filled in stub files

Added flags to messages, moved classes out of __init__.py, and documented that the inbox is where Imbox.messages are from without any folder argument.

.gitignore

@@ -31,3 +31,6 @@ nosetests.xml
 
 example.*
 example.py
+
+# PyCharm
+.idea/

CHANGELOG.md

@@ -1,3 +1,16 @@
+## 0.9.5 (5 December 2017)
+
+IMPROVEMENTS:
+
+ * `date__on` support: ([#109](https://github.com/martinrusev/imbox/pull/109))
+ * Starttls support: ([#108](https://github.com/martinrusev/imbox/pull/108))
+ * Mark emails as flagged/starred: ([#107](https://github.com/martinrusev/imbox/pull/107))
+ * Messages filter can use date objects instead of stringified dates: ([#104](https://github.com/martinrusev/imbox/pull/104))
+ * Fix attachment parsing when a semicolon character ends the Content-Disposition line: ([#100](https://github.com/martinrusev/imbox/pull/100))
+ * Parsing - UnicecodeDecodeError() fixes: ([#96](https://github.com/martinrusev/imbox/pull/96))
+ * Imbox() `with` support: ([#92](https://github.com/martinrusev/imbox/pull/92))
+
+
 ## 0.9 (18 September 2017)
 
 IMPROVEMENTS:

MANIFEST.in

@@ -1,3 +1,5 @@
 include LICENSE
 include MANIFEST.in
 include README.rst
+include CHANGELOG.md
+graft tests

README.rst

@@ -12,7 +12,7 @@ Python library for reading IMAP mailboxes and converting email content to machin
 Requirements
 ------------
 
-Python (3.2, 3.3, 3.4, 3.5, 3.6)
+Python (3.3, 3.4, 3.5, 3.6)
 
 
 Installation
@@ -34,32 +34,55 @@ Usage
             username='username',
             password='password',
             ssl=True,
-            ssl_context=None) as imbox:
+            ssl_context=None,
+            starttls=False) as imbox:
 
-        # Gets all messages
-        all_messages = imbox.messages()
+        # Get all folders
+        status, folders_with_additional_info = imbox.folders()
+
+        # Gets all messages from the inbox
+        all_inbox_messages = imbox.messages()
 
         # Unread messages
-        unread_messages = imbox.messages(unread=True)
+        unread_inbox_messages = imbox.messages(unread=True)
+
+        # Flagged messages
+        inbox_flagged_messages = imbox.messages(flagged=True)
+
+        # Un-flagged messages
+        inbox_unflagged_messages = imbox.messages(unflagged=True)
+
+        # Flagged messages
+        flagged_messages = imbox.messages(flagged=True)
+
+        # Un-flagged messages
+        unflagged_messages = imbox.messages(unflagged=True)
 
         # Messages sent FROM
-        messages_from = imbox.messages(sent_from='martin@amon.cx')
+        inbox_messages_from = imbox.messages(sent_from='sender@example.org')
 
         # Messages sent TO
-        messages_from = imbox.messages(sent_to='martin@amon.cx')
+        inbox_messages_to = imbox.messages(sent_to='receiver@example.org')
 
         # Messages received before specific date
-        messages_from = imbox.messages(date__lt='31-July-2013')
+        inbox_messages_received_before = imbox.messages(date__lt=datetime.date(2018, 7, 31))
 
         # Messages received after specific date
-        messages_from = imbox.messages(date__gt='30-July-2013')
+        inbox_messages_received_after = imbox.messages(date__gt=datetime.date(2018, 7, 30))
+
+        # Messages received on a specific date
+        inbox_messages_received_on_date = imbox.messages(date__on=datetime.date(2018, 7, 30))
 
         # Messages from a specific folder
-        messages_folder = imbox.messages(folder='Social')
+        messages_from_folder = imbox.messages(folder='Social')
 
+        # Messages whose subjects contain a string
+        inbox_messages_subject_christmas = imbox.messages(subject='Christmas')
 
+        # Messages from a specific folder
+        messages_in_folder_social = imbox.messages(folder='Social')
 
-        for uid, message in all_messages:
+        for uid, message in all_inbox_messages:
         # Every message is an object with the following keys
 
             message.sent_from
@@ -113,5 +136,34 @@ Usage
             'subject': u 'Hello John, How are you today'
             }
 
+        # With the message id, several actions on the message are available:
+        # delete the message
+        imbox.delete(uid)
+
+        # mark the message as read
+        imbox.mark_seen(uid)
+
+
+
+Changelog
+---------
 
 `Changelog <https://github.com/martinrusev/imbox/blob/master/CHANGELOG.md>`_
+
+
+Running the tests
+-----------------
+
+You can run the imbox tests with ``tox``.
+
+Requirements:
+ * the supported python versions
+ * ``tox``. Tox is packaged in Debian and derivatives distributions.
+
+On Ubuntu, you can install several python versions with:
+
+.. code:: sh
+
+    sudo add-apt-repository ppa:deadsnakes/ppa
+    sudo apt update
+    sudo apt install python3.X

imbox/__init__.py

@@ -1,89 +1,8 @@
-from imbox.imap import ImapTransport
-from imbox.parser import parse_email
-from imbox.query import build_search_query
+from imbox.imbox import Imbox
 
-import logging
-logger = logging.getLogger(__name__)
+__version_info__ = (0, 9, 5)
+__version__ = '.'.join([str(x) for x in __version_info__])
 
 
-class Imbox:
+__all__ = ['Imbox']
 
-    def __init__(self, hostname, username=None, password=None, ssl=True,
-                 port=None, ssl_context=None, policy=None):
-
-        self.server = ImapTransport(hostname, ssl=ssl, port=port,
-                                    ssl_context=ssl_context)
-        self.hostname = hostname
-        self.username = username
-        self.password = password
-        self.parser_policy = policy
-        self.connection = self.server.connect(username, password)
-        logger.info("Connected to IMAP Server with user {username} on {hostname}{ssl}".format(
-            hostname=hostname, username=username, ssl=(" over SSL" if ssl else "")))
-
-    def __enter__(self):
-        return self
-
-    def __exit__(self, type, value, traceback):
-        self.logout()
-
-    def logout(self):
-        self.connection.close()
-        self.connection.logout()
-        logger.info("Disconnected from IMAP Server {username}@{hostname}".format(
-            hostname=self.hostname, username=self.username))
-
-    def query_uids(self, **kwargs):
-        query = build_search_query(**kwargs)
-        message, data = self.connection.uid('search', None, query)
-        if data[0] is None:
-            return []
-        return data[0].split()
-
-    def fetch_by_uid(self, uid):
-        message, data = self.connection.uid('fetch', uid, '(BODY.PEEK[])')
-        logger.debug("Fetched message for UID {}".format(int(uid)))
-        raw_email = data[0][1]
-
-        email_object = parse_email(raw_email, policy=self.parser_policy)
-
-        return email_object
-
-    def fetch_list(self, **kwargs):
-        uid_list = self.query_uids(**kwargs)
-        logger.debug("Fetch all messages for UID in {}".format(uid_list))
-
-        for uid in uid_list:
-            yield (uid, self.fetch_by_uid(uid))
-
-    def mark_seen(self, uid):
-        logger.info("Mark UID {} with \\Seen FLAG".format(int(uid)))
-        self.connection.uid('STORE', uid, '+FLAGS', '(\\Seen)')
-
-    def delete(self, uid):
-        logger.info("Mark UID {} with \\Deleted FLAG and expunge.".format(int(uid)))
-        mov, data = self.connection.uid('STORE', uid, '+FLAGS', '(\\Deleted)')
-        self.connection.expunge()
-
-    def copy(self, uid, destination_folder):
-        logger.info("Copy UID {} to {} folder".format(int(uid), str(destination_folder)))
-        return self.connection.uid('COPY', uid, destination_folder)
-
-    def move(self, uid, destination_folder):
-        logger.info("Move UID {} to {} folder".format(int(uid), str(destination_folder)))
-        if self.copy(uid, destination_folder):
-            self.delete(uid)
-
-    def messages(self, *args, **kwargs):
-        folder = kwargs.get('folder', False)
-        msg = ""
-
-        if folder:
-            self.connection.select(folder)
-            msg = " from folder '{}'".format(folder)
-
-        logger.info("Fetch list of messages{}".format(msg))
-        return self.fetch_list(**kwargs)
-
-    def folders(self):
-        return self.connection.list()

imbox/imap.py

@@ -8,24 +8,20 @@ logger = logging.getLogger(__name__)
 
 class ImapTransport:
 
-    def __init__(self, hostname, port=None, ssl=True, ssl_context=None):
+    def __init__(self, hostname, port=None, ssl=True, ssl_context=None, starttls=False):
         self.hostname = hostname
-        self.port = port
-        kwargs = {}
 
         if ssl:
-            self.transport = IMAP4_SSL
-            if not self.port:
-                self.port = 993
+            self.port = port or 993
             if ssl_context is None:
                 ssl_context = pythonssllib.create_default_context()
-            kwargs["ssl_context"] = ssl_context
+            self.server = IMAP4_SSL(self.hostname, self.port, ssl_context=ssl_context)
         else:
-            self.transport = IMAP4
-            if not self.port:
-                self.port = 143
+            self.port = port or 143
+            self.server = IMAP4(self.hostname, self.port)
 
-        self.server = self.transport(self.hostname, self.port, **kwargs)
+        if starttls:
+            self.server.starttls()
         logger.debug("Created IMAP4 transport for {host}:{port}"
                      .format(host=self.hostname, port=self.port))
 

imbox/imap.pyi

@@ -0,0 +1,13 @@
+from imaplib import IMAP4, IMAP4_SSL
+from ssl import SSLContext
+from typing import Optional, Union, Tuple, List
+
+
+class ImapTransport:
+
+    def __init__(self, hostname: str, port: Optional[int], ssl: bool,
+                 ssl_context: Optional[SSLContext], starttls: bool) -> None: ...
+
+    def list_folders(self) -> Tuple[str, List[bytes]]: ...
+
+    def connect(self, username: str, password: str) -> Union[IMAP4, IMAP4_SSL]: ...

imbox/imbox.py

@@ -0,0 +1,72 @@
+from imbox.imap import ImapTransport
+from imbox.messages import Messages
+
+import logging
+
+logger = logging.getLogger(__name__)
+
+
+class Imbox:
+
+    def __init__(self, hostname, username=None, password=None, ssl=True,
+                 port=None, ssl_context=None, policy=None, starttls=False):
+
+        self.server = ImapTransport(hostname, ssl=ssl, port=port,
+                                    ssl_context=ssl_context, starttls=starttls)
+        self.hostname = hostname
+        self.username = username
+        self.password = password
+        self.parser_policy = policy
+        self.connection = self.server.connect(username, password)
+        logger.info("Connected to IMAP Server with user {username} on {hostname}{ssl}".format(
+            hostname=hostname, username=username, ssl=(" over SSL" if ssl or starttls else "")))
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, type, value, traceback):
+        self.logout()
+
+    def logout(self):
+        self.connection.close()
+        self.connection.logout()
+        logger.info("Disconnected from IMAP Server {username}@{hostname}".format(
+            hostname=self.hostname, username=self.username))
+
+    def mark_seen(self, uid):
+        logger.info("Mark UID {} with \\Seen FLAG".format(int(uid)))
+        self.connection.uid('STORE', uid, '+FLAGS', '(\\Seen)')
+
+    def mark_flag(self, uid):
+        logger.info("Mark UID {} with \\Flagged FLAG".format(int(uid)))
+        self.connection.uid('STORE', uid, '+FLAGS', '(\\Flagged)')
+
+    def delete(self, uid):
+        logger.info("Mark UID {} with \\Deleted FLAG and expunge.".format(int(uid)))
+        self.connection.expunge()
+
+    def copy(self, uid, destination_folder):
+        logger.info("Copy UID {} to {} folder".format(int(uid), str(destination_folder)))
+        return self.connection.uid('COPY', uid, destination_folder)
+
+    def move(self, uid, destination_folder):
+        logger.info("Move UID {} to {} folder".format(int(uid), str(destination_folder)))
+        if self.copy(uid, destination_folder):
+            self.delete(uid)
+
+    def messages(self, **kwargs):
+        folder = kwargs.get('folder', False)
+
+        if folder:
+            self.connection.select(folder)
+            msg = " from folder '{}'".format(folder)
+        else:
+            msg = " from inbox"
+
+        logger.info("Fetch list of messages{}".format(msg))
+        return Messages(connection=self.connection,
+                        parser_policy=self.parser_policy,
+                        **kwargs)
+
+    def folders(self):
+        return self.connection.list()

imbox/imbox.pyi

@@ -0,0 +1,31 @@
+import datetime
+from email._policybase import Policy
+from inspect import Traceback
+from ssl import SSLContext
+from typing import Optional, Union, Tuple, List
+
+
+class Imbox:
+
+    def __init__(self, hostname: str, username: Optional[str], password: Optional[str], ssl: bool,
+                 port: Optional[int], ssl_context: Optional[SSLContext], policy: Optional[Policy], starttls: bool): ...
+
+    def __enter__(self) -> 'Imbox': ...
+
+    def __exit__(self, type: Exception, value: str, traceback: Traceback) -> None: ...
+
+    def logout(self) -> None: ...
+
+    def mark_seen(self, uid: bytes) -> None: ...
+
+    def mark_flag(self, uid: bytes) -> None: ...
+
+    def delete(self, uid: bytes) -> None: ...
+
+    def copy(self, uid: bytes, destination_folder: Union[bytes, str]) -> Tuple[str, Union[list, List[None, bytes]]]: ...
+
+    def move(self, uid: bytes, destination_folder: Union[bytes, str]) -> None: ...
+
+    def messages(self, **kwargs: Union[bool, str, datetime.date]) -> 'Messages': ...
+
+    def folders(self) -> Tuple[str, List[bytes]]: ...

imbox/messages.py

@@ -0,0 +1,62 @@
+from imbox.parser import fetch_email_by_uid
+from imbox.query import build_search_query
+
+import logging
+
+logger = logging.getLogger(__name__)
+
+
+class Messages:
+
+    def __init__(self,
+                 connection,
+                 parser_policy,
+                 **kwargs):
+
+        self.connection = connection
+        self.parser_policy = parser_policy
+        self.kwargs = kwargs
+        self._uid_list = self._query_uids(**kwargs)
+
+        logger.debug("Fetch all messages for UID in {}".format(self._uid_list))
+
+    def _fetch_email(self, uid):
+        return fetch_email_by_uid(uid=uid,
+                                  connection=self.connection,
+                                  parser_policy=self.parser_policy)
+
+    def _query_uids(self, **kwargs):
+        query_ = build_search_query(**kwargs)
+        message, data = self.connection.uid('search', None, query_)
+        if data[0] is None:
+            return []
+        return data[0].split()
+
+    def _fetch_email_list(self):
+        for uid in self._uid_list:
+            yield uid, self._fetch_email(uid)
+
+    def __repr__(self):
+        if len(self.kwargs) > 0:
+            return 'Messages({})'.format('\n'.join('{}={}'.format(key, value)
+                                                   for key, value in self.kwargs.items()))
+        return 'Messages(ALL)'
+
+    def __iter__(self):
+        return self._fetch_email_list()
+
+    def __next__(self):
+        return self
+
+    def __len__(self):
+        return len(self._uid_list)
+
+    def __getitem__(self, index):
+        uids = self._uid_list[index]
+
+        if not isinstance(uids, list):
+            uid = uids
+            return uid, self._fetch_email(uid)
+
+        return [(uid, self._fetch_email(uid))
+                for uid in uids]

imbox/messages.pyi

@@ -0,0 +1,28 @@
+import datetime
+from email._policybase import Policy
+from imaplib import IMAP4, IMAP4_SSL
+from typing import Union, List, Generator, Tuple
+
+
+class Messages:
+
+    def __init__(self,
+                 connection: Union[IMAP4, IMAP4_SSL],
+                 parser_policy: Policy,
+                 **kwargs: Union[bool, str, datetime.date]) -> None: ...
+
+    def _fetch_email(self, uid: bytes) -> 'Struct': ...
+
+    def _query_uids(self, **kwargs: Union[bool, str, datetime.date]) -> List[bytes]: ...
+
+    def _fetch_email_list(self) -> Generator[Tuple[bytes, 'Struct']]: ...
+
+    def __repr__(self) -> str: ...
+
+    def __iter__(self) -> Generator[Tuple[bytes, 'Struct']]: ...
+
+    def __next__(self) -> 'Messages': ...
+
+    def __len__(self) -> int: ...
+
+    def __getitem__(self, index) -> Union['Struct', List['Struct']]: ...

imbox/parser.py

@@ -1,14 +1,17 @@
+import imaplib
 import io
 import re
 import email
 import base64
 import quopri
+import sys
 import time
 from datetime import datetime
 from email.header import decode_header
 from imbox.utils import str_encode, str_decode
 
 import logging
+
 logger = logging.getLogger(__name__)
 
 
@@ -33,7 +36,10 @@ def decode_mail_header(value, default_charset='us-ascii'):
         return str_decode(str_encode(value, default_charset, 'replace'), default_charset)
     else:
         for index, (text, charset) in enumerate(headers):
-            logger.debug("Mail header no. {}: {} encoding {}".format(index, str_decode(text, charset or 'utf-8'), charset))
+            logger.debug("Mail header no. {index}: {data} encoding {charset}".format(
+                index=index,
+                data=str_decode(text, charset or 'utf-8', 'replace'),
+                charset=charset))
             try:
                 headers[index] = str_decode(text, charset or default_charset,
                                             'replace')
@@ -54,7 +60,7 @@ def get_mail_addresses(message, header_name):
     for index, (address_name, address_email) in enumerate(addresses):
         addresses[index] = {'name': decode_mail_header(address_name),
                             'email': address_email}
-        logger.debug("{} Mail addressees in message: <{}> {}".format(header_name.upper(), address_name, address_email))
+        logger.debug("{} Mail address in message: <{}> {}".format(header_name.upper(), address_name, address_email))
     return addresses
 
 
@@ -63,13 +69,13 @@ def decode_param(param):
     values = v.split('\n')
     value_results = []
     for value in values:
-        match = re.search(r'=\?((?:\w|-)+)\?(Q|B)\?(.+)\?=', value)
+        match = re.search(r'=\?((?:\w|-)+)\?([QB])\?(.+)\?=', value)
         if match:
             encoding, type_, code = match.groups()
             if type_ == 'Q':
                 value = quopri.decodestring(code)
             elif type_ == 'B':
-                value = base64.decodestring(code)
+                value = base64.decodebytes(code.encode())
             value = str_encode(value, encoding)
             value_results.append(value)
             if value_results:
@@ -82,7 +88,11 @@ def parse_attachment(message_part):
     # Check again if this is a valid attachment
     content_disposition = message_part.get("Content-Disposition", None)
     if content_disposition is not None and not message_part.is_multipart():
-        dispositions = content_disposition.strip().split(";")
+        dispositions = [
+            disposition.strip()
+            for disposition in content_disposition.split(";")
+            if disposition.strip()
+        ]
 
         if dispositions[0].lower() in ["attachment", "inline"]:
             file_data = message_part.get_payload(decode=True)
@@ -97,13 +107,14 @@ def parse_attachment(message_part):
                 attachment['filename'] = filename
 
             for param in dispositions[1:]:
-                name, value = decode_param(param)
+                if param:
+                    name, value = decode_param(param)
 
-                if 'file' in name:
-                    attachment['filename'] = value
+                    if 'file' in name:
+                        attachment['filename'] = value[1:-1] if value.startswith('"') else value
 
-                if 'create-date' in name:
-                    attachment['create-date'] = value
+                    if 'create-date' in name:
+                        attachment['create-date'] = value
 
             return attachment
 
@@ -119,9 +130,31 @@ def decode_content(message):
         return content
 
 
+def fetch_email_by_uid(uid, connection, parser_policy):
+    message, data = connection.uid('fetch', uid, '(BODY.PEEK[] FLAGS)')
+    logger.debug("Fetched message for UID {}".format(int(uid)))
+
+    raw_headers, raw_email = data[0]
+
+    email_object = parse_email(raw_email, policy=parser_policy)
+    flags = parse_flags(raw_headers.decode())
+    email_object.__dict__['flags'] = flags
+
+    return email_object
+
+
+def parse_flags(headers):
+    """Copied from https://github.com/girishramnani/gmail/blob/master/gmail/message.py"""
+    if len(headers) == 0:
+        return []
+    if sys.version_info[0] == 3:
+        headers = bytes(headers, "ascii")
+    return list(imaplib.ParseFlags(headers))
+
+
 def parse_email(raw_email, policy=None):
     if isinstance(raw_email, bytes):
-        raw_email = str_encode(raw_email, 'utf-8')
+        raw_email = str_encode(raw_email, 'utf-8', errors='ignore')
     if policy is not None:
         email_parse_kwargs = dict(policy=policy)
     else:
@@ -132,9 +165,7 @@ def parse_email(raw_email, policy=None):
     except UnicodeEncodeError:
         email_message = email.message_from_string(raw_email.encode('utf-8'), **email_parse_kwargs)
     maintype = email_message.get_content_maintype()
-    parsed_email = {}
-
-    parsed_email['raw_email'] = raw_email
+    parsed_email = {'raw_email': raw_email}
 
     body = {
         "plain": [],
@@ -154,7 +185,7 @@ def parse_email(raw_email, policy=None):
                 content = decode_content(part)
 
             is_inline = content_disposition is None \
-                or content_disposition == "inline"
+                        or content_disposition.startswith("inline")
             if content_type == "text/plain" and is_inline:
                 body['plain'].append(content)
             elif content_type == "text/html" and is_inline:

imbox/parser.pyi

@@ -0,0 +1,30 @@
+import datetime
+from email._policybase import Policy
+from email.message import Message
+from imaplib import IMAP4_SSL
+import io
+from typing import Union, Dict, List, KeysView, Tuple, Optional
+
+
+class Struct:
+    def __init__(self, **entries: Union[
+        str, datetime.datetime, Dict[str, str], list, List[Dict[str, str]]
+    ]) -> None: ...
+
+    def keys(self) -> KeysView: ...
+
+    def __repr__(self) -> str: ...
+
+def decode_mail_header(value: str, default_charset: str) -> str: ...
+
+def get_mail_addresses(message: Message, header_name: str) -> List[Dict[str, str]]: ...
+
+def decode_param(param: str) -> Tuple[str, str]: ...
+
+def parse_attachment(message_part: Message) -> Optional[Dict[str, Union[int, str, io.BytesIO]]]: ...
+
+def decode_content(message: Message) -> str: ...
+
+def fetch_email_by_uid(uid: bytes, connection: IMAP4_SSL, parser_policy: Optional[Policy]) -> Struct: ...
+
+def parse_email(raw_email: bytes, policy: Optional[Policy]) -> Struct: ...

imbox/query.py

@@ -4,27 +4,24 @@ import logging
 
 logger = logging.getLogger(__name__)
 
-IMAP_MONTHS = ["Jan", "Feb", "Mar", "Apr", "May", "Jun",
-               "Jul", "Aug", "Sep", "Oct", "Nov", "Dec"]
-
 
 def format_date(date):
-
-    return "%s-%s-%s" % (date.day, IMAP_MONTHS[date.month - 1], date.year)
+    if isinstance(date, datetime.date):
+        return date.strftime('%d-%b-%Y')
+    return date
 
 
 def build_search_query(**kwargs):
 
     # Parse keyword arguments
     unread = kwargs.get('unread', False)
+    unflagged = kwargs.get('unflagged', False)
+    flagged = kwargs.get('flagged', False)
     sent_from = kwargs.get('sent_from', False)
     sent_to = kwargs.get('sent_to', False)
     date__gt = kwargs.get('date__gt', False)
-    if type(date__gt) is datetime.date:
-        date__gt = format_date(date__gt)
     date__lt = kwargs.get('date__lt', False)
-    if type(date__lt) is datetime.date:
-        date__lt = format_date(date__lt)
+    date__on = kwargs.get('date__on', False)
     subject = kwargs.get('subject')
 
     query = []
@@ -32,6 +29,12 @@ def build_search_query(**kwargs):
     if unread:
         query.append("(UNSEEN)")
 
+    if unflagged:
+        query.append("(UNFLAGGED)")
+
+    if flagged:
+        query.append("(FLAGGED)")
+
     if sent_from:
         query.append('(FROM "%s")' % sent_from)
 
@@ -39,10 +42,13 @@ def build_search_query(**kwargs):
         query.append('(TO "%s")' % sent_to)
 
     if date__gt:
-        query.append('(SINCE "%s")' % date__gt)
+        query.append('(SINCE "%s")' % format_date(date__gt))
 
     if date__lt:
-        query.append('(BEFORE "%s")' % date__lt)
+        query.append('(BEFORE "%s")' % format_date(date__lt))
+    
+    if date__on:
+        query.append('(ON "%s")' % format_date(date__on))
 
     if subject is not None:
         query.append('(SUBJECT "%s")' % subject)

imbox/query.pyi

@@ -0,0 +1,7 @@
+import datetime
+from typing import Union
+
+
+def format_date(date: Union[str, datetime.date]) -> str: ...
+
+def build_search_query(**kwargs: Union[bool, str, datetime.date]) -> str: ...

imbox/utils.py

@@ -1,14 +1,16 @@
 import logging
 logger = logging.getLogger(__name__)
 
+
 def str_encode(value='', encoding=None, errors='strict'):
     logger.debug("Encode str {} with and errors {}".format(value, encoding, errors))
     return str(value, encoding, errors)
 
+
 def str_decode(value='', encoding=None, errors='strict'):
     if isinstance(value, str):
         return bytes(value, encoding, errors).decode('utf-8')
     elif isinstance(value, bytes):
         return value.decode(encoding or 'utf-8', errors=errors)
     else:
-        raise TypeError( "Cannot decode '{}' object".format(value.__class__) )
+        raise TypeError("Cannot decode '{}' object".format(value.__class__))

imbox/utils.pyi

@@ -0,0 +1,6 @@
+from typing import Optional, Union
+
+
+def str_encode(value: Union[str, bytes], encoding: Optional[str], errors: str) -> str: ...
+
+def str_decode(value: Union[str, bytes], encoding: Optional[str], errors: str) -> Union[str, bytes]: ...

setup.py

@@ -1,7 +1,9 @@
 from setuptools import setup
 import os
 
-version = '0.9'
+import imbox
+
+version = imbox.__version__
 
 
 def read(filename):
@@ -22,10 +24,10 @@ setup(
     zip_safe=False,
     classifiers=(
         'Programming Language :: Python',
-        'Programming Language :: Python :: 3.2',
         'Programming Language :: Python :: 3.3',
         'Programming Language :: Python :: 3.4',
         'Programming Language :: Python :: 3.5',
         'Programming Language :: Python :: 3.6'
     ),
+    test_suite='tests',
 )

tests/8422.msg

@@ -0,0 +1,22 @@
+Delivered-To: receiver@example.com
+Return-Path: <sender@example.com>
+Date: Thu, 20 Jul 2017 07:34:22 -0500
+Message-ID: <59705CFE.A95F.0016.0@journeys.com>
+Subject: Following up Re: Looking to connect, let's schedule a call!
+From: sender@example.com
+To: "Receiver" <receiver@example.com>
+Mime-Version: 1.0
+Content-Type: multipart/mixed; boundary="=__PartBD85995F.0__="
+
+This is a MIME message. If you are reading this text, you may want to 
+consider changing to a mail reader or gateway that understands how to 
+properly handle MIME multipart messages.
+
+--=__PartBD85995F.0__=
+Content-Type: multipart/alternative; boundary="=__PartBD85995F.1__="
+
+--=__PartBD85995F.1__=
+Content-Type: text/plain; charset=Windows-1252
+Content-Transfer-Encoding: 8bit
+
+Following up on my previous message. I’d love to connect you with

tests/parser_tests.py

@@ -1,6 +1,7 @@
 import unittest
 from imbox.parser import *
 
+import os
 import sys
 if sys.version_info.minor < 3:
     SMTP = False
@@ -8,6 +9,9 @@ else:
     from email.policy import SMTP
 
 
+TEST_DIR = os.path.dirname(os.path.abspath(__file__))
+
+
 raw_email = """Delivered-To: johndoe@gmail.com
 X-Originating-Email: [martin@amon.cx]
 Message-ID: <test0@example.com>
@@ -80,6 +84,197 @@ Content-Transfer-Encoding: quoted-printable
 """
 
 
+raw_email_encoded_multipart = b"""Delivered-To: receiver@example.com
+Return-Path: <kkoudelka@wallvet.com>
+Date: Tue, 08 Aug 2017 08:15:11 -0700
+From: <kkoudelka@wallvet.com>
+To: interviews+347243@gethappie.me
+Message-Id: <20170808081511.2b876c018dd94666bcc18e28cf079afb.99766f164b.wbe@email24.godaddy.com>
+Subject: RE: Kari, are you open to this?
+Mime-Version: 1.0
+Content-Type: multipart/related;
+	boundary="=_7c18e0b95b772890a22ed6c0f810a434"
+
+--=_7c18e0b95b772890a22ed6c0f810a434
+Content-Transfer-Encoding: quoted-printable
+Content-Type: text/html; charset="utf-8"
+
+<html><body><span style=3D"font-family:Verdana; color:#000; font-size:10pt;=
+"><div>Hi Richie,</div></span></body></html>
+--=_7c18e0b95b772890a22ed6c0f810a434
+Content-Transfer-Encoding: base64
+Content-Type: image/jpeg; charset=binary;
+ name="sigimg0"; 
+Content-Disposition: inline;
+ filename="sigimg0"; 
+
+/9j/4AAQSkZJRgABAQAAAQABAAD//gA+Q1JFQVRPUjogZ2QtanBlZyB2MS4wICh1c2luZyBJSkcg
+jt0JaKhjm3xq23GR60UuZBZn/9k=
+--=_7c18e0b95b772890a22ed6c0f810a434
+Content-Transfer-Encoding: base64
+Content-Type: image/jpeg; charset=binary;
+ name="sigimg1"; 
+Content-Disposition: inline;
+ filename="sigimg1"; 
+
+/9j/4AAQSkZJRgABAQAAAQABAAD//gA+Q1JFQVRPUjogZ2QtanBlZyB2MS4wICh1c2luZyBJSkcg
+SlBFRyB2NjIpLCBkZWZhdWx0IHF1YWxpdHkK/9sAQwAIBgYHBgUIBwcHCQkICgwUDQwLCwwZEhMP
+ooooA//Z
+--=_7c18e0b95b772890a22ed6c0f810a434--
+
+"""
+
+
+raw_email_encoded_bad_multipart = b"""Delivered-To: receiver@example.com
+Return-Path: <sender@example.com>
+From: sender@example.com
+To: "Receiver" <receiver@example.com>, "Second\r\n Receiver" <recipient@example.com>
+Subject: Re: Looking to connect with you...
+Date: Thu, 20 Apr 2017 15:32:52 +0000
+Message-ID: <BN6PR16MB179579288933D60C4016D078C31B0@BN6PR16MB1795.namprd16.prod.outlook.com>
+Content-Type: multipart/related;
+	boundary="_004_BN6PR16MB179579288933D60C4016D078C31B0BN6PR16MB1795namp_";
+	type="multipart/alternative"
+MIME-Version: 1.0
+--_004_BN6PR16MB179579288933D60C4016D078C31B0BN6PR16MB1795namp_
+Content-Type: multipart/alternative;
+	boundary="_000_BN6PR16MB179579288933D60C4016D078C31B0BN6PR16MB1795namp_"
+--_000_BN6PR16MB179579288933D60C4016D078C31B0BN6PR16MB1795namp_
+Content-Type: text/plain; charset="utf-8"
+Content-Transfer-Encoding: base64
+SGkgRGFuaWVsbGUsDQoNCg0KSSBhY3R1YWxseSBhbSBoYXBweSBpbiBteSBjdXJyZW50IHJvbGUs
+Y3J1aXRlciB8IENoYXJsb3R0ZSwgTkMNClNlbnQgdmlhIEhhcHBpZQ0KDQoNCg==
+--_000_BN6PR16MB179579288933D60C4016D078C31B0BN6PR16MB1795namp_
+Content-Type: text/html; charset="utf-8"
+Content-Transfer-Encoding: base64
+PGh0bWw+DQo8aGVhZD4NCjxtZXRhIGh0dHAtZXF1aXY9IkNvbnRlbnQtVHlwZSIgY29udGVudD0i
+CjwvZGl2Pg0KPC9kaXY+DQo8L2JvZHk+DQo8L2h0bWw+DQo=
+--_000_BN6PR16MB179579288933D60C4016D078C31B0BN6PR16MB1795namp_--
+--_004_BN6PR16MB179579288933D60C4016D078C31B0BN6PR16MB1795namp_
+Content-Type: image/png; name="=?utf-8?B?T3V0bG9va0Vtb2ppLfCfmIoucG5n?="
+Content-Description: =?utf-8?B?T3V0bG9va0Vtb2ppLfCfmIoucG5n?=
+Content-Disposition: inline;
+	filename="=?utf-8?B?T3V0bG9va0Vtb2ppLfCfmIoucG5n?="; size=488;
+	creation-date="Thu, 20 Apr 2017 15:32:52 GMT";
+	modification-date="Thu, 20 Apr 2017 15:32:52 GMT"
+Content-ID: <254962e2-f05c-40d1-aa11-0d34671b056c>
+Content-Transfer-Encoding: base64
+iVBORw0KGgoAAAANSUhEUgAAABMAAAATCAYAAAByUDbMAAAAGXRFWHRTb2Z0d2FyZQBBZG9iZSBJ
+cvED9AIR3TCAAAMAqh+p+YMVeBQAAAAASUVORK5CYII=
+--_004_BN6PR16MB179579288933D60C4016D078C31B0BN6PR16MB1795namp_--
+"""
+
+
+raw_email_encoded_another_bad_multipart = b"""Delivered-To: receiver@example.com
+Return-Path: <sender@example.com>
+Mime-Version: 1.0
+Date: Wed, 22 Mar 2017 15:21:55 -0500
+Message-ID: <58D29693.192A.0075.1@wimort.com>
+Subject: Re: Reaching Out About Peoples Home Equity
+From: sender@example.com
+To: receiver@example.com
+Content-Type: multipart/alternative; boundary="____NOIBTUQXSYRVOOAFLCHY____"
+
+
+--____NOIBTUQXSYRVOOAFLCHY____
+Content-Type: text/plain; charset=iso-8859-15
+Content-Transfer-Encoding: quoted-printable
+Content-Disposition: inline;
+	modification-date="Wed, 22 Mar 2017 15:21:55 -0500"
+
+Chloe,
+ 
+--____NOIBTUQXSYRVOOAFLCHY____
+Content-Type: multipart/related; boundary="____XTSWHCFJMONXSVGPVDLY____"
+
+
+--____XTSWHCFJMONXSVGPVDLY____
+Content-Type: text/html; charset=iso-8859-15
+Content-Transfer-Encoding: quoted-printable
+Content-Disposition: inline;
+	modification-date="Wed, 22 Mar 2017 15:21:55 -0500"
+
+<HTML xmlns=3D"http://www.w3.org/1999/xhtml">
+<BODY style=3D"COLOR: black; FONT: 10pt Segoe UI; MARGIN: 4px 4px 1px" =
+leftMargin=3D0 topMargin=3D0 offset=3D"0" marginwidth=3D"0" marginheight=3D=
+"0">
+<DIV>Chloe,</DIV>
+<IMG src=3D"cid:VFXVGHA=
+GXNMI.36b3148cbf284ba18d35bdd8386ac266" width=3D1 height=3D1> </BODY></HTML=
+>
+--____XTSWHCFJMONXSVGPVDLY____
+Content-ID: <TLUACRGXVUBY.IMAGE_3.gif>
+Content-Type: image/gif
+Content-Transfer-Encoding: base64
+
+R0lGODlhHgHCAPf/AIOPr9GvT7SFcZZjVTEuMLS1tZKUlJN0Znp4eEA7PV1aWvz8+8V6Zl1BNYxX
+HvOZ1/zmOd95agUEADs=
+--____XTSWHCFJMONXSVGPVDLY____
+Content-ID: <VFXVGHAGXNMI.36b3148cbf284ba18d35bdd8386ac266>
+Content-Type: image/xxx
+Content-Transfer-Encoding: base64
+
+R0lGODlhAQABAPAAAAAAAAAAACH5BAEAAAAALAAAAAABAAEAAAICRAEAOw==
+--____XTSWHCFJMONXSVGPVDLY____--
+
+--____NOIBTUQXSYRVOOAFLCHY____--
+"""
+
+
+raw_email_with_trailing_semicolon_to_disposition_content = b"""Delivered-To: receiver@example.com
+Return-Path: <sender@example.com>
+Mime-Version: 1.0
+Date: Wed, 22 Mar 2017 15:21:55 -0500
+Message-ID: <58D29693.192A.0075.1@wimort.com>
+Subject: Re: Reaching Out About Peoples Home Equity
+From: sender@example.com
+To: receiver@example.com
+Content-Type: multipart/alternative; boundary="____NOIBTUQXSYRVOOAFLCHY____"
+
+
+--____NOIBTUQXSYRVOOAFLCHY____
+Content-Type: text/plain; charset=iso-8859-15
+Content-Transfer-Encoding: quoted-printable
+Content-Disposition: inline;
+	modification-date="Wed, 22 Mar 2017 15:21:55 -0500"
+
+Hello Chloe
+
+--____NOIBTUQXSYRVOOAFLCHY____
+Content-Type: multipart/related; boundary="____XTSWHCFJMONXSVGPVDLY____"
+
+
+--____XTSWHCFJMONXSVGPVDLY____
+Content-Type: text/html; charset=iso-8859-15
+Content-Transfer-Encoding: quoted-printable
+Content-Disposition: inline;
+	modification-date="Wed, 22 Mar 2017 15:21:55 -0500"
+
+<HTML xmlns=3D"http://www.w3.org/1999/xhtml">
+<BODY>
+<DIV>Hello Chloe</DIV>
+</BODY>
+</HTML>
+--____XTSWHCFJMONXSVGPVDLY____
+Content-Type: application/octet-stream; name="abc.xyz"
+Content-Description: abc.xyz
+Content-Disposition: attachment; filename="abc.xyz";
+Content-Transfer-Encoding: base64
+
+R0lGODlhHgHCAPf/AIOPr9GvT7SFcZZjVTEuMLS1tZKUlJN0Znp4eEA7PV1aWvz8+8V6Zl1BNYxX
+HvOZ1/zmOd95agUEADs=
+--____XTSWHCFJMONXSVGPVDLY____
+Content-ID: <VFXVGHAGXNMI.36b3148cbf284ba18d35bdd8386ac266>
+Content-Type: image/xxx
+Content-Transfer-Encoding: base64
+
+R0lGODlhAQABAPAAAAAAAAAAACH5BAEAAAAALAAAAAABAAEAAAICRAEAOw==
+--____XTSWHCFJMONXSVGPVDLY____--
+
+--____NOIBTUQXSYRVOOAFLCHY____--
+"""
+
+
 class TestParser(unittest.TestCase):
 
     def test_parse_email(self):
@@ -96,16 +291,40 @@ class TestParser(unittest.TestCase):
         self.assertEqual('Выписка по карте', parsed_email.subject)
         self.assertEqual('Выписка по карте 1234', parsed_email.body['html'][0])
 
+    def test_parse_email_invalid_unicode(self):
+        parsed_email = parse_email(open(os.path.join(TEST_DIR, '8422.msg'), 'rb').read())
+        self.assertEqual("Following up Re: Looking to connect, let's schedule a call!", parsed_email.subject)
+
+    def test_parse_email_inline_body(self):
+        parsed_email = parse_email(raw_email_encoded_another_bad_multipart)
+        self.assertEqual("Re: Reaching Out About Peoples Home Equity", parsed_email.subject)
+        self.assertTrue(parsed_email.body['plain'])
+        self.assertTrue(parsed_email.body['html'])
+
+    def test_parse_email_multipart(self):
+        parsed_email = parse_email(raw_email_encoded_multipart)
+        self.assertEqual("RE: Kari, are you open to this?", parsed_email.subject)
+
+    def test_parse_email_bad_multipart(self):
+        parsed_email = parse_email(raw_email_encoded_bad_multipart)
+        self.assertEqual("Re: Looking to connect with you...", parsed_email.subject)
+
     def test_parse_email_ignores_header_casing(self):
         self.assertEqual('one', parse_email('Message-ID: one').message_id)
         self.assertEqual('one', parse_email('Message-Id: one').message_id)
         self.assertEqual('one', parse_email('Message-id: one').message_id)
         self.assertEqual('one', parse_email('message-id: one').message_id)
 
-    # TODO - Complete the test suite
     def test_parse_attachment(self):
-        pass
+        parsed_email = parse_email(raw_email_with_trailing_semicolon_to_disposition_content)
+        self.assertEqual(1, len(parsed_email.attachments))
+        attachment = parsed_email.attachments[0]
+        self.assertEqual('application/octet-stream', attachment['content-type'])
+        self.assertEqual(71, attachment['size'])
+        self.assertEqual('abc.xyz', attachment['filename'])
+        self.assertTrue(attachment['content'])
 
+    # TODO - Complete the test suite
     def test_decode_mail_header(self):
         pass
 
@@ -117,6 +336,9 @@ class TestParser(unittest.TestCase):
         from_message_object = email.message_from_string("From: John Smith <johnsmith@gmail.com>")
         self.assertEqual([{'email': 'johnsmith@gmail.com', 'name': 'John Smith'}], get_mail_addresses(from_message_object, 'from'))
 
+        invalid_encoding_in_from_message_object = email.message_from_string("From: =?UTF-8?Q?C=E4cilia?= <caciliahxg827m@example.org>")
+        self.assertEqual([{'email': 'caciliahxg827m@example.org', 'name': 'C<EFBFBD>cilia'}], get_mail_addresses(invalid_encoding_in_from_message_object, 'from'))
+
     def test_parse_email_with_policy(self):
         if not SMTP:
             return

tests/query_tests.py

@@ -15,22 +15,36 @@ class TestQuery(unittest.TestCase):
         res = build_search_query(unread=True)
         self.assertEqual(res, "(UNSEEN)")
 
+    def test_unflagged(self):
+
+        res = build_search_query(unflagged=True)
+        self.assertEqual(res, "(UNFLAGGED)")
+
+    def test_flagged(self):
+
+        res = build_search_query(flagged=True)
+        self.assertEqual(res, "(FLAGGED)")
+
     def test_sent_from(self):
 
         res = build_search_query(sent_from='test@example.com')
-        self.assertEqual(res, "(FROM \"test@example.com\")")
+        self.assertEqual(res, '(FROM "test@example.com")')
 
     def test_sent_to(self):
 
         res = build_search_query(sent_to='test@example.com')
-        self.assertEqual(res, "(TO \"test@example.com\")")
+        self.assertEqual(res, '(TO "test@example.com")')
 
     def test_date__gt(self):
 
         res = build_search_query(date__gt=date(2014, 12, 31))
-        self.assertEqual(res, "(SINCE \"31-Dec-2014\")")
+        self.assertEqual(res, '(SINCE "31-Dec-2014")')
 
     def test_date__lt(self):
 
         res = build_search_query(date__lt=date(2014, 1, 1))
-        self.assertEqual(res, "(BEFORE \"1-Jan-2014\")")
+        self.assertEqual(res, '(BEFORE "01-Jan-2014")')
+
+    def test_date__on(self):
+        res = build_search_query(date__on=date(2014, 1, 1))
+        self.assertEqual(res, '(ON "01-Jan-2014")')

tox.ini

@@ -0,0 +1,6 @@
+[tox]
+envlist = py33,py34,py35,py36
+
+[testenv]
+deps=nose
+commands=nosetests -v