Make build.sh usable without the gh-pages branch

- Add missing `https://` to the default Microsoft Graph URL

CHANGELOG.md

@@ -1,10 +1,22 @@
 Changelog
 =========
 
+8.18.1
+------
+
+- Add missing `https://` to the default Microsoft Graph URL
+
+8.18.0
+------
+
+- Add support for Microsoft national clouds via Graph API base URL (PR #590)
+- Avoid stopping processing when an invalid DMARC report is encountered (PR #587)
+- Increase `http.client._MAXHEADERS` from `100` to `200` to avoid errors connecting to Elasticsearch/OpenSearch (PR #589)
+
 8.17.0
 ------
 
-- Ignore duplicate aggregate DMARC reports with the same `org_name` and `report_id` seen within the same hour (Fixes [#539](https://github.com/domainaware/parsedmarc/issues/539))
+- Ignore duplicate aggregate DMARC reports with the same `org_name` and `report_id` seen within the same hour (Fixes #535)
 - Fix saving SMTP TLS reports to OpenSearch (PR #585 closed issue #576)
 - Add 303 entries to `base_reverse_dns_map.csv`
 

build.sh

@@ -14,7 +14,9 @@ cd docs
 make clean 
 make html
 touch build/html/.nojekyll
-cp -rf build/html/* ../../parsedmarc-docs/
+if [  -d "./../parsedmarc-docs" ]; then
+  cp -rf build/html/* ../../parsedmarc-docs/
+fi
 cd ..
 ./sortmaps.py
 python3 tests.py

docs/source/usage.md

@@ -208,6 +208,8 @@ The full set of configuration options are:
   - `mailbox` - str: The mailbox name. This defaults to the
       current user if using the UsernamePassword auth method, but
       could be a shared mailbox if the user has access to the mailbox
+  - `graph_url` - str: Microsoft Graph URL.  Allows for use of National Clouds (ex Azure Gov)
+      (Default: https://graph.microsoft.com)
   - `token_file` - str: Path to save the token file
       (Default: `.token`)
   - `allow_unencrypted_storage` - bool: Allows the Azure Identity

parsedmarc/__init__.py

@@ -39,7 +39,7 @@ from parsedmarc.utils import is_outlook_msg, convert_outlook_msg
 from parsedmarc.utils import parse_email
 from parsedmarc.utils import timestamp_to_human, human_timestamp_to_datetime
 
-__version__ = "8.17.0"
+__version__ = "8.18.1"
 
 logger.debug("parsedmarc v{0}".format(__version__))
 
@@ -272,7 +272,7 @@ def _parse_smtp_tls_failure_details(failure_details):
         return new_failure_details
 
     except KeyError as e:
-        raise InvalidSMTPTLSReport(f"Missing required failure details field:" f" {e}")
+        raise InvalidSMTPTLSReport(f"Missing required failure details field: {e}")
     except Exception as e:
         raise InvalidSMTPTLSReport(str(e))
 
@@ -284,7 +284,7 @@ def _parse_smtp_tls_report_policy(policy):
         policy_type = policy["policy"]["policy-type"]
         failure_details = []
         if policy_type not in policy_types:
-            raise InvalidSMTPTLSReport(f"Invalid policy type " f"{policy_type}")
+            raise InvalidSMTPTLSReport(f"Invalid policy type {policy_type}")
         new_policy = OrderedDict(policy_domain=policy_domain, policy_type=policy_type)
         if "policy-string" in policy["policy"]:
             if isinstance(policy["policy"]["policy-string"], list):
@@ -332,9 +332,7 @@ def parse_smtp_tls_report_json(report):
                 raise Exception(f"Missing required field: {required_field}]")
         if not isinstance(report["policies"], list):
             policies_type = type(report["policies"])
-            raise InvalidSMTPTLSReport(
-                f"policies must be a list, " f"not {policies_type}"
-            )
+            raise InvalidSMTPTLSReport(f"policies must be a list, not {policies_type}")
         for policy in report["policies"]:
             policies.append(_parse_smtp_tls_report_policy(policy))
 
@@ -1246,11 +1244,11 @@ def parse_report_email(
                         field_name = match[0].lower().replace(" ", "-")
                         fields[field_name] = match[1].strip()
 
-                    feedback_report = "Arrival-Date: {}\n" "Source-IP: {}" "".format(
+                    feedback_report = "Arrival-Date: {}\nSource-IP: {}".format(
                         fields["received-date"], fields["sender-ip-address"]
                     )
                 except Exception as e:
-                    error = "Unable to parse message with " 'subject "{0}": {1}'.format(
+                    error = 'Unable to parse message with subject "{0}": {1}'.format(
                         subject, e
                     )
                     raise InvalidDMARCReport(error)
@@ -1294,10 +1292,10 @@ def parse_report_email(
                     "is not a valid "
                     "aggregate DMARC report: {1}".format(subject, e)
                 )
-                raise ParserError(error)
+                raise InvalidDMARCReport(error)
 
             except Exception as e:
-                error = "Unable to parse message with " 'subject "{0}": {1}'.format(
+                error = 'Unable to parse message with subject "{0}": {1}'.format(
                     subject, e
                 )
                 raise ParserError(error)
@@ -1331,7 +1329,7 @@ def parse_report_email(
         return result
 
     if result is None:
-        error = 'Message with subject "{0}" is ' "not a valid report".format(subject)
+        error = 'Message with subject "{0}" is not a valid report'.format(subject)
         raise InvalidDMARCReport(error)
 
 
@@ -1666,7 +1664,7 @@ def get_dmarc_reports_from_mailbox(
                     aggregate_reports.append(parsed_email["report"])
                 else:
                     logger.debug(
-                        "Skipping duplicate aggregate report " f"with ID: {report_id}"
+                        f"Skipping duplicate aggregate report with ID: {report_id}"
                     )
                 aggregate_report_msg_uids.append(msg_uid)
             elif parsed_email["report_type"] == "forensic":
@@ -1708,7 +1706,7 @@ def get_dmarc_reports_from_mailbox(
 
                 except Exception as e:
                     message = "Error deleting message UID"
-                    e = "{0} {1}: " "{2}".format(message, msg_uid, e)
+                    e = "{0} {1}: {2}".format(message, msg_uid, e)
                     logger.error("Mailbox error: {0}".format(e))
         else:
             if len(aggregate_report_msg_uids) > 0:

parsedmarc/cli.py

@@ -14,6 +14,7 @@ import json
 from ssl import CERT_NONE, create_default_context
 from multiprocessing import Pipe, Process
 import sys
+import http.client
 from tqdm import tqdm
 
 from parsedmarc import (
@@ -48,6 +49,8 @@ from parsedmarc.log import logger
 from parsedmarc.utils import is_mbox, get_reverse_dns
 from parsedmarc import SEEN_AGGREGATE_REPORT_IDS
 
+http.client._MAXHEADERS = 200  # pylint:disable=protected-access
+
 formatter = logging.Formatter(
     fmt="%(levelname)8s:%(filename)s:%(lineno)d:%(message)s",
     datefmt="%Y-%m-%d:%H:%M:%S",
@@ -396,7 +399,7 @@ def _main():
     arg_parser.add_argument(
         "-c",
         "--config-file",
-        help="a path to a configuration file " "(--silent implied)",
+        help="a path to a configuration file (--silent implied)",
     )
     arg_parser.add_argument(
         "file_path",
@@ -404,7 +407,7 @@ def _main():
         help="one or more paths to aggregate or forensic "
         "report files, emails, or mbox files'",
     )
-    strip_attachment_help = "remove attachment payloads from forensic " "report output"
+    strip_attachment_help = "remove attachment payloads from forensic report output"
     arg_parser.add_argument(
         "--strip-attachment-payloads", help=strip_attachment_help, action="store_true"
     )
@@ -447,14 +450,14 @@ def _main():
     arg_parser.add_argument(
         "-t",
         "--dns_timeout",
-        help="number of seconds to wait for an answer " "from DNS (default: 2.0)",
+        help="number of seconds to wait for an answer from DNS (default: 2.0)",
         type=float,
         default=2.0,
     )
     arg_parser.add_argument(
         "--offline",
         action="store_true",
-        help="do not make online queries for geolocation " " or  DNS",
+        help="do not make online queries for geolocation  or  DNS",
     )
     arg_parser.add_argument(
         "-s", "--silent", action="store_true", help="only print errors"
@@ -528,6 +531,7 @@ def _main():
         graph_tenant_id=None,
         graph_mailbox=None,
         graph_allow_unencrypted_storage=False,
+        graph_url="https://graph.microsoft.com",
         hec=None,
         hec_token=None,
         hec_index=None,
@@ -730,7 +734,7 @@ def _main():
             if "host" in imap_config:
                 opts.imap_host = imap_config["host"]
             else:
-                logger.error("host setting missing from the " "imap config section")
+                logger.error("host setting missing from the imap config section")
                 exit(-1)
             if "port" in imap_config:
                 opts.imap_port = imap_config.getint("port")
@@ -746,14 +750,12 @@ def _main():
             if "user" in imap_config:
                 opts.imap_user = imap_config["user"]
             else:
-                logger.critical("user setting missing from the " "imap config section")
+                logger.critical("user setting missing from the imap config section")
                 exit(-1)
             if "password" in imap_config:
                 opts.imap_password = imap_config["password"]
             else:
-                logger.critical(
-                    "password setting missing from the " "imap config section"
-                )
+                logger.critical("password setting missing from the imap config section")
                 exit(-1)
             if "reports_folder" in imap_config:
                 opts.mailbox_reports_folder = imap_config["reports_folder"]
@@ -822,21 +824,20 @@ def _main():
                     opts.graph_user = graph_config["user"]
                 else:
                     logger.critical(
-                        "user setting missing from the " "msgraph config section"
+                        "user setting missing from the msgraph config section"
                     )
                     exit(-1)
                 if "password" in graph_config:
                     opts.graph_password = graph_config["password"]
                 else:
                     logger.critical(
-                        "password setting missing from the " "msgraph config section"
+                        "password setting missing from the msgraph config section"
                     )
                 if "client_secret" in graph_config:
                     opts.graph_client_secret = graph_config["client_secret"]
                 else:
                     logger.critical(
-                        "client_secret setting missing from the "
-                        "msgraph config section"
+                        "client_secret setting missing from the msgraph config section"
                     )
                     exit(-1)
 
@@ -849,7 +850,7 @@ def _main():
                     opts.graph_tenant_id = graph_config["tenant_id"]
                 else:
                     logger.critical(
-                        "tenant_id setting missing from the " "msgraph config section"
+                        "tenant_id setting missing from the msgraph config section"
                     )
                     exit(-1)
 
@@ -858,8 +859,7 @@ def _main():
                     opts.graph_client_secret = graph_config["client_secret"]
                 else:
                     logger.critical(
-                        "client_secret setting missing from the "
-                        "msgraph config section"
+                        "client_secret setting missing from the msgraph config section"
                     )
                     exit(-1)
 
@@ -867,7 +867,7 @@ def _main():
                 opts.graph_client_id = graph_config["client_id"]
             else:
                 logger.critical(
-                    "client_id setting missing from the " "msgraph config section"
+                    "client_id setting missing from the msgraph config section"
                 )
                 exit(-1)
 
@@ -875,10 +875,13 @@ def _main():
                 opts.graph_mailbox = graph_config["mailbox"]
             elif opts.graph_auth_method != AuthMethod.UsernamePassword.name:
                 logger.critical(
-                    "mailbox setting missing from the " "msgraph config section"
+                    "mailbox setting missing from the msgraph config section"
                 )
                 exit(-1)
 
+            if "graph_url" in graph_config:
+                opts.graph_url = graph_config["graph_url"]
+
             if "allow_unencrypted_storage" in graph_config:
                 opts.graph_allow_unencrypted_storage = graph_config.getboolean(
                     "allow_unencrypted_storage"
@@ -890,7 +893,7 @@ def _main():
                 opts.elasticsearch_hosts = _str_to_list(elasticsearch_config["hosts"])
             else:
                 logger.critical(
-                    "hosts setting missing from the " "elasticsearch config section"
+                    "hosts setting missing from the elasticsearch config section"
                 )
                 exit(-1)
             if "timeout" in elasticsearch_config:
@@ -928,7 +931,7 @@ def _main():
                 opts.opensearch_hosts = _str_to_list(opensearch_config["hosts"])
             else:
                 logger.critical(
-                    "hosts setting missing from the " "opensearch config section"
+                    "hosts setting missing from the opensearch config section"
                 )
                 exit(-1)
             if "timeout" in opensearch_config:
@@ -964,21 +967,21 @@ def _main():
                 opts.hec = hec_config["url"]
             else:
                 logger.critical(
-                    "url setting missing from the " "splunk_hec config section"
+                    "url setting missing from the splunk_hec config section"
                 )
                 exit(-1)
             if "token" in hec_config:
                 opts.hec_token = hec_config["token"]
             else:
                 logger.critical(
-                    "token setting missing from the " "splunk_hec config section"
+                    "token setting missing from the splunk_hec config section"
                 )
                 exit(-1)
             if "index" in hec_config:
                 opts.hec_index = hec_config["index"]
             else:
                 logger.critical(
-                    "index setting missing from the " "splunk_hec config section"
+                    "index setting missing from the splunk_hec config section"
                 )
                 exit(-1)
             if "skip_certificate_verification" in hec_config:
@@ -991,9 +994,7 @@ def _main():
             if "hosts" in kafka_config:
                 opts.kafka_hosts = _str_to_list(kafka_config["hosts"])
             else:
-                logger.critical(
-                    "hosts setting missing from the " "kafka config section"
-                )
+                logger.critical("hosts setting missing from the kafka config section")
                 exit(-1)
             if "user" in kafka_config:
                 opts.kafka_username = kafka_config["user"]
@@ -1008,21 +1009,20 @@ def _main():
                 opts.kafka_aggregate_topic = kafka_config["aggregate_topic"]
             else:
                 logger.critical(
-                    "aggregate_topic setting missing from the " "kafka config section"
+                    "aggregate_topic setting missing from the kafka config section"
                 )
                 exit(-1)
             if "forensic_topic" in kafka_config:
                 opts.kafka_forensic_topic = kafka_config["forensic_topic"]
             else:
                 logger.critical(
-                    "forensic_topic setting missing from the " "kafka config section"
+                    "forensic_topic setting missing from the kafka config section"
                 )
             if "smtp_tls_topic" in kafka_config:
                 opts.kafka_smtp_tls_topic = kafka_config["smtp_tls_topic"]
             else:
                 logger.critical(
-                    "forensic_topic setting missing from the "
-                    "splunk_hec config section"
+                    "forensic_topic setting missing from the splunk_hec config section"
                 )
 
         if "smtp" in config.sections():
@@ -1030,7 +1030,7 @@ def _main():
             if "host" in smtp_config:
                 opts.smtp_host = smtp_config["host"]
             else:
-                logger.critical("host setting missing from the " "smtp config section")
+                logger.critical("host setting missing from the smtp config section")
                 exit(-1)
             if "port" in smtp_config:
                 opts.smtp_port = smtp_config.getint("port")
@@ -1042,23 +1042,21 @@ def _main():
             if "user" in smtp_config:
                 opts.smtp_user = smtp_config["user"]
             else:
-                logger.critical("user setting missing from the " "smtp config section")
+                logger.critical("user setting missing from the smtp config section")
                 exit(-1)
             if "password" in smtp_config:
                 opts.smtp_password = smtp_config["password"]
             else:
-                logger.critical(
-                    "password setting missing from the " "smtp config section"
-                )
+                logger.critical("password setting missing from the smtp config section")
                 exit(-1)
             if "from" in smtp_config:
                 opts.smtp_from = smtp_config["from"]
             else:
-                logger.critical("from setting missing from the " "smtp config section")
+                logger.critical("from setting missing from the smtp config section")
             if "to" in smtp_config:
                 opts.smtp_to = _str_to_list(smtp_config["to"])
             else:
-                logger.critical("to setting missing from the " "smtp config section")
+                logger.critical("to setting missing from the smtp config section")
             if "subject" in smtp_config:
                 opts.smtp_subject = smtp_config["subject"]
             if "attachment" in smtp_config:
@@ -1071,7 +1069,7 @@ def _main():
             if "bucket" in s3_config:
                 opts.s3_bucket = s3_config["bucket"]
             else:
-                logger.critical("bucket setting missing from the " "s3 config section")
+                logger.critical("bucket setting missing from the s3 config section")
                 exit(-1)
             if "path" in s3_config:
                 opts.s3_path = s3_config["path"]
@@ -1096,9 +1094,7 @@ def _main():
             if "server" in syslog_config:
                 opts.syslog_server = syslog_config["server"]
             else:
-                logger.critical(
-                    "server setting missing from the " "syslog config section"
-                )
+                logger.critical("server setting missing from the syslog config section")
                 exit(-1)
             if "port" in syslog_config:
                 opts.syslog_port = syslog_config["port"]
@@ -1149,17 +1145,17 @@ def _main():
             if "host" in gelf_config:
                 opts.gelf_host = gelf_config["host"]
             else:
-                logger.critical("host setting missing from the " "gelf config section")
+                logger.critical("host setting missing from the gelf config section")
                 exit(-1)
             if "port" in gelf_config:
                 opts.gelf_port = gelf_config["port"]
             else:
-                logger.critical("port setting missing from the " "gelf config section")
+                logger.critical("port setting missing from the gelf config section")
                 exit(-1)
             if "mode" in gelf_config:
                 opts.gelf_mode = gelf_config["mode"]
             else:
-                logger.critical("mode setting missing from the " "gelf config section")
+                logger.critical("mode setting missing from the gelf config section")
                 exit(-1)
 
         if "webhook" in config.sections():
@@ -1185,8 +1181,7 @@ def _main():
         try:
             fh = logging.FileHandler(opts.log_file, "a")
             formatter = logging.Formatter(
-                "%(asctime)s - "
-                "%(levelname)s - [%(filename)s:%(lineno)d] - %(message)s"
+                "%(asctime)s - %(levelname)s - [%(filename)s:%(lineno)d] - %(message)s"
             )
             fh.setFormatter(formatter)
             logger.addHandler(fh)
@@ -1294,7 +1289,7 @@ def _main():
 
     if opts.hec:
         if opts.hec_token is None or opts.hec_index is None:
-            logger.error("HEC token and HEC index are required when " "using HEC URL")
+            logger.error("HEC token and HEC index are required when using HEC URL")
             exit(1)
 
         verify = True
@@ -1457,7 +1452,7 @@ def _main():
         try:
             if opts.imap_user is None or opts.imap_password is None:
                 logger.error(
-                    "IMAP user and password must be specified if" "host is specified"
+                    "IMAP user and password must be specified ifhost is specified"
                 )
 
             ssl = True
@@ -1496,6 +1491,7 @@ def _main():
                 password=opts.graph_password,
                 token_file=opts.graph_token_file,
                 allow_unencrypted_storage=opts.graph_allow_unencrypted_storage,
+                graph_url=opts.graph_url,
             )
 
         except Exception:

parsedmarc/mail/gmail.py

@@ -63,9 +63,7 @@ class GmailConnection(MailboxConnection):
             ).execute()
         except HttpError as e:
             if e.status_code == 409:
-                logger.debug(
-                    f"Folder {folder_name} already exists, " f"skipping creation"
-                )
+                logger.debug(f"Folder {folder_name} already exists, skipping creation")
             else:
                 raise e
 

parsedmarc/mail/graph.py

@@ -89,6 +89,7 @@ class MSGraphConnection(MailboxConnection):
         self,
         auth_method: str,
         mailbox: str,
+        graph_url: str,
         client_id: str,
         client_secret: str,
         username: str,
@@ -108,7 +109,10 @@ class MSGraphConnection(MailboxConnection):
             token_path=token_path,
             allow_unencrypted_storage=allow_unencrypted_storage,
         )
-        client_params = {"credential": credential}
+        client_params = {
+            "credential": credential,
+            "cloud": graph_url,
+        }
         if not isinstance(credential, ClientSecretCredential):
             scopes = ["Mail.ReadWrite"]
             # Detect if mailbox is shared
@@ -137,16 +141,16 @@ class MSGraphConnection(MailboxConnection):
         request_url = f"/users/{self.mailbox_name}/mailFolders{sub_url}"
         resp = self._client.post(request_url, json=request_body)
         if resp.status_code == 409:
-            logger.debug(f"Folder {folder_name} already exists, " f"skipping creation")
+            logger.debug(f"Folder {folder_name} already exists, skipping creation")
         elif resp.status_code == 201:
             logger.debug(f"Created folder {folder_name}")
         else:
-            logger.warning(f"Unknown response " f"{resp.status_code} {resp.json()}")
+            logger.warning(f"Unknown response {resp.status_code} {resp.json()}")
 
     def fetch_messages(self, folder_name: str, **kwargs) -> List[str]:
         """Returns a list of message UIDs in the specified folder"""
         folder_id = self._find_folder_id_from_folder_path(folder_name)
-        url = f"/users/{self.mailbox_name}/mailFolders/" f"{folder_id}/messages"
+        url = f"/users/{self.mailbox_name}/mailFolders/{folder_id}/messages"
         since = kwargs.get("since")
         if not since:
             since = None
@@ -185,7 +189,7 @@ class MSGraphConnection(MailboxConnection):
         resp = self._client.patch(url, json={"isRead": "true"})
         if resp.status_code != 200:
             raise RuntimeWarning(
-                f"Failed to mark message read" f"{resp.status_code}: {resp.json()}"
+                f"Failed to mark message read{resp.status_code}: {resp.json()}"
             )
 
     def fetch_message(self, message_id: str, **kwargs):
@@ -193,7 +197,7 @@ class MSGraphConnection(MailboxConnection):
         result = self._client.get(url)
         if result.status_code != 200:
             raise RuntimeWarning(
-                f"Failed to fetch message" f"{result.status_code}: {result.json()}"
+                f"Failed to fetch message{result.status_code}: {result.json()}"
             )
         mark_read = kwargs.get("mark_read")
         if mark_read:
@@ -205,7 +209,7 @@ class MSGraphConnection(MailboxConnection):
         resp = self._client.delete(url)
         if resp.status_code != 204:
             raise RuntimeWarning(
-                f"Failed to delete message " f"{resp.status_code}: {resp.json()}"
+                f"Failed to delete message {resp.status_code}: {resp.json()}"
             )
 
     def move_message(self, message_id: str, folder_name: str):
@@ -215,7 +219,7 @@ class MSGraphConnection(MailboxConnection):
         resp = self._client.post(url, json=request_body)
         if resp.status_code != 201:
             raise RuntimeWarning(
-                f"Failed to move message " f"{resp.status_code}: {resp.json()}"
+                f"Failed to move message {resp.status_code}: {resp.json()}"
             )
 
     def keepalive(self):
@@ -250,7 +254,7 @@ class MSGraphConnection(MailboxConnection):
         filter = f"?$filter=displayName eq '{folder_name}'"
         folders_resp = self._client.get(url + filter)
         if folders_resp.status_code != 200:
-            raise RuntimeWarning(f"Failed to list folders." f"{folders_resp.json()}")
+            raise RuntimeWarning(f"Failed to list folders.{folders_resp.json()}")
         folders: list = folders_resp.json()["value"]
         matched_folders = [
             folder for folder in folders if folder["displayName"] == folder_name

parsedmarc/mail/imap.py

@@ -85,7 +85,5 @@ class IMAPConnection(MailboxConnection):
                 logger.warning("IMAP connection timeout. Reconnecting...")
                 sleep(check_timeout)
             except Exception as e:
-                logger.warning(
-                    "IMAP connection error. {0}. " "Reconnecting...".format(e)
-                )
+                logger.warning("IMAP connection error. {0}. Reconnecting...".format(e))
                 sleep(check_timeout)

parsedmarc/utils.py

@@ -344,7 +344,7 @@ def get_service_from_reverse_dns_base_domain(
 
     if not (offline or always_use_local_file) and len(reverse_dns_map) == 0:
         try:
-            logger.debug(f"Trying to fetch " f"reverse DNS map from {url}...")
+            logger.debug(f"Trying to fetch reverse DNS map from {url}...")
             csv_file.write(requests.get(url).text)
             csv_file.seek(0)
             load_csv(csv_file)