileName = obj.name if obj.name.startswith('.') else '*.' +obj.name.split('.')[-1]
            if fileName not in FILE_TYPES:
                FILE_TYPES.append(fileName)
              
        if (obj.is_dir()) and (not obj.path.endswith(IGNORED_DIRECTORIES)):
            global DIRECOTRY_COUNTER
            DIRECOTRY_COUNTER += 1
            listDirectories(obj.path)

ROOT_DIR = sys.argv[1] if os.access(sys.argv[1], os.F_OK) else ROOT_DIR
print('Searching for unique file types in {}'.format(ROOT_DIR))

start = time.time()
listDirectories(ROOT_DIR)
end = time.time()

print('\r\nFetched {} directories and {} files in {}s'.format(DIRECOTRY_COUNTER, FILE_COUNTER, end - start))
print('Found {} unique file types'.format(len(FILE_TYPES)))
input('\r\nPress Enter to show...')

for t in FILE_TYPES:
    print(t)