def retrieve_new_emails(client: IMAPClient,
user: EmailAuthUser,
logger: Optional[logging.Logger] = None) -> List[int]:
"""Retrieve new unseen emails for the user."""
logger = logger or get_logger(CLIENT)
mails: List[int] = []
try:
client.login(user.email_address, user.password)
logger.info("Logged in successful for %s", user)
except LoginError:
logger.exception("Credentials could not be verified for '%s'.",
user.email_address)
client.shutdown()
raise
except socket.error:
logger.exception("Could not connect to the mail server.")
client.shutdown()
raise
except Exception as exc_info: # pylint: disable=broad-except
logger.exception("Unknown Exception occured. %s", str(exc_info))
client.shutdown()
raise
else:
create_folder_if_not_exists(client, EmailFolders.IMPORTANT, logger)
create_folder_if_not_exists(client, EmailFolders.URGENT, logger)
search_key = b"UNSEEN UNKEYWORD " + FLAG_TO_CHECK
client.select_folder(EmailFolders.INBOX, readonly=True)
logger.debug("Searching for unseen emails flagged '%s'", FLAG_TO_CHECK)
mails = client.search(search_key)
return mails
def test_get_logger_on_logger_set_on_config(config_env_file):
"""Test get_logger return logger when no config is set."""
os.environ["ZIPPY_CONFIG_FILE"] = str(config_env_file.absolute().resolve())
from zippy.utils.log_handler import get_logger
logger = get_logger("testA")
assert logger.getEffectiveLevel() == logging.DEBUG
assert logger.name == "testA"
def test_get_logger_on_no_key_found(config_file):
"""Test get_logger if no module name exist in config."""
os.environ["ZIPPY_CONFIG_FILE"] = str(config_file.absolute().resolve())
from zippy.utils.log_handler import get_logger
logger = get_logger("testA")
assert logger.getEffectiveLevel() == logging.WARNING
assert logger.name == "testA"
def test_get_logger_positive(config_file):
"""Test get_logger happy path."""
os.environ["ZIPPY_CONFIG_FILE"] = str(config_file.absolute().resolve())
from zippy.utils.log_handler import get_logger
logger = get_logger("test")
assert logger.getEffectiveLevel() == logging.DEBUG
assert logger.name == "test"
def main(
config: dict,
user: EmailAuthUser,
client: Optional[IMAPClient] = None,
logger: Optional[logging.Logger] = None,
):
"""Handle all updates for a specific users."""
logger = logger or get_logger(CLIENT)
client = client or get_client(config, logger=logger)
with client:
unprocessed_mails = retrieve_new_emails(client, user, logger)
processed_messages = process_mails(client, unprocessed_mails, logger)
online_train_all(processed_messages)
def process_mail(client: IMAPClient, uid: int, message_data: dict,
logger: logging.Logger) -> ProcessedMessage:
"""Process mail."""
output = get_logger(OUTPUT)
email_message = email.message_from_bytes(message_data[MESSAGE_FORMAT])
*msg, threshold = rank_message(email_message)
processed_msg = ProcessedMessage(*msg)
output.info({
"rank": processed_msg.rank,
"important": processed_msg.important,
"intent": processed_msg.intent,
"subject": email_message["Subject"],
"uid": uid,
"from": email_message["From"],
"to": email_message["To"],
"threshold": threshold,
})
if processed_msg.important and not processed_msg.intent:
shift_mail(
client=client,
uid=uid,
source=EmailFolders.INBOX,
destination=EmailFolders.IMPORTANT,
logger=logger,
)
elif processed_msg.important and processed_msg.intent:
shift_mail(
client=client,
uid=uid,
source=EmailFolders.INBOX,
destination=EmailFolders.URGENT,
logger=logger,
)
else:
mark_processed(client=client, uid=uid, logger=logger)
return processed_msg
"""Module for online training new email messages."""
import pathlib
import nltk
import numpy as np
import pandas as pd
from sklearn.feature_extraction.text import CountVectorizer
from zippy.utils.log_handler import get_logger
LOGGER = get_logger("client")
MODEL_DIR = pathlib.Path(__file__).parents[3] / "output/models/simplerank"
try:
VEC = CountVectorizer(stop_words=nltk.corpus.stopwords.words("english"))
except LookupError:
nltk.download("stopwords")
VEC = CountVectorizer(stop_words=nltk.corpus.stopwords.words("english"))
def load_weights(user="******"):
"""Load weights from the CSV."""
from_wt = pd.read_csv(MODEL_DIR / user / "from_weight.csv")
thread_sender_wt = pd.read_csv(MODEL_DIR / user /
"thread_senders_weight.csv")
thread_wt = pd.read_csv(MODEL_DIR / user / "thread_weights.csv")
thread_term_wt = pd.read_csv(MODEL_DIR / user / "thread_term_weights.csv")
msg_term_wt = pd.read_csv(MODEL_DIR / user / "msg_terms_weight.csv")
def wrapper(*args, **kwargs):
log = logger or get_logger(CLIENT)
log.info("LOG: Running job %s", func.__name__)
result = func(*args, **kwargs)
log.info("LOG: Job '%s' completed", func.__name__)
return result