extract-msg 0.48.2__py3-none-any.whl

extract_msg/constants/__init__.py

@@ -0,0 +1,294 @@
+"""
+The constants used in extract_msg. If you modify any of these without explicit
+instruction to do so from one of the contributers, please do not complain about
+bugs.
+"""
+
+__all__ = [
+    # Modules.
+    'ps',
+    're',
+    'st',
+
+    # Constants.
+    'DATE_FORMAT',
+    'DEFAULT_CLSID',
+    'DT_FORMAT',
+    'FIXED_LENGTH_PROPS',
+    'FIXED_LENGTH_PROPS_STRING',
+    'HEADER_FORMAT',
+    'HEADER_FORMAT_TYPE',
+    'HEADER_FORMAT_VALUE_TYPE',
+    'KNOWN_CLASS_TYPES',
+    'KNOWN_FILE_FLAGS',
+    'MAINDOC',
+    'MSG_PATH',
+    'MULTIPLE_16_BYTES',
+    'MULTIPLE_16_BYTES_HEX',
+    'MULTIPLE_2_BYTES',
+    'MULTIPLE_2_BYTES_HEX',
+    'MULTIPLE_4_BYTES',
+    'MULTIPLE_4_BYTES_HEX',
+    'MULTIPLE_8_BYTES',
+    'MULTIPLE_8_BYTES_HEX',
+    'NEEDS_ARG',
+    'NULL_DATE',
+    'PTYPES',
+    'PYTPFLOATINGTIME_START',
+    'REFUSED_CLASS_TYPES',
+    'REPOSITORY_URL',
+    'SAVE_TYPE',
+    'VARIABLE_LENGTH_PROPS',
+    'VARIABLE_LENGTH_PROPS_STRING',
+]
+
+
+from datetime import datetime
+from typing import (
+        Any, Callable, Dict, Final, List, Optional, Tuple, Type, TypeVar, Union
+    )
+
+from . import ps, re, st
+from ..enums import SaveType
+from ..null_date import NullDate
+
+
+_T = TypeVar('_T')
+
+
+# Constants for formating datetime objects.
+DATE_FORMAT: Final[str] = '%d %B, %Y'
+DT_FORMAT: Final[str] = '%a, %d %b %Y %H:%M:%S %z'
+
+
+# Typing Constants.
+HEADER_FORMAT_VALUE_TYPE = Union[str, Tuple[Union[str, None], bool], None]
+# Basically a dict of HEADER_FORMAT_TYPE and dicts containing them.
+HEADER_FORMAT_TYPE = Optional[Dict[str, Union[HEADER_FORMAT_VALUE_TYPE, Dict[str, HEADER_FORMAT_VALUE_TYPE]]]]
+SAVE_TYPE = Tuple[SaveType, Union[List[str], str, None]]
+MSG_PATH = Union[str, List[str], Tuple[str]]
+# Type used for the getXAs methods' overrideClass argument.
+OVERRIDE_CLASS = Union[Type[_T], Callable[[Any], _T]]
+
+
+FIXED_LENGTH_PROPS: Final[Tuple[int, ...]] = (
+    0x0000,
+    0x0001,
+    0x0002,
+    0x0003,
+    0x0004,
+    0x0005,
+    0x0006,
+    0x0007,
+    0x000A,
+    0x000B,
+    0x0014,
+    0x0040,
+)
+
+FIXED_LENGTH_PROPS_STRING: Final[Tuple[str, ...]] = (
+    '0000',
+    '0001',
+    '0002',
+    '0003',
+    '0004',
+    '0005',
+    '0006',
+    '0007',
+    '000A',
+    '000B',
+    '0014',
+    '0040',
+)
+
+VARIABLE_LENGTH_PROPS: Final[Tuple[int, ...]] = (
+    0x000D,
+    0x001E,
+    0x001F,
+    0x0048,
+    0x00FB,
+    0x00FD,
+    0x00FE,
+    0X0102,
+    0x1002,
+    0x1003,
+    0x1004,
+    0x1005,
+    0x1006,
+    0x1007,
+    0x1014,
+    0x101E,
+    0x101F,
+    0x1040,
+    0x1048,
+    0x1102,
+)
+
+VARIABLE_LENGTH_PROPS_STRING: Final[Tuple[str, ...]] = (
+    '000D',
+    '001E',
+    '001F',
+    '0048',
+    '00FB',
+    '00FD',
+    '00FE',
+    '0102',
+    '1002',
+    '1003',
+    '1004',
+    '1005',
+    '1006',
+    '1007',
+    '1014',
+    '101E',
+    '101F',
+    '1040',
+    '1048',
+    '1102',
+)
+
+# Multiple type properties that take up 2 bytes.
+MULTIPLE_2_BYTES: Final[Tuple[str, ...]] = (
+    '1002',
+)
+
+MULTIPLE_2_BYTES_HEX: Final[Tuple[int, ...]] = (
+    0x1002,
+)
+
+# Multiple type properties that take up 4 bytes.
+MULTIPLE_4_BYTES: Final[Tuple[str, ...]] = (
+    '1003',
+    '1004',
+)
+
+MULTIPLE_4_BYTES_HEX: Final[Tuple[int, ...]] = (
+    0x1003,
+    0x1004,
+)
+
+# Multiple type properties that take up 8 bytes.
+MULTIPLE_8_BYTES: Final[Tuple[str, ...]] = (
+    '1005',
+    '1006',
+    '1007',
+    '1014',
+    '1040',
+)
+
+MULTIPLE_8_BYTES_HEX: Final[Tuple[int, ...]] = (
+    0x1005,
+    0x1006,
+    0x1007,
+    0x1014,
+    0x1040,
+)
+
+# Multiple type properties that take up 16 bytes.
+MULTIPLE_16_BYTES: Final[Tuple[str, ...]] = (
+    '1048',
+)
+
+MULTIPLE_16_BYTES_HEX: Final[Tuple[int, ...]] = (
+    0x1048,
+)
+
+
+# Used to format the header for saving only the header.
+HEADER_FORMAT: Final[str] = """From: {From}
+To: {To}
+Cc: {Cc}
+Bcc: {Bcc}
+Subject: {subject}
+Date: {Date}
+Message-ID: {Message-Id}
+"""
+
+
+KNOWN_CLASS_TYPES: Final[Tuple[str, ...]] = (
+    'ipm.activity',
+    'ipm.appointment', # [MS-OXOCAL]
+    'ipm.contact', # [MS-OXOCNTC]
+    'ipm.configuration', # [MS-OXOCFG]
+    'ipm.distlist',
+    'ipm.document',
+    'ipm.ole.class',
+    'ipm.outlook.recall',
+    'ipm.note',
+    'ipm.post',
+    'ipm.stickynote',
+    'ipm.recall.report',
+    'ipm.remote',
+    'ipm.report',
+    'ipm.resend',
+    'ipm.schedule',
+    'ipm.task',
+    'ipm.taskrequest',
+    'report',
+)
+
+# Each item is a tuple of the lowercase class type and the issue number
+# associated with it.
+REFUSED_CLASS_TYPES: Final[Tuple[Tuple[str, str], ...]] = (
+    ('ipm.outlook.recall', '235'),
+)
+
+PYTPFLOATINGTIME_START: Final[datetime] = datetime(1899, 12, 30)
+# Do modifications before assigning to the constant.
+_tmp = NullDate(4500, 8, 31, 23, 59)
+_tmp.filetime = 915046235400000000
+NULL_DATE: Final[NullDate] = _tmp
+
+# Constants used for argparse stuff.
+KNOWN_FILE_FLAGS: Final[Tuple[str, ...]] = (
+    '--out-name',
+)
+NEEDS_ARG: Final[Tuple[str, ...]]= (
+    '--out-name',
+)
+REPOSITORY_URL: Final[str] = 'https://github.com/TeamMsgExtractor/msg-extractor'
+MAINDOC: Final[str] = f"""extract_msg:
+\tExtracts emails and attachments saved in Microsoft Outlook's .msg files.
+
+{REPOSITORY_URL}"""
+
+# Default class ID for the root entry for OleWriter. This should be
+# referencing Outlook if I understand it correctly.
+DEFAULT_CLSID: Final[bytes] = b'\x0b\r\x02\x00\x00\x00\x00\x00\xc0\x00\x00\x00\x00\x00\x00F'
+
+PTYPES: Final[Dict[int, str]]= {
+    0x0000: 'PtypUnspecified',
+    0x0001: 'PtypNull',
+    0x0002: 'PtypInteger16',  # Signed short.
+    0x0003: 'PtypInteger32',  # Signed int.
+    0x0004: 'PtypFloating32',  # Float.
+    0x0005: 'PtypFloating64',  # Double.
+    0x0006: 'PtypCurrency',
+    0x0007: 'PtypFloatingTime',
+    0x000A: 'PtypErrorCode',
+    0x000B: 'PtypBoolean',
+    0x000D: 'PtypObject/PtypEmbeddedTable/Storage',
+    0x0014: 'PtypInteger64',  # Signed longlong.
+    0x001E: 'PtypString8',
+    0x001F: 'PtypString',
+    0x0040: 'PtypTime', # Use filetimeToUtc to convert to unix time stamp.
+    0x0048: 'PtypGuid',
+    0x00FB: 'PtypServerId',
+    0x00FD: 'PtypRestriction',
+    0x00FE: 'PtypRuleAction',
+    0x0102: 'PtypBinary',
+    0x1002: 'PtypMultipleInteger16',
+    0x1003: 'PtypMultipleInteger32',
+    0x1004: 'PtypMultipleFloating32',
+    0x1005: 'PtypMultipleFloating64',
+    0x1006: 'PtypMultipleCurrency',
+    0x1007: 'PtypMultipleFloatingTime',
+    0x1014: 'PtypMultipleInteger64',
+    0x101E: 'PtypMultipleString8',
+    0x101F: 'PtypMultipleString',
+    0x1040: 'PtypMultipleTime',
+    0x1048: 'PtypMultipleGuid',
+    0x1102: 'PtypMultipleBinary',
+}
+
+# END CONSTANTS.

extract_msg/constants/ps.py

@@ -0,0 +1,46 @@
+"""
+Property set identifier constants.
+"""
+
+__all__ = [
+    'PSETID_ADDRESS',
+    'PSETID_AIRSYNC',
+    'PSETID_APPOINTMENT',
+    'PSETID_ATTACHMENT',
+    'PSETID_CALENDAR_ASSISTANT',
+    'PSETID_COMMON',
+    'PSETID_LOG',
+    'PSETID_MEETING',
+    'PSETID_MESSAGING',
+    'PSETID_NOTE',
+    'PSETID_POSTRSS',
+    'PSETID_SHARING',
+    'PSETID_TASK',
+    'PSETID_UNIFIEDMESSAGING',
+    'PSETID_XMLEXTRACTEDENTITIES',
+    'PS_INTERNET_HEADERS',
+    'PS_MAPI',
+    'PS_PUBLIC_STRINGS',
+]
+
+from typing import Final
+
+
+PS_MAPI: Final[str] = '{00020328-0000-0000-C000-000000000046}'
+PS_PUBLIC_STRINGS: Final[str] = '{00020329-0000-0000-C000-000000000046}'
+PSETID_COMMON: Final[str] = '{00062008-0000-0000-C000-000000000046}'
+PSETID_ADDRESS: Final[str] = '{00062004-0000-0000-C000-000000000046}'
+PS_INTERNET_HEADERS: Final[str] = '{00020386-0000-0000-C000-000000000046}'
+PSETID_APPOINTMENT: Final[str] = '{00062002-0000-0000-C000-000000000046}'
+PSETID_MEETING: Final[str] = '{6ED8DA90-450B-101B-98DA-00AA003F1305}'
+PSETID_LOG: Final[str] = '{0006200A-0000-0000-C000-000000000046}'
+PSETID_MESSAGING: Final[str] = '{41F28F13-83F4-4114-A584-EEDB5A6B0BFF}'
+PSETID_NOTE: Final[str] = '{0006200E-0000-0000-C000-000000000046}'
+PSETID_POSTRSS: Final[str] = '{00062041-0000-0000-C000-000000000046}'
+PSETID_TASK: Final[str] = '{00062003-0000-0000-C000-000000000046}'
+PSETID_UNIFIEDMESSAGING: Final[str] = '{4442858E-A9E3-4E80-B900-317A210CC15B}'
+PSETID_AIRSYNC: Final[str] = '{71035549-0739-4DCB-9163-00F0580DBBDF}'
+PSETID_SHARING: Final[str] = '{00062040-0000-0000-C000-000000000046}'
+PSETID_XMLEXTRACTEDENTITIES: Final[str] = '{23239608-685D-4732-9C55-4C95CB4E8E33}'
+PSETID_ATTACHMENT: Final[str] = '{96357F7F-59E1-47D0-99A7-46515C183B54}'
+PSETID_CALENDAR_ASSISTANT: Final[str] = '{11000E07-B51B-40D6-AF21-CAA85EDAB1D0}'

extract_msg/constants/re.py

@@ -0,0 +1,42 @@
+"""
+Regular expression constants.
+"""
+
+
+__all__ = [
+    'HTML_BODY_START',
+    'HTML_SAN_SPACE',
+    'INVALID_FILENAME_CHARS',
+    'INVALID_OLE_PATH',
+    'RTF_ENC_BODY_START',
+]
+
+
+import re
+
+from typing import Final
+
+
+# Allow better typing in versions above 3.8.
+import sys
+if sys.version_info >= (3, 9):
+    _RE_STR_TYPE = re.Pattern[str]
+    _RE_BYTES_TYPE = re.Pattern[bytes]
+else:
+    _RE_STR_TYPE = re.Pattern
+    _RE_BYTES_TYPE = re.Pattern
+
+
+# Characters that are invalid in a filename.
+INVALID_FILENAME_CHARS: Final[_RE_STR_TYPE] = re.compile(r'[\\/:*?"<>|]')
+# Regular expression to find sections of spaces for htmlSanitize.
+HTML_SAN_SPACE: Final[_RE_STR_TYPE] = re.compile('  +')
+# Regular expression to find the start of the html body.
+HTML_BODY_START: Final[_RE_BYTES_TYPE] = re.compile(b'<body[^>]*>')
+# Regular expression to find the start of the html body in encapsulated RTF.
+# This is used for one of the pattern types that makes life easy.
+RTF_ENC_BODY_START: Final[_RE_BYTES_TYPE] = re.compile(br'\{\\\*\\htmltag[0-9]* ?<body[^>]*>\}')
+# Used in the vaildation of OLE paths. Any of these characters in a name make it
+# invalid.
+INVALID_OLE_PATH: Final[_RE_STR_TYPE] = re.compile(r'[:/\\!]')
+

extract_msg/constants/st.py

@@ -0,0 +1,110 @@
+"""
+Struct constants.
+"""
+
+__all__ = [
+    'ST_BC_FIELD_INFO',
+    'ST_BC_HEAD',
+    'ST_BE_F32',
+    'ST_BE_F64',
+    'ST_BE_I16',
+    'ST_BE_I32',
+    'ST_BE_I64',
+    'ST_BE_I8',
+    'ST_BE_UI16',
+    'ST_BE_UI32',
+    'ST_BE_UI64',
+    'ST_BE_UI8',
+    'ST_CF_DIR_ENTRY',
+    'ST_GUID',
+    'ST_LE_F32',
+    'ST_LE_F64',
+    'ST_LE_I16',
+    'ST_LE_I32',
+    'ST_LE_I64',
+    'ST_LE_I8',
+    'ST_LE_UI16',
+    'ST_LE_UI32',
+    'ST_LE_UI64',
+    'ST_LE_UI8',
+    'ST_NP_ENT',
+    'ST_PEID',
+    'ST_PROP_BASE',
+    'ST_PROP_VAR',
+    'ST_PROPSTORE_HEADER',
+    'ST_RGB',
+    'ST_SBO_I8',
+    'ST_SBO_I16',
+    'ST_SBO_I32',
+    'ST_SBO_I64',
+    'ST_SBO_UI8',
+    'ST_SBO_UI16',
+    'ST_SBO_UI32',
+    'ST_SBO_UI64',
+    'ST_SYSTEMTIME',
+    'ST_TZ',
+]
+
+
+import struct
+
+from typing import Final
+
+
+# Define pre-compiled structs to make unpacking slightly faster.
+# General structs.
+ST_PROPSTORE_HEADER: Final[struct.Struct] = struct.Struct('<8x4I')
+ST_PROP_BASE: Final[struct.Struct] = struct.Struct('<2HI')
+# Struct used for unpacking a system time.
+ST_SYSTEMTIME: Final[struct.Struct] = struct.Struct('<8H')
+# Struct used for unpacking a GUID from bytes.
+ST_GUID: Final[struct.Struct] = struct.Struct('<IHH8s')
+# Struct for unpacking a TimeZoneStruct from bytes.
+ST_TZ: Final[struct.Struct] = struct.Struct('<iiiH16sH16s')
+# Struct for packing a compount file directory entry.
+ST_CF_DIR_ENTRY: Final[struct.Struct] = struct.Struct('<64sHBBIII16sIQQIQ')
+# Struct used for unpacking the entries in the entry stream
+ST_NP_ENT: Final[struct.Struct] = struct.Struct('<IHH')
+# Structs used by prop.py
+ST_PROP_VAR: Final[struct.Struct] = struct.Struct('<2I')
+# PermanentEntryID parsing struct
+ST_PEID: Final[struct.Struct] = struct.Struct('<B3x16s4xI')
+# Struct for unpacking the first part of the BusinessCardDisplayDefinition
+# structure.
+ST_BC_HEAD: Final[struct.Struct] = struct.Struct('BBBBBBBBBBBxB')
+# Struct for completely unpacking the FieldInfo structure.
+ST_BC_FIELD_INFO: Final[struct.Struct] = struct.Struct('HBBBxHBBBxBBBx')
+# Structs for parsing basic types.
+ST_LE_I8: Final[struct.Struct] = struct.Struct('<b')
+ST_LE_I16: Final[struct.Struct] = struct.Struct('<h')
+ST_LE_I32: Final[struct.Struct] = struct.Struct('<i')
+ST_LE_I64: Final[struct.Struct] = struct.Struct('<q')
+ST_LE_UI8: Final[struct.Struct] = struct.Struct('<B')
+ST_LE_UI16: Final[struct.Struct] = struct.Struct('<H')
+ST_LE_UI32: Final[struct.Struct] = struct.Struct('<I')
+ST_LE_UI64: Final[struct.Struct] = struct.Struct('<Q')
+ST_LE_F32: Final[struct.Struct] = struct.Struct('<f')
+ST_LE_F64: Final[struct.Struct] = struct.Struct('<d')
+ST_BE_I8: Final[struct.Struct] = struct.Struct('>b')
+ST_BE_I16: Final[struct.Struct] = struct.Struct('>h')
+ST_BE_I32: Final[struct.Struct] = struct.Struct('>i')
+ST_BE_I64: Final[struct.Struct] = struct.Struct('>q')
+ST_BE_UI8: Final[struct.Struct] = struct.Struct('>B')
+ST_BE_UI16: Final[struct.Struct] = struct.Struct('>H')
+ST_BE_UI32: Final[struct.Struct] = struct.Struct('>I')
+ST_BE_UI64: Final[struct.Struct] = struct.Struct('>Q')
+ST_BE_F32: Final[struct.Struct]  = struct.Struct('>f')
+ST_BE_F64: Final[struct.Struct] = struct.Struct('>d')
+# Structs that use the system byte order, where consistency on a single system
+# is all that matters. Mainly used for quick casts between signed and unsigned.
+ST_SBO_I8: Final[struct.Struct] = struct.Struct('@b')
+ST_SBO_I16: Final[struct.Struct] = struct.Struct('@h')
+ST_SBO_I32: Final[struct.Struct] = struct.Struct('@i')
+ST_SBO_I64: Final[struct.Struct] = struct.Struct('@q')
+ST_SBO_UI8: Final[struct.Struct] = struct.Struct('@B')
+ST_SBO_UI16: Final[struct.Struct] = struct.Struct('@H')
+ST_SBO_UI32: Final[struct.Struct] = struct.Struct('@I')
+ST_SBO_UI64: Final[struct.Struct] = struct.Struct('@Q')
+# Struct for an RGB value that, in little endian, would be an int written in hex
+# as 0x00BBGGRR.
+ST_RGB: Final[struct.Struct] = struct.Struct('<BBBx')

extract_msg/data/logging-config/logging-nt.json

@@ -0,0 +1,64 @@
+{
+    "version": 1,
+    "disable_existing_loggers": false,
+    "formatters": {
+        "simple": {
+            "format": "%(asctime)s - %(name)s - %(levelname)s - %(message)s"
+        }
+    },
+    "handlers": {
+        "console": {
+            "class": "logging.StreamHandler",
+            "level": "DEBUG",
+            "formatter": "simple",
+            "stream": "ext://sys.stdout"
+        },
+        "info_file_handler": {
+            "class": "logging.handlers.RotatingFileHandler",
+            "level": "INFO",
+            "formatter": "simple",
+            "filename": "%LOCALAPPDATA%/extract_msg/extract_msg.log",
+            "maxBytes": 10485760,
+            "backupCount": 20,
+            "encoding": "utf8"
+        },
+        "error_file_handler": {
+            "class": "logging.handlers.RotatingFileHandler",
+            "level": "ERROR",
+            "formatter": "simple",
+            "filename": "%LOCALAPPDATA%/extract_msg/extract_msg.log",
+            "maxBytes": 10485760,
+            "backupCount": 20,
+            "encoding": "utf8"
+        },
+        "warning_file_handler": {
+            "class": "logging.handlers.RotatingFileHandler",
+            "level": "WARNING",
+            "formatter": "simple",
+            "filename": "%LOCALAPPDATA%/extract_msg/extract_msg.log",
+            "maxBytes": 10485760,
+            "backupCount": 20,
+            "encoding": "utf8"
+        },
+        "developer_file_handler": {
+            "class": "logging.handlers.RotatingFileHandler",
+            "level": "DEVELOPER",
+            "formatter": "simple",
+            "filename": "%LOCALAPPDATA%/extract_msg/extract_msg.log",
+            "maxBytes": 10485760,
+            "backupCount": 20,
+            "encoding": "utf8"
+        }
+    },
+    "loggers": {
+        "my_module": {
+            "level": "ERROR",
+            "handlers": ["console"],
+            "propagate": "no"
+        }
+    },
+    "root": {
+        "level": "DEVELOPER",
+        "handlers": ["console", "info_file_handler", "error_file_handler", "warning_file_handler", "developer_file_handler"]
+    }
+}

extract_msg/data/logging-config/logging-posix.json

@@ -0,0 +1,64 @@
+{
+    "version": 1,
+    "disable_existing_loggers": false,
+    "formatters": {
+        "simple": {
+            "format": "%(asctime)s - %(name)s - %(levelname)s - %(message)s"
+        }
+    },
+    "handlers": {
+        "console": {
+            "class": "logging.StreamHandler",
+            "level": "DEBUG",
+            "formatter": "simple",
+            "stream": "ext://sys.stdout"
+        },
+        "info_file_handler": {
+            "class": "logging.handlers.RotatingFileHandler",
+            "level": "INFO",
+            "formatter": "simple",
+            "filename": "/var/log/extract_msg/extract_msg.log",
+            "maxBytes": 10485760,
+            "backupCount": 20,
+            "encoding": "utf8"
+        },
+        "error_file_handler": {
+            "class": "logging.handlers.RotatingFileHandler",
+            "level": "ERROR",
+            "formatter": "simple",
+            "filename": "/var/log/extract_msg/extract_msg.log",
+            "maxBytes": 10485760,
+            "backupCount": 20,
+            "encoding": "utf8"
+        },
+        "warning_file_handler": {
+            "class": "logging.handlers.RotatingFileHandler",
+            "level": "WARNING",
+            "formatter": "simple",
+            "filename": "/var/log/extract_msg/extract_msg.log",
+            "maxBytes": 10485760,
+            "backupCount": 20,
+            "encoding": "utf8"
+        },
+        "developer_file_handler": {
+            "class": "logging.handlers.RotatingFileHandler",
+            "level": "DEVELOPER",
+            "formatter": "simple",
+            "filename": "/var/log/extract_msg/extract_msg.log",
+            "maxBytes": 10485760,
+            "backupCount": 20,
+            "encoding": "utf8"
+        }
+    },
+    "loggers": {
+        "my_module": {
+            "level": "ERROR",
+            "handlers": ["console"],
+            "propagate": "no"
+        }
+    },
+    "root": {
+        "level": "DEVELOPER",
+        "handlers": ["console", "info_file_handler", "error_file_handler", "warning_file_handler", "developer_file_handler"]
+    }
+}