feat: scan all cmlab member's directory

clean.py

@@ -1,111 +0,0 @@
-'''
-
-Clean the direction in CML Server (tmp2 or home directory)
-
-
-Suppose that the CLEAN_PATH may seems like this.
-And we will only check the directories & files under the user directory with "depth=1"
-
-CLEAN_PATH
-    |
-    +---- user1 (user directory)
-    |       |
-    |       +--- file1 (depth=1, will be check)
-    |       |
-    |       +--- Project1 (depth=1, will be check)
-    |       |       |
-    |       |       +--- subset1-1 (depth=2, won't be check)
-    |       |
-    |       +--- Project2 (depth=1)
-    |               |
-    |               +--- subset2-1 (depth=2)
-    |               |
-    |               +--- subset2-2 (depth=2)
-    |
-    +---- user2 (user directory)
-            |
-            +--- Project1 (depth=1)
-                    |
-                    +--- subset1-1 (depth=2)
-
-'''
-import os
-import subprocess
-import smtplib
-from email.mime.text import MIMEText
-
-CLEAN_PATH = "/home/master/12"
-KEEP_FILES = [".bashrc", ".bash_logout", ".bash_profile", "public-html", ".ssh", ".profile", "examples.desktop", ".bash_history", ".config", ".zshrc", ".vimrc", ".gnupg", ".oh-my-zsh", ".vim", ".viminfo", ".vscode-server", ".gitconfig", ".local"]
-WHITE_LIST = ["mysql", "unix_manager", "unixmanager215", "gangxing", "how083012", "yulunliu", "amie12349", "minliwu", "chenyuyang"]
-
-DATE = '2024-02-01'
-
-GOOGLE_OWNER = "snsd0805@cmlab.csie.ntu.edu.tw"
-GOOGLE_CODE = os.environ['GOOGLE_CODE']
-
-
-class Mailer():
-    def __init__(self, google_owner, google_code):
-        self.google_owner = google_owner
-        self.google_code = google_code
-
-    def send(self, sender, receiver, subject, msg):
-        msg = MIMEText(msg, 'html')
-        msg['Subject'] = subject
-        msg['From'] = sender
-        msg['To'] = receiver
-
-        smtp = smtplib.SMTP('smtp.gmail.com', 587)
-        smtp.ehlo()
-        smtp.starttls()
-        smtp.login(self.google_owner, self.google_code)
-
-        status = smtp.send_message(msg)
-        smtp.quit()
-
-        if status == {}:
-            return True
-        else:
-            return False
-
-class Cleaner():
-    def __init__(self, root, date, mailer, white_list=WHITE_LIST, keep_files=KEEP_FILES):
-        self.root = root
-        self.date = date
-        self.white_list = white_list
-        self.keep_files = keep_files
-        self.mailer = mailer
-
-    def run(self, test=True, send_mail=False):
-        for user in sorted(os.listdir(self.root)):
-            print(user)
-            if user in self.white_list:
-                continue
-
-            for obj in sorted(os.listdir(f'{self.root}/{user}')):
-                if obj in self.keep_files:
-                    continue
-
-                obj = obj.replace(' ', '\ ')
-
-                # stat {file or dir}:
-                #       get Create, modified information
-                # grep Modify:
-                #       get Modified time
-                # awk:
-                #       -F ' ': use space to split the data
-                #       print $2: column 2 (get date)
-                modify_date = subprocess.getoutput(f"stat {self.root}/{user}/{obj} | grep Modify | awk -F ' ' '{{print $2}}'" )
-
-                needRemove = modify_date < DATE
-                print(f'    {obj}')
-                print(f'        {modify_date} {needRemove}')
-
-
-
-
-
-mailer = Mailer(GOOGLE_OWNER, GOOGLE_CODE)
-mailer.send('unix_manager@cmlab.csie.ntu.edu.tw', 'snsd0805@cmlab.csie.ntu.edu.tw', '[測試郵件] test', 'Test mail')
-cleaner = Cleaner('/home/master/12', "2024/02/01", mailer)
-cleaner.run()

cleaner.py

@@ -0,0 +1,82 @@
+import os
+import subprocess
+
+MAIL_MESSAGE = '''
+Hi, {}<br>
+<br>
+因 CML Server 掛載家目錄(~)的 NAS 空間即將不足，我們將於 4/28(日) 00:00 開始清除長時間沒有修改紀錄的資料夾<br>
+<br>
+本次清理，我們將移除 2024/02/01 以後沒有修改紀錄的資料夾、檔案<br>
+以下是您家目錄中可能被移除的內容：<br>
+<ul>
+{}
+</ul>
+<br>
+若是仍要保留的資料請隨意對其做修改，以更新修改紀錄的時間。<br>
+造成不便，敬請見諒！<br><br>
+
+<br>
+* 信件由網管自動化工具寄出，若為錯誤寄送請忽略此信件。<br>
+<br>
+Best,<br>
+CMLab Unix Manager, Ting-Jun Wang<br>
+CMLab, National Taiwan University<br>
+Email: unix_manager@cmlab.csie.ntu.edu.tw<br>
+<pre>
+            __  __     _
+           / / / /__  (_)_ __
+   __  ___/ /_/ / _ \/ /\ \ /
+  /  |/  /\____/_//_/_//_\_\____ ____
+ / /|_/ / _ `/ _ \/ _ `/ _ `/ -_) __/
+/_/  /_/\_,_/_//_/\_,_/\_, /\__/_/
+                      /___/
+</pre>
+'''
+
+class Cleaner():
+    def __init__(self, user_dir, date, mailer, white_list, keep_files):
+        self.user_dir = user_dir
+        self.date = date
+        self.white_list = white_list
+        self.keep_files = keep_files
+        self.mailer = mailer
+
+    def run(self, test=True, send_mail=False):
+        username = self.user_dir.split('/')[-1]
+        if username in self.white_list:
+            print(f'    WHITE LIST: {username}')
+            return
+
+        remove_objs = []
+        for obj in sorted(os.listdir(self.user_dir)):
+            if obj in self.keep_files:
+                continue
+
+            obj = obj.replace(' ', '\ ')
+
+            # stat {file or dir}:
+            #       get Create, modified information
+            # grep Modify:
+            #       get Modified time
+            # awk:
+            #       -F ' ': use space to split the data
+            #       print $2: column 2 (get date)
+            modify_date = subprocess.getoutput(f"stat {self.user_dir}/{obj} | grep Modify | awk -F ' ' '{{print $2}}'" )
+
+            if modify_date < self.date:
+                remove_objs.append((obj, modify_date))
+                print(f'    {modify_date} {obj}')
+        
+        if send_mail and len(remove_objs) != 0:
+            if test:
+                receiver = 'snsd0805@cmlab.csie.ntu.edu.tw'
+            else:
+                receiver = f'{username}@cmlab.csie.ntu.edu.tw'
+            sender = 'unix_manager@cmlab.csie.ntu.edu.tw'
+            subject = f'[網管通知] CML Server 清理預計被移除的內容({username})'
+            msg = ''
+            for obj, date in remove_objs:
+                msg += f'<li>{date} - {obj}</li>'
+            msg = MAIL_MESSAGE.format(username, msg)
+
+            self.mailer.send(sender, receiver, subject, msg)

mailer.py

@@ -0,0 +1,27 @@
+import smtplib
+from email.mime.text import MIMEText
+
+class Mailer():
+    def __init__(self, google_owner, google_code):
+        self.google_owner = google_owner
+        self.google_code = google_code
+
+    def send(self, sender, receiver, subject, msg):
+        msg = MIMEText(msg, 'html')
+        msg['Subject'] = subject
+        msg['From'] = sender
+        msg['To'] = receiver
+
+        smtp = smtplib.SMTP('smtp.gmail.com', 587)
+        smtp.ehlo()
+        smtp.starttls()
+        smtp.login(self.google_owner, self.google_code)
+
+        status = smtp.send_message(msg)
+        smtp.quit()
+
+        if status == {}:
+            return True
+        else:
+            return False
+

main.py

@@ -0,0 +1,101 @@
+'''
+
+Clean the direction in CML Server (tmp2 or home directory)
+
+
+Suppose that the CLEAN_PATH may seems like this.
+And we will only check the directories & files under the user directory with "depth=1"
+
+CLEAN_PATH
+    |
+    +---- user1 (user directory)
+    |       |
+    |       +--- file1 (depth=1, will be check)
+    |       |
+    |       +--- Project1 (depth=1, will be check)
+    |       |       |
+    |       |       +--- subset1-1 (depth=2, won't be check)
+    |       |
+    |       +--- Project2 (depth=1)
+    |               |
+    |               +--- subset2-1 (depth=2)
+    |               |
+    |               +--- subset2-2 (depth=2)
+    |
+    +---- user2 (user directory)
+            |
+            +--- Project1 (depth=1)
+                    |
+                    +--- subset1-1 (depth=2)
+
+'''
+
+from mailer import Mailer
+from cleaner import Cleaner
+import os
+
+
+CLEAN_PATH = "/home/master/12"
+KEEP_FILES = [".bashrc", ".bash_logout", ".bash_profile", "public-html", ".ssh", ".profile", "examples.desktop", ".bash_history", ".config", ".zshrc", ".vimrc", ".gnupg", ".oh-my-zsh", ".vim", ".viminfo", ".vscode-server", ".gitconfig", ".local", ".zsh_history", ".zsh_profile"]
+WHITE_LIST = ["mysql", "unix_manager", "unixmanager215", "gangxing", "how083012", "yulunliu", "amie12349", "minliwu", "chenyuyang", "unixmanager5xx"]
+
+DATE = '2024-02-01'
+
+GOOGLE_OWNER = "snsd0805@cmlab.csie.ntu.edu.tw"
+GOOGLE_CODE = os.environ['GOOGLE_CODE']
+
+
+def get_master():
+    ans = []
+    for year in os.listdir(f'/home/master/'):
+        if os.path.isdir(f'/home/master/{year}'):
+            for user in os.listdir(f'/home/master/{year}'):
+                ans.append(f'/home/master/{year}/{user}')
+    return ans
+
+def get_phd():
+    ans = []
+    for dir_name in os.listdir(f'/home/phd/'):
+        if dir_name.isnumeric():
+            year = dir_name
+            for user in os.listdir(f'/home/phd/{year}'):
+                ans.append(f'/home/phd/{year}/{user}')
+        else:
+            ans.append(f'/home/phd/{dir_name}')
+    return ans
+
+def get_ra():
+    ans = []
+    for user in os.listdir(f'/home/ra/'):
+        ans.append(f'/home/ra/{user}')
+    return ans
+
+def get_extra():
+    ans = []
+    for user in os.listdir(f'/home/extra/'):
+        ans.append(f'/home/extra/{user}')
+    return ans
+
+
+def main(clean_type, test=True, send_mail=False):
+    mailer = Mailer(GOOGLE_OWNER, GOOGLE_CODE)
+
+    if clean_type == 'home':
+        student_list = []
+        for student_type in ['master', 'phd', 'ra', 'extra']:
+            func = globals()[f'get_{student_type}']
+            student_list += func()
+
+        for index, student_dir in enumerate(student_list):
+            print(f'{index+1}/{len(student_list)} {student_dir}')
+            cleaner = Cleaner(student_dir, "2024-02-01", mailer, WHITE_LIST, KEEP_FILES)
+            cleaner.run(test=test, send_mail=send_mail)
+
+    # mailer.send('unix_manager@cmlab.csie.ntu.edu.tw', 'snsd0805@cmlab.csie.ntu.edu.tw', '[測試郵件] test', 'Test mail')
+
+    # cleaner = Cleaner('/home/master/12', "2024/02/01", mailer, WHITE_LIST, KEEP_FILES)
+    # cleaner.run(test=test, send_mail=send_mail)
+
+
+if __name__ == "__main__":
+    main('home', test=False, send_mail=True)