Compare commits

..

2 Commits

Author SHA1 Message Date
copilot-swe-agent[bot]
b1e76c8569 Fix IMAP since option date format to use RFC 3501 compliant DD-Mon-YYYY format
Co-authored-by: seanthegeek <44679+seanthegeek@users.noreply.github.com>
2025-12-29 17:49:27 +00:00
copilot-swe-agent[bot]
a5de313a57 Initial plan 2025-12-29 17:43:14 +00:00
9 changed files with 16 additions and 210 deletions

View File

@@ -1,18 +1,5 @@
# Changelog
## 9.0.8
### Fixes
- Fix logging configuration not propagating to child parser processes (#646).
- Update `mailsuite` dependency to `?=1.11.1` to solve issues with iCloud IMAP (#493).
## 9.0.7
## Fixes
- Fix IMAP `since` option (#PR 645 closes issues #581 and #643).
## 9.0.6
### Fixes

View File

@@ -29,14 +29,3 @@ token_file = /etc/example/token.json
include_spam_trash = True
paginate_messages = True
scopes = https://www.googleapis.com/auth/gmail.modify
[msgraph]
auth_method = ClientSecret
client_id = 12345678-90ab-cdef-1234-567890abcdef
client_secret = your-client-secret-here
tenant_id = 12345678-90ab-cdef-1234-567890abcdef
mailbox = dmarc-reports@example.com
# Use standard folder names - they work across all locales
# and avoid "Default folder Root not found" errors
reports_folder = Inbox
archive_folder = Archive

View File

@@ -229,18 +229,6 @@ The full set of configuration options are:
username, you must grant the app `Mail.ReadWrite.Shared`.
:::
:::{tip}
When configuring folder names (e.g., `reports_folder`, `archive_folder`),
you can use standard folder names like `Inbox`, `Archive`, `Sent Items`, etc.
These will be automatically mapped to Microsoft Graph's well-known folder names,
which works reliably across different mailbox locales and avoids issues with
uninitialized or shared mailboxes. Supported folder names include:
- English: Inbox, Sent Items, Deleted Items, Drafts, Junk Email, Archive, Outbox
- German: Posteingang, Gesendete Elemente, Gelöschte Elemente, Entwürfe, Junk-E-Mail, Archiv
- French: Boîte de réception, Éléments envoyés, Éléments supprimés, Brouillons, Courrier indésirable, Archives
- Spanish: Bandeja de entrada, Elementos enviados, Elementos eliminados, Borradores, Correo no deseado
:::
:::{warning}
If you are using the `ClientSecret` auth method, you need to
grant the `Mail.ReadWrite` (application) permission to the

View File

@@ -595,12 +595,18 @@ def parsed_smtp_tls_reports_to_csv_rows(
if "mx_host_patterns" in policy:
record["mx_host_patterns"] = "|".join(policy["mx_host_patterns"])
successful_record = record.copy()
successful_record["policy_domain"] = policy["policy_domain"]
successful_record["policy_type"] = policy["policy_type"]
successful_record["policy_domain"] = policy[
"policy_domain"
]
successful_record["policy_type"] = policy[
"policy_type"
]
successful_record["successful_session_count"] = policy[
"successful_session_count"
]
successful_record["failed_session_count"] = policy["failed_session_count"]
successful_record["failed_session_count"] = policy[
"failed_session_count"
]
rows.append(successful_record)
if "failure_details" in policy:
for failure_details in policy["failure_details"]:
@@ -1942,9 +1948,7 @@ def get_dmarc_reports_from_mailbox(
"Only days and weeks values in 'since' option are \
considered for IMAP connections. Examples: 2d or 1w"
)
since = (datetime.now(timezone.utc) - timedelta(minutes=_since)).strftime(
"%d-%b-%Y"
)
since = (datetime.now(timezone.utc) - timedelta(minutes=_since)).strftime("%d-%b-%Y")
current_time = datetime.now(timezone.utc).strftime("%d-%b-%Y")
elif isinstance(connection, MSGraphConnection):
since = (

View File

@@ -46,7 +46,6 @@ from parsedmarc.mail import (
MSGraphConnection,
)
from parsedmarc.mail.graph import AuthMethod
from parsedmarc.types import ParsingResults
from parsedmarc.utils import get_base_domain, get_reverse_dns, is_mbox
# Increase the max header limit for very large emails. `_MAXHEADERS` is a
@@ -68,48 +67,6 @@ def _str_to_list(s):
return list(map(lambda i: i.lstrip(), _list))
def _configure_logging(log_level, log_file=None):
"""
Configure logging for the current process.
This is needed for child processes to properly log messages.
Args:
log_level: The logging level (e.g., logging.DEBUG, logging.WARNING)
log_file: Optional path to log file
"""
# Get the logger
from parsedmarc.log import logger
# Set the log level
logger.setLevel(log_level)
# Add StreamHandler with formatter if not already present
# Check if we already have a StreamHandler to avoid duplicates
# Use exact type check to distinguish from FileHandler subclass
has_stream_handler = any(type(h) is logging.StreamHandler for h in logger.handlers)
if not has_stream_handler:
formatter = logging.Formatter(
fmt="%(levelname)8s:%(filename)s:%(lineno)d:%(message)s",
datefmt="%Y-%m-%d:%H:%M:%S",
)
handler = logging.StreamHandler()
handler.setFormatter(formatter)
logger.addHandler(handler)
# Add FileHandler if log_file is specified
if log_file:
try:
fh = logging.FileHandler(log_file, "a")
formatter = logging.Formatter(
"%(asctime)s - %(levelname)s - [%(filename)s:%(lineno)d] - %(message)s"
)
fh.setFormatter(formatter)
logger.addHandler(fh)
except (IOError, OSError, PermissionError) as error:
logger.warning("Unable to write to log file: {}".format(error))
def cli_parse(
file_path,
sa,
@@ -122,29 +79,8 @@ def cli_parse(
reverse_dns_map_url,
normalize_timespan_threshold_hours,
conn,
log_level=logging.ERROR,
log_file=None,
):
"""Separated this function for multiprocessing
Args:
file_path: Path to the report file
sa: Strip attachment payloads flag
nameservers: List of nameservers
dns_timeout: DNS timeout
ip_db_path: Path to IP database
offline: Offline mode flag
always_use_local_files: Always use local files flag
reverse_dns_map_path: Path to reverse DNS map
reverse_dns_map_url: URL to reverse DNS map
normalize_timespan_threshold_hours: Timespan threshold
conn: Pipe connection for IPC
log_level: Logging level for this process
log_file: Optional path to log file
"""
# Configure logging in this child process
_configure_logging(log_level, log_file)
"""Separated this function for multiprocessing"""
try:
file_results = parse_report_file(
file_path,
@@ -1525,10 +1461,6 @@ def _main():
if n_procs < 1:
n_procs = 1
# Capture the current log level to pass to child processes
current_log_level = logger.level
current_log_file = opts.log_file
for batch_index in range((len(file_paths) + n_procs - 1) // n_procs):
processes = []
connections = []
@@ -1554,8 +1486,6 @@ def _main():
opts.reverse_dns_map_url,
opts.normalize_timespan_threshold_hours,
child_conn,
current_log_level,
current_log_file,
),
)
processes.append(process)
@@ -1758,13 +1688,13 @@ def _main():
logger.exception("Mailbox Error")
exit(1)
parsing_results: ParsingResults = {
results = {
"aggregate_reports": aggregate_reports,
"forensic_reports": forensic_reports,
"smtp_tls_reports": smtp_tls_reports,
}
process_reports(parsing_results)
process_reports(results)
if opts.smtp_host:
try:
@@ -1778,7 +1708,7 @@ def _main():
else _str_to_list(str(opts.smtp_to))
)
email_results(
parsing_results,
results,
opts.smtp_host,
opts.smtp_from,
smtp_to_value,

View File

@@ -1,3 +1,3 @@
__version__ = "9.0.8"
__version__ = "9.0.6"
USER_AGENT = f"parsedmarc/{__version__}"

View File

@@ -20,59 +20,6 @@ from msgraph.core import GraphClient
from parsedmarc.log import logger
from parsedmarc.mail.mailbox_connection import MailboxConnection
# Mapping of common folder names to Microsoft Graph well-known folder names
# This avoids the "Default folder Root not found" error on uninitialized mailboxes
WELL_KNOWN_FOLDER_MAP = {
# English names
"inbox": "inbox",
"sent items": "sentitems",
"sent": "sentitems",
"sentitems": "sentitems",
"deleted items": "deleteditems",
"deleted": "deleteditems",
"deleteditems": "deleteditems",
"trash": "deleteditems",
"drafts": "drafts",
"junk email": "junkemail",
"junk": "junkemail",
"junkemail": "junkemail",
"spam": "junkemail",
"archive": "archive",
"outbox": "outbox",
"conversation history": "conversationhistory",
"conversationhistory": "conversationhistory",
# German names
"posteingang": "inbox",
"gesendete elemente": "sentitems",
"gesendet": "sentitems",
"gelöschte elemente": "deleteditems",
"gelöscht": "deleteditems",
"entwürfe": "drafts",
"junk-e-mail": "junkemail",
"archiv": "archive",
"postausgang": "outbox",
# French names
"boîte de réception": "inbox",
"éléments envoyés": "sentitems",
"envoyés": "sentitems",
"éléments supprimés": "deleteditems",
"supprimés": "deleteditems",
"brouillons": "drafts",
"courrier indésirable": "junkemail",
"archives": "archive",
"boîte d'envoi": "outbox",
# Spanish names
"bandeja de entrada": "inbox",
"elementos enviados": "sentitems",
"enviados": "sentitems",
"elementos eliminados": "deleteditems",
"eliminados": "deleteditems",
"borradores": "drafts",
"correo no deseado": "junkemail",
"archivar": "archive",
"bandeja de salida": "outbox",
}
class AuthMethod(Enum):
DeviceCode = 1
@@ -183,13 +130,6 @@ class MSGraphConnection(MailboxConnection):
self.mailbox_name = mailbox
def create_folder(self, folder_name: str):
# Check if this is a well-known folder - they already exist and cannot be created
if "/" not in folder_name:
well_known_name = WELL_KNOWN_FOLDER_MAP.get(folder_name.lower())
if well_known_name:
logger.debug(f"Folder '{folder_name}' is a well-known folder, skipping creation")
return
sub_url = ""
path_parts = folder_name.split("/")
if len(path_parts) > 1: # Folder is a subFolder
@@ -306,12 +246,6 @@ class MSGraphConnection(MailboxConnection):
parent_folder_id = folder_id
return self._find_folder_id_with_parent(path_parts[-1], parent_folder_id)
else:
# Check if this is a well-known folder name (case-insensitive)
well_known_name = WELL_KNOWN_FOLDER_MAP.get(folder_name.lower())
if well_known_name:
# Use well-known folder name directly to avoid querying uninitialized mailboxes
logger.debug(f"Using well-known folder name '{well_known_name}' for '{folder_name}'")
return well_known_name
return self._find_folder_id_with_parent(folder_name, None)
def _find_folder_id_with_parent(

View File

@@ -48,7 +48,7 @@ dependencies = [
"imapclient>=2.1.0",
"kafka-python-ng>=2.2.2",
"lxml>=4.4.0",
"mailsuite>=1.11.1",
"mailsuite>=1.11.0",
"msgraph-core==0.2.2",
"opensearch-py>=2.4.2,<=3.0.0",
"publicsuffixlist>=0.10.0",

View File

@@ -156,32 +156,6 @@ class Test(unittest.TestCase):
parsedmarc.parsed_smtp_tls_reports_to_csv(parsed_report)
print("Passed!")
def testMSGraphWellKnownFolders(self):
"""Test MSGraph well-known folder name mapping"""
from parsedmarc.mail.graph import WELL_KNOWN_FOLDER_MAP
# Test English folder names
assert WELL_KNOWN_FOLDER_MAP.get("inbox") == "inbox"
assert WELL_KNOWN_FOLDER_MAP.get("sent items") == "sentitems"
assert WELL_KNOWN_FOLDER_MAP.get("deleted items") == "deleteditems"
assert WELL_KNOWN_FOLDER_MAP.get("archive") == "archive"
# Test case insensitivity - simulating how the code actually uses it
# This is what happens when user config has "reports_folder = Inbox"
assert WELL_KNOWN_FOLDER_MAP.get("inbox") == "inbox"
assert WELL_KNOWN_FOLDER_MAP.get("Inbox".lower()) == "inbox" # User's exact config
assert WELL_KNOWN_FOLDER_MAP.get("INBOX".lower()) == "inbox"
assert WELL_KNOWN_FOLDER_MAP.get("Archive".lower()) == "archive"
# Test German folder names
assert WELL_KNOWN_FOLDER_MAP.get("posteingang") == "inbox"
assert WELL_KNOWN_FOLDER_MAP.get("Posteingang".lower()) == "inbox" # Capitalized
assert WELL_KNOWN_FOLDER_MAP.get("archiv") == "archive"
# Test that custom folders don't match
assert WELL_KNOWN_FOLDER_MAP.get("custom_folder") is None
assert WELL_KNOWN_FOLDER_MAP.get("my_reports") is None
if __name__ == "__main__":
unittest.main(verbosity=2)