From 6af6eed48dec34eb9c64c3bed28b27bf156b4b72 Mon Sep 17 00:00:00 2001 From: Stacy Brock Date: Wed, 14 Apr 2021 14:43:34 -0700 Subject: [PATCH] Improve filter-rules.py Changelist: - Add init_filters() func as an entry point when filters are reloaded from a file. This is a good place to put any code that should be loaded once instead of on a per-message basis. - Added helper functions for moving messages and checking if a message has any matches in a list of strings. - Created local object caches for normalized keyword lists and mailbox folder IDs. This puts the data closer to where it is needed, and removes the need for passing 'self' around when filtering. - Added filter rule to catch unactionable email. --- filter-rules.py | 109 +++++++++++++++++++++++++++++++++--------------- 1 file changed, 76 insertions(+), 33 deletions(-) diff --git a/filter-rules.py b/filter-rules.py index 66ef6c6..b831675 100644 --- a/filter-rules.py +++ b/filter-rules.py @@ -11,29 +11,19 @@ BLOCK_EMAIL = [ ] BLOCK_KEYWORDS = [ - 'charitable fund drive', - 'election reminder', 'email preferences', 'end these email updates', - 'food drive', - 'food share', - 'general election', - 'giving tuesday', - 'hardship leave donations needed', 'manage your preferences', 'modify your preferences', 'opt-out', 'opt out', 'prefer not to receive', 'prefer not to see', - 'register to vote' 'Samsung SDS America', 'sidekickopen', - 'special election', 'subscription preferences', 'survey', 'unsubscribe', - 'voter registration', 'want to receive', 'webinar', 'whitepaper', @@ -42,7 +32,21 @@ BLOCK_KEYWORDS = [ BLOCK_DOMAINS = [ 'customeriomail.com', - 'mailgun.net' + 'mailgun.net', + 'norstar.net' +] + +BLOCK_ANNOYING = [ + 'charitable fund drive', + 'election reminder', + 'food drive', + 'food share', + 'general election', + 'giving tuesday', + 'hardship leave donations needed', + 'register to vote' + 'special election', + 'voter registration' ] ALLOW = [ @@ -65,11 +69,25 @@ ALLOW = [ 'microsoft.com' ] +normalized = {} +folder_cache = {} + def filter_message(self, message): # normalize message attributes normalized_to = [x.address.lower() for x in message.to] normalized_from = message.sender.address.lower() - normalized_subject = message.subject.lower() + + # filter unactionable coresys emails + if 'coresys@lists.oregonstate.edu' in normalized_from: + unactionable = [x.lower() for x in [ + 'DWPRODRAW Verification', + 'BFPDB Reimbursement Job Monitor - OK', + 'DSDB Important Job Monitor - OK' + ]] + if is_in_message(unactionable, message): + self._log_result(message, 'moving to unactionable') + move_message(message, 'unactionable') + return # filter alerts-sig if (normalized_from in ['mcc-b11-stor1@oregonstate.edu', @@ -78,7 +96,7 @@ def filter_message(self, message): 'isilon@storage.sig.oregonstate.edu'] or 'alarm.DatastoreDiskUsageAlarm' in message.subject): self._log_result(message, 'moving to alerts-sig') - message.move(self._folders['alerts-sig']) + move_message(message, 'alerts-sig') return # filter conference spam @@ -90,7 +108,7 @@ def filter_message(self, message): # filter backup-nightly if 'backup-nightly@lists.oregonstate.edu' in message.to: self._log_result(message, 'moving to backup-nightly') - message.move(self._folders['backup-nightly']) + move_message(message, 'backup-nightly') return # delete HP alert spam @@ -99,6 +117,12 @@ def filter_message(self, message): message.delete() return + # delete other spam + if is_in_message(normalized['BLOCK_ANNOYING'], message): + self._log_result(message, 'junking spam containing annoying content') + move_message(message, 'Junk Email') + return + # keep messages from allowed emails and domains for good in ALLOW: if good in normalized_from and normalized_from not in BLOCK_EMAIL: @@ -109,22 +133,14 @@ def filter_message(self, message): # junk messages from blocked senders if normalized_from in BLOCK_EMAIL: self._log_result(message, 'junking spam from blocked sender') - message.move(self._folders['Junk Email']) + move_message(message, 'Junk Email') return - # junk messages with blocked keywords in message body - is_spam = False - message_body = message.body.lower() - for phrase in self._normalized['BLOCK_KEYWORDS']: - if phrase in normalized_subject: - is_spam = True - break - if phrase in message_body: - is_spam = True - break - if is_spam: + # junk messages with blocked keywords + if is_in_message(normalized['BLOCK_KEYWORDS'], message): self._log_result(message, 'junking spam containing blocked keyword') - message.move(self._folders['Junk Email']) + move_message(message, 'Junk Email') + return # process message headers into a sane data structure @@ -144,21 +160,49 @@ def filter_message(self, message): break if is_spam: self._log_result(message, 'junking spam from blocked domain') - message.move(self._folders['Junk Email']) + move_message(message, 'Junk Email') return # junk known spam headers if (get_header('X-Spam-Flag' == 'YES', headers) or int(get_header('X-MS-Exchange-Organization-SCL', headers)) >= 5): self._log_result(message, 'junking spam with known header') - message.move(self._folders['Junk Email']) + move_message(message, 'Junk Email') return # KEEP MESSAGE self._log_result(message, 'keeping message, passed all filter checks') -def normalize_lists(self): - self._normalized['BLOCK_KEYWORDS'] = [x.lower() for x in BLOCK_KEYWORDS] +def is_in_message(list_, message): + """search a message for a list of strings + + Returns True if any string in the list is found in the message + """ + is_found = False + message_body = message.body.lower() + message_subject = message.subject.lower() + + for term in list_: + if term in message_subject: + is_found = True + break + if term in message_body: + is_found = True + break + return is_found + +def move_message(message, folder): + if folder not in folder_cache: + return False + message.move(folder_cache[folder]) + +def init_filters(self): + # hack to copy a dict from parent object into local object + for k,v in self._folders.items(): + folder_cache[k] = v + + normalized['BLOCK_KEYWORDS'] = [x.lower() for x in BLOCK_KEYWORDS] + normalized['BLOCK_ANNOYING'] = [x.lower() for x in BLOCK_ANNOYING] def get_header(header_key, headers): vals = [] @@ -174,9 +218,8 @@ def get_header(header_key, headers): return False def search_headers(search, headers): - vals = [] - is_found = False for header in headers: for val in header.values(): if search in val: return True + return False