Covers #448, fix logic, more complete form of #447

Author: TheElementalOfDestruction

CHANGELOG.md

@@ -3,6 +3,9 @@
 * Fix an issue in `extract_msg.utils.msgPathToString()` that prevented backslashes from being replaced with forward slashes.
 * Change the behavior of `extract_msg.utils.minutesToDurationStr()` to properly use plurals.
 * Fixed issue in `extract_msg.utils.unwrapMsg()` that would prevent it from working on signed messages due to an API change.
+* Added new exception `MimetypeFailureError`.
+* Modified the logic of `MessageBase.asEmailMessage()` to use `AttachmentBase/SignedAttachment.name` instead of `getFilename()` which only exists on AttachmentBase.
+* Modified the logic of `MessageBase.htmlBodyPrepared()` to properly put the mimetype in image tags to ensure rendering. Logic was also modified to use `encode` instead of `prettify` to reduce computation and output size.
 
 **v0.52.0**
 * [[TeamMsgExtractor #444](https://github.com/TeamMsgExtractor/msg-extractor/issues/444)] Fix typo in string that prevented HTML body from generating from the plain text body properly.

extract_msg/__init__.py

@@ -27,7 +27,7 @@
 #    along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
 __author__ = 'Destiny Peterson & Matthew Walker'
-__date__ = '2024-10-24'
+__date__ = '2024-11-23'
 __version__ = '0.53.0'
 
 __all__ = [

extract_msg/exceptions.py

@@ -92,6 +92,11 @@ class InvalidPropertyIdError(ExMsgBaseException):
     The provided property ID was invalid.
     """
 
+class MimetypeFailureError(ExMsgBaseException):
+    """
+    The mimetype was unable to be properly determined when it was mandatory.
+    """
+
 class NotWritableError(ExMsgBaseException):
     """
     Modification was attempted on an instance that is not writable.

extract_msg/msg_classes/message_base.py

@@ -39,7 +39,8 @@
     )
 from ..exceptions import (
         ConversionError, DataNotFoundError, DeencapMalformedData,
-        DeencapNotEncapsulated, IncompatibleOptionsError, WKError
+        DeencapNotEncapsulated, IncompatibleOptionsError, MimetypeFailureError,
+        WKError
     )
 from .msg import MSGFile
 from ..structures.report_tag import ReportTag
@@ -178,7 +179,7 @@ def asEmailMessage(self) -> EmailMessage:
             if att.dataType:
                 if hasattr(att.dataType, 'asEmailMessage'):
                     # Replace the extension with '.eml'.
-                    filename = att.getFilename()
+                    filename = att.name or ''
                     if filename.lower().endswith('.msg'):
                         filename = filename[:-4] + '.eml'
                     msgMain.add_attachment(
@@ -1198,12 +1199,35 @@ def htmlBodyPrepared(self) -> Optional[bytes]:
         for tag in tags:
             # Iterate through the attachments until we get the right one.
             cid = tag['src'][4:]
-            data = next((attachment.data for attachment in self.attachments if attachment.cid == cid), None)
+            att = next((attachment for attachment in self.attachments if hasattr(attachment, 'cid') and attachment.cid == cid), None)
             # If we found anything, inject it.
-            if data:
-                tag['src'] = (b'data:image;base64,' + base64.b64encode(data)).decode('utf-8')
+            if att and isinstance(att.data, bytes):
+                # Try to get the mimetype. If we can't, see if the item has an
+                # extension and guess the mimtype for a few known ones.
+                mime = att.mimetype
+                if not mime:
+                    ext = (att.name or '').split('.')[-1].lower()
+                    if ext == 'png':
+                        mime = 'image/png'
+                    elif ext == 'jpg' or ext == 'jpeg':
+                        mime = 'image/jpeg'
+                    elif ext == 'gif':
+                        mime = 'image/gif'
+                    elif ext == 'tiff' or ext == 'tif':
+                        mime = 'image/tif'
+                    elif ext == 'bmp':
+                        mime = 'image/bmp'
+                    elif ext == 'svg':
+                        mime = 'image/svg+xml'
+                # Final check.
+                if mime:
+                    tag['src'] = (b'data:' + mime.encode() + b';base64,' + base64.b64encode(att.data)).decode('utf-8')
+                else:
+                    # We don't know what to actually put for this item, and we
+                    # really should never end up here, so throw an error.
+                    raise MimetypeFailureError('Could not get the mimetype to use for htmlBodyPrepared.')
 
-        return soup.prettify('utf-8')
+        return soup.encode('utf-8')
 
     @functools.cached_property
     def htmlInjectableHeader(self) -> str: