CV Type file corrector

cv-files-corrector.py

###########################################################################################
####################################   Requirements   #####################################
# Python version 3.10 and newest(didn't tested on older versions)
# pip install python-magic
#################################### How does it work #####################################
# Traverses through files in specified directory, checking the metadata of each file:
#  - if file type is not pdf/docx/doc/txt - REMOVES it,
#  - if extension file different from the type in metadata(pdf/docx/doc/txt) - RENAMES it.
####################################    How to Run    #####################################
# 1. save the script to your system.
# 2. python cv-files-corrector.py /path/to/folder/with/cv
# 3. await to "Done!" message.
###########################################################################################

import magic
import os
import sys

args = sys.argv

if len(args) <= 1:
  print("Directory with resumes should be specified!")
  sys.exit(1)

dir_path = sys.argv[1]

def should_be_removed(file_mime_type, path):
  removed = False
  for removable_type in types_to_remove:
    if removable_type in file_mime_type:
     
      os.remove(path)
      removed = True
      return removed
  return removed

def get_file_type(path):
  mime = magic.Magic(mime=True)
  file_mime_type = mime.from_file(path)
  if "pdf" in file_mime_type:
    return "pdf"
  elif "msword" in file_mime_type:
    return "doc"
  elif "vnd.openxmlformats-officedocument.wordprocessingml.document" in file_mime_type:
    return "docx"
  elif "text/plain" in file_mime_type:
    return "txt"
  else:
    print(f"Removing unsupportable file {path} with type {file_mime_type}")
    return None

files = os.listdir(dir_path)

for file in files:
  filecreds = os.path.basename(file).split(".")
  if len(filecreds) == 1:
    print(f"Removing unsupported file {file}")
    os.remove(f"{dir_path}/{file}")
    continue
  filename = filecreds[0]
  filename_type = filecreds[1]
  full_path = f"{dir_path}/{file}"
  file_type = get_file_type(full_path)
  if file_type is not None and file_type != filename_type:
    new_file_path = f"{dir_path}/{filename}.{file_type}"
    os.rename(full_path, new_file_path)
    print(f"Renamed {full_path} to {new_file_path}")

print("Done!")