# Copyright (c) 2012 Web Notes Technologies Pvt Ltd (http://erpnext.com) # # MIT License (MIT) # # Permission is hereby granted, free of charge, to any person obtaining a # copy of this software and associated documentation files (the "Software"), # to deal in the Software without restriction, including without limitation # the rights to use, copy, modify, merge, publish, distribute, sublicense, # and/or sell copies of the Software, and to permit persons to whom the # Software is furnished to do so, subject to the following conditions: # # The above copyright notice and this permission notice shall be included in # all copies or substantial portions of the Software. # # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, # INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A # PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT # HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF # CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE # OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. # from __future__ import unicode_literals """ This module contains classes for managing incoming emails """ class IncomingMail: """ Single incoming email object. Extracts, text / html and attachments from the email """ def __init__(self, content): """ Parse the incoming mail content """ import email self.mail = email.message_from_string(content) self.text_content = '' self.html_content = '' self.attachments = [] self.parse() def get_text_content(self): """ Returns the text parts of the email. If None, then HTML parts """ return self.text_content or self.html_content def get_charset(self, part): """ Guesses character set """ charset = part.get_content_charset() if not charset: import chardet charset = chardet.detect(str(part))['encoding'] return charset def get_payload(self, part, charset): """ get utf-8 encoded part content """ try: return unicode(part.get_payload(decode=True),str(charset),"ignore") except LookupError, e: return part.get_payload() def get_attachment(self, part, charset): """ Extracts an attachment """ self.attachments.append({ 'content-type': part.get_content_type(), 'filename': part.get_filename(), 'content': part.get_payload(decode=True), }) def parse(self): """ Extracts text, html and attachments from the mail """ for part in self.mail.walk(): self.process_part(part) def get_thread_id(self): """ Extracts thread id of the message between first [] from the subject """ import re subject = self.mail.get('Subject', '') return re.findall('(?<=\[)[\w/-]+', subject) def process_part(self, part): """ Process a single part of an email """ content_type = part.get_content_type() charset = part.get_content_charset() if not charset: charset = self.get_charset(part) if content_type == 'text/plain': self.text_content += self.get_payload(part, charset) if content_type == 'text/html': self.html_content += self.get_payload(part, charset) if part.get_filename(): self.get_attachment(part, charset) class POP3Mailbox: """ A simple pop3 mailbox, abstracts connection and mail extraction To use, subclass it and override method process_message(from, subject, text, thread_id) """ def __init__(self, settings_doc): """ settings_doc must contain use_ssl, host, username, password (by name or object) """ if isinstance(settings_doc, basestring): from webnotes.model.doc import Document self.settings = Document(settings_doc, settings_doc) else: self.settings = settings_doc def connect(self): """ Connects to the mailbox """ import poplib if self.settings.use_ssl: self.pop = poplib.POP3_SSL(self.settings.host) else: self.pop = poplib.POP3(self.settings.host) self.pop.user(self.settings.username) self.pop.pass_(self.settings.password) def get_messages(self): """ Loads messages from the mailbox and calls process_message for each message """ import webnotes if not self.check_mails(): return # nothing to do self.connect() num = num_copy = len(self.pop.list()[1]) # track if errors arised errors = False # WARNING: Hard coded max no. of messages to be popped if num > 20: num = 20 for m in xrange(1, num+1): msg = self.pop.retr(m) try: webnotes.conn.begin() self.process_message(IncomingMail(b'\n'.join(msg[1]))) self.pop.dele(m) webnotes.conn.commit() except: from webnotes.utils.scheduler import log # log performs rollback and logs error in scheduler log log("receive.get_messages") errors = True webnotes.conn.begin() # WARNING: Delete message number 101 onwards from the pop list # This is to avoid having too many messages entering the system num = num_copy if num > 100 and not errors: for m in xrange(101, num+1): self.pop.dele(m) self.pop.quit() def check_mails(self): """ To be overridden If mailbox is to be scanned, returns true """ return True def process_message(self, mail): """ To be overriden """ pass