MANUAL file expiry deletion script

.gitignore

@@ -0,0 +1,2 @@
+/venv
+/infra_file_auto_expiry/source/__pycache__

file_expiry.sh

@@ -0,0 +1,14 @@
+#!/bin/bash
+
+set -e
+
+PYTHON=$(which python) # get python location
+MAIN_SCRIPT=$(realpath "infra_file_auto_expiry/source/main.py")
+STORAGE_FOLDER=""
+
+echo "Automatic File Expiry Tool"
+echo "Removes expired files - unused for 30 days or more"
+echo "Python location is: $PYTHON"
+echo "Running $MAIN_SCRIPT"
+
+(crontab -l ; echo "0 0 * * * $PYTHON $MAIN_SCRIPT \"$STORAGE_FOLDER\"") | crontab -

infra_file_auto_expiry/source/main.py

@@ -0,0 +1,17 @@
+from utils import *
+import sys
+
+
+def main(base_folder):
+    delete_expired_files(base_folder, temp_folder, 0)
+
+if __name__ == "__main__":
+    if len(sys.argv) != 3:
+
+        print("Usage: python file_expiry.py <folder to scan> <temp>")
+        sys.exit(1)
+
+    base_folder = sys.argv[1]
+    temp_folder = sys.argv[2]
+
+    main(base_folder=base_folder)

infra_file_auto_expiry/source/utils.py

@@ -0,0 +1,96 @@
+import pwd
+import shutil
+import os
+import time
+
+def is_expired_file(path, days_for_expire=30):
+    """
+    Checks the last time a file or folder has been accessed. If it has not 
+    been accessed in the days specified, then return True. False if otherwise.
+
+    string days: The full path to the file that is being checked
+    int days: The amount of days since last access that indicates that a file
+        has expired. 
+    """
+    current_time = time.time()
+    # get last access time (includes change and modification times)
+    file_stat = os.stat(path)
+
+    # compare days for expiry (in seconds) to the time since last access
+    return ((days_for_expire * 24 * 60 * 60) < (current_time - file_stat.st_atime) and
+        (days_for_expire * 24 * 60 * 60) < (current_time - file_stat.st_ctime) and 
+        (days_for_expire * 24 * 60 * 60) < (current_time - file_stat.st_mtime))
+
+def is_expired_folder(folder_path, days_for_expire=30):
+    """
+    Goes through all files in a folder, and deletes the ones that
+    have expired
+
+    string base_folder: The folder containing the files to delete
+    """
+    if not os.path.isdir(folder_path):
+        print("Bad folder path:", folder_path)
+        return False
+
+    all_files_expired = True
+    for entry in os.scandir(folder_path):
+        if entry.is_dir():
+            if not is_expired_folder(entry.path, days_for_expire):
+                return False
+        elif entry.is_file() and not is_expired(entry.path, days_for_expire):
+            all_files_expired = False
+    return all_files_expired
+
+def is_expired(path, days_for_expire=30):
+    if os.path.isfile(path):
+        return is_expired_file(path, days_for_expire)
+    if os.path.isdir(path):
+        return is_expired_folder(path, days_for_expire)
+
+    return False
+
+def get_file_creator(path):
+    """
+    Gets the file creators username
+
+    ls -l filepath command on linux returns something like this:
+    -rw-rw-r-- 1 machung machung 4 Feb 17 05:14 /home/machung/test.txt
+    So we select the file owner username from this command output. 
+
+    string file_path: The absolute path of the file
+    """
+    try:
+        # Get the UID of the file or directory owner
+        uid = os.stat(path).st_uid
+        # Get the username associated with the UID
+        username = pwd.getpwuid(uid).pw_name
+        return username
+    except FileNotFoundError:
+        print(f"Error: File '{path}' not found.")
+        return None
+    except Exception as e:
+        print(f"Error: {e}")
+        return None
+
+def notify_file_creator(file_creator):
+    """
+    TODO: implement proper notification system
+    """
+    print(f"Deleting file by ", file_creator)
+
+def scan_folder_for_expired(folder_path, days_for_expire=30):
+    for entry in os.scandir(folder_path):
+        if is_expired(entry.path, days_for_expire):
+            yield entry.path 
+
+def delete_expired_files(folder_path, temp_folder, days_for_expire=30):
+    if not os.path.isdir(folder_path) or not os.path.isdir(temp_folder):
+        print("Base folder does not exist ")
+        return
+
+    for path in scan_folder_for_expired(folder_path, days_for_expire):
+        notify_file_creator(get_file_creator(path))
+        shutil.move(path, temp_folder)
+
+    shutil.rmtree(temp_folder)
+

infra_file_auto_expiry/tests/test_utils.py

@@ -0,0 +1,49 @@
+import unittest
+import os
+import sys
+from unittest.mock import patch
+
+module_path = os.path.dirname(
+    os.path.dirname(os.path.abspath(__file__))
+)
+sys.path.append(module_path)
+
+from source.utils import *
+
+
+class TestUtils(unittest.TestCase):
+    @patch("pwd.getpwuid")
+    @patch("os.stat")
+    def test_get_file_creator(self, patch_os, patch_pwd):
+        """
+        Tests retrieving the user name of a file owner
+        """
+        # Successfully retrieves file owner
+        patch_os.return_value.st_uid=0
+        patch_pwd.return_value.pw_name="tester_account"
+        self.assertEqual(get_file_creator("/home/machung/test.txt"), "tester_account")
+
+
+    @patch('os.stat')
+    @patch('time.time')
+    def test_is_expired(self, patch_time, patch_os):
+        """
+        Tests the is_expired_file function
+        """
+        time_to_expire_days = 30
+        patch_os.return_value.st_atime = 5 * 24 * 60 * 60
+        patch_time.return_value = 50 * 24 * 60 * 60 
+
+        # Days since last access is 50 - 5 = 45 > 30
+        # The file should be expired
+        self.assertTrue(is_expired("test_name", time_to_expire_days))
+
+        patch_time.return_value = 10 * 24 * 60 * 60 
+        # Days since last access is 10 - 5 = 5 < 30
+        # The file should not be expired
+        self.assertFalse(is_expired("test_name", time_to_expire_days))
+
+    def test_notify_file_creator(self):
+        pass
+if __name__ == '__main__':
+    unittest.main()