extract-msg

v0.33.0

• [TeamMsgExtractor #212] Added support for Task objects.
• Added additional parameter overrideClass to all _ensureSetX type functions. This parameter is a class to initialize using the data, if provided. The data will be provided as the first argument to the __init__ function of the class if the data is not None. If the data is done, the value set and returned from _ensureSetX will be None. This can be overriden using the second added parameter, which defaults to this behavior. Simply set preserveNone to False to force the data to be passed to the class. Keep in mind that this means the class will receive None as it's first parameter.
• Updated Named class to make it more like the dictionary it is build on top of. Specifically, added keys, values, and items as functions.
• Fixed issue in Named where old code wasn't deleted, causing some named properties to be completely omitted.
• Improved efficiency of Named by making it so get no longer copied the entire dictionary repeatedly while looking for the property.
• Fixed typo in Attachment that caused embedded msg files to still regenerate the Named instance even though they should have been using the parent's.
• Updated fields in MSGFile to use enums.
• Added additional properties to MSGFile.
• Significant cleanup.

v0.32.0

• [TeamMsgExtractor #217] Redesigned named properties to be much more efficient. All in all, load times for MSG files are significantly improved all around.
  • Named properties load dynamically.
  • Named properties use a single shared instance for embedded msg files to contain all of the entries, and tiny individual instances for actually accessing based on what is accessing.
  • Named has been updated so it's methods are more standardized.
  • Named is no longer usable to directly access property values. Instead, access to them is done through NamedProperties.
  • _registerNamedProperty has been removed due to no longer being part of the named properties framework.
• Actually removed the exception I meant to remove in 0.31.0.
• Changed Attachment.type to an enum instead of a string. This makes it easier to see all of the possible values.
• Added option to allow attachment saving to be skipped when calling Message.save.

v0.31.1

• Updated signed attachment mimetype property from mime to mimetype to match with the regular attachment property.

v0.31.0

• [TeamMsgExtractor #223] First implementation of support for signed messages.
• [TeamMsgExtractor #256] Extended the properties of the attachment class to allow for access to more data.
• Fixed some issues with the contact class (some of the code was left unfinished).
• Converted many arguments for msg files to keyword arguments (**kwargs). This is a breaking change if your code provides anything except the path to the msg file. This will, however, make it easier to add arguments without the function signature being polluted with numerous arguments.
• Updated exception blocks to ensure they would only catch exceptions that are instances of Exception or a child of Exception. A few were missing this check and so might have caught unwanted exceptions.
• Shifted some minor parts down towards the AttachmentBase class where possible to allow more usability with broken and unsupported attachments. Nothing about these properties required the attachment to be fully processed, which is why they were moved down. From the perspective of using the Attachment class nothing has changed. The following properties have been moved: attachmentEncoding, additionalInformation, cid/contentId, longFilename, renderingPosition, and shortFilename.
• Added link to readme for supporting development.
• Finally found the behavior to use for missing encoding. As such, MissingEncodingError is being removed from the module entirely, as it is no longer a thing.
• Moved many constant sets to enums. This makes things a lot more organized.

v0.30.14

• Fixed major bug in MSGFile.listDir that would cause it to give the wrong data due to caching. I forgot to have it cache a different result for each set of options, so it would always give the same result regardless of options after the first access.
• Minor updates to setup.py.
• Fixed URLs in README.

v0.30.13

• [TeamMsgExtractor #257] Fixed missing documentation for customPath keyword argument to Attachment.save.

v0.30.12

• [TeamMsgExtractor #253] Fixed various docstring issues.

v0.30.11

• [TeamMsgExtractor #249] Fixed an error with opening the body causing the file to throw an uncaught exception in the __init__ function of MessageBase. The error may still come up later, but you will still have general access to the instance.
• Fixed an issue with part of the command line documentation.

v0.30.10

• [TeamMsgExtractor #249] Fixed exception catching not properly accessing the exception (forgot to go through one of the submodules to access exception).
• Updated docstring for MessageBase.deencapsulatedRtf.

v0.30.9

• Fixed the behavior of Properties.get so it actually behaves like a dict (that was the intent of it, but I did it the wrong way for some reason).
• Fixed a type that caused an exception when no HTML body could be found nor generated.

v0.30.8

• Update imapclient requirement to >=2.1.0 instead of ==2.1.0. Currently there are no changes that would prevent current future versions from working.

v0.30.7

• [TeamMsgExtractor #239] Fixed msg.py not having import pathlib.
• After going through the details of the example MSG files provided with the module, specifically unicode.msg, I now am glad I decided to put in some fail-safes in the HTML body processing. One of them does not have an <html>, <head>, nor <body> tag, and so would have had an error. This will actually prevent the header from injecting properly as well, so a bit of validation before was made necessary to ensure the HTML saving would still work.
• Added new exception BadHtmlError.
• Added new function utils.validateHtml.
• Updated README credits.
• Changed header logic to generate manually if the header data has been stripped (filled with null bytes) and not just if the stream does not exist.

v0.30.6

• Small adjustments to internal code to make it a bit better.
• Added Message.getSaveBody, Message.getSaveHtmlBody, and Message.getSaveRtfBody. These three functions generate their respective bodies that will be used when saving the file, allowing you to retrieve the final products without having to write them to the disk first. All arguments that are passed to Message.save that would influence the respective bodies are passed to their respective functions.
• I thought I added the documentation for attachmentsOnly to Message.save but apparently it was missing. Not sure what happened there but I made sure it was added this time.
• Added new option to Message.save called charset. This is used in the preparation of the HTML body when using preparedHtml. This is also usable with --charset CHARSET from the command line.

v0.30.5

• [TeamMsgExtractor #225] Added the ability to generate the HTML body from the RTF body if it is encapsulated HTML. If there is no RTF body, then it will do a very basic generation from the plain text body. This process is automatically performed if the HTML body is missing.
• Added the ability for the plain text body to sometimes generate from the RTF body if the plain text body does not exist.
• Added documentation to Message.save for the attachmentsOnly option.

v0.30.4

• [TeamMsgExtractor #233] Added option to Message.save to only save the attachments (attachmentsOnly). This can also be used from the command line with the option --attachments-only.
• Corrected error string for incompatible options in Message.save.
• Fixed if blocks being nested weirdly in Message.save instead of being chained with elif. Probably an artifact left from adding support for HTML and RTF.

v0.30.3

• [TeamMsgExtractor #232] Updated list of known MSG class types so the module would correctly give UnsupportedMSGTypeError instead of UnrecognizedMSGTypeError.

v0.30.2

• Fixed typo in utils.knownMsgClass.
• Updated contributing guidelines and pull request template. Pull requests were updated to match the new structure of the module while the guidelines were updated for clarity.

v0.30.1

• [TeamMsgExtractor #102] Added property MessageBase.htmlBodyPrepared which provides the HTML body that has been prepared for actual use or conversion to PDF. All attachments that can be injected into the body directly will be injected.
• Corrected a mistake in the documentation of Message.save.
• Fixed issue where the command line parser was not checking for raw in conjunction with other saving options.
• Changed utils.setupLogging to use pathlib.
• Improved reliability of the logic in utils.setupLogging.
• Removed function utils.getContFileDir.
• Cleaned up plain Exception being raised at the end of utils.injectRtfHeader if the injection failed. This now raises RuntimeError, an error that should be reported so the injection system can be improved.
• Added parsing for PtypServerId to utils.parseType.
• Added FolderID, MessageID, and ServerID to data.
• Added zip output to the command line.
• Added support for PtypMultipleFloatingTime to utils.parseType.
• Improved documentation of many functions with the exceptions they may raise.
• Changed the way zip files are handled so that files written to them actually have a modification date now.

v0.30.0

• Removed all support for Python 2. This caused a lot of things to be moved around and changed from indirect references to direct references, so it's possible something fell through the cracks. I'm doing my best to test it, but let me know if you have an issue.
• Changed classes to now prefer super() over direct superclass initialization.
• Removed explicit object subclassing (it's implicit in Python 3 so we don't need it anymore).
• Converted most .formats into f strings.
• Improved consistency of docstrings. It's not perfect, but it should at least be better.
• Started the addition of type hints to functions and methods.
• Updated utils.bytesToGuid to make it faster and more efficient.
• Renamed utils.msgEpoch to utils.filetimeToUtc to be more descriptive.
• Updated internal variable names to be more consistent.
• Improvements to the way __main__ works. This does not affect the output it will generate, only the efficiency and readability.

v0.29.6

• Corrected error string for incompatible options in Message.save.

v0.29.5

• Updated list of known MSG class types so the module would correctly give UnsupportedMSGTypeError instead of UnrecognizedMSGTypeError.

v0.29.4

• Fixed typo in utils.knownMsgClass.
• Updated contributing guidelines and pull request template.

v0.29.3

• [TeamMsgExtractor #226] Fix typo in command parsing that prevented the usage of allowFallback.
• Fixed main still manually navigating to a new directory with os.chdir instead of using customPath.
• Fixed issue in main where the --html option was being using for both html and rtf. This meant if you wanted rtf it would not have used it, and if you wanted html it would have thrown an error.
• Fixed --out-name having no effect.
• Fixed --out having no effect.

v0.29.2

• Fixed issue where the RTF injection was accidentally doing HTML escapes for non-encapsulated streams and not doing escapes for encapsulated streams.
• Fixed name error in Message.save causing bad logic. For context, the internal variable zip was renamed to _zip to avoid a name conflict with the built-in function. Some instances of it were missed.

v0.29.1

• [TeamMsgExtractor #198] Added a feature to save the header in it's own file (prefers the full raw header if it can find it, otherwise puts in a generated one) that was actually supposed to be in v0.29.0 but I forgot, lol.

v0.29.0

• [TeamMsgExtractor #207] Made it so that unspecified dates are handled properly. For clarification, an unspecified date is a custom value in MSG files for dates that means that the date is unspecified. It is distinctly different from a property not existing, which will still return None. For unspecified dates, datetime.datetime.max is returned. While perhaps not the best solution, it will have to do for now.
• Fixed an issue where utils.parseType was returning a string for the date when it makes more sense to return an actual datetime instance.
• [TeamMsgExtractor #165] [TeamMsgExtractor #191] Completely redesigned all existing save functions. You can now properly save to custom locations under custom file names. This change may break existing code for several reasons. First, all arguments have been changed to keyword arguments. Second, a few keyword arguments have been renamed to better fit the naming conventions.
• [TeamMsgExtractor #200] Changed imports to use relative imports instead of hard imports where applicable.
• Updated the save functions to no longer rely on the current working directory to save things. The module now does what it can to use hard pathing so that if you spontaneously change working directory it will not cause problems. This should also allow for saving to be threaded, if I am correct.
• [TeamMsgExtractor #197] Added new property Message.defaultFolderName. This property returns the default name to be used for a Message if none of the options change the name.
• [TeamMsgExtractor #201] Fixed an issue where if the class type was all caps it would not be recognized. According to the documentation the comparisons should have been case insensitive, but I must have misread it at some point.
• [TeamMsgExtractor #202] Module will now handle path lengths in a semi-intelligent way to determine how best to save the MSG files. Default path length max is 255.
• [TeamMsgExtractor #203] Fixed an issue where having multiple "." characters in your file name would cause the directories to be incorrectly named when using the useFileName (now useMsgFilename) argument in the save function.
• [TeamMsgExtractor #204] Fixed an issue where the failsafe name used by attachments wasn't being encoded before hand causing encoding errors.
• MSG files with a type of simply IPM will now be returned as MSGFile by openMsg, as this specifies that no format has been specified.
• [TeamMsgExtractor #214] Attachments that error because the MSG class type wasn't recognized or isn't supported will now correctly be UnsupportedAttachment instead of BrokenAttachment.
• Improved internal code in many functions to make them faster and more efficient.
• openMsg will now tell you if a class type is simply unsupported rather than unrecognized. If it is found in the list, the function will raise UnsupportedMSGTypeError.
• Added caching to MSGFile.listDir. I found that if you have larger files this single function might be taking up over half of the processing time because of how many times it is used in the module.
• Fully implemented raw saving.
• Extended the Contact class to have more properties.
• Added new function MSGFile._ensureSetTyped which acts like the other ensure set functions but doesn't require you to know the type. Prefer to use other ensure set function when you know exactly what type it will be.
• Changed Message.saveRaw to MSGFile.saveRaw.
• Changed MSGFile.saveRaw to take a path and save the contents to a zip file.
• Corrected the help doc to reflect the current repository (was still on mattgwwalker).
• Fixed a bug that would cause an exception on trying to access the RTF body on a file that didn't have one. This is now correctly returning None.
• The raw keyword of Message.save now actually works.
• Added property Attachment.randomFilename which allows you to get the randomly generated name for attachments that don't have a usable one otherwise.
• Added function Attachment.regenerateRandomName for creating a new random name if necessary.
• Added function Attachment.getFilename. This function is used to get the name an attachment will be saved with given the specified arguments. Arguments are identical to Attachment.save.
• Changed pull requests to reflect new style.
• Added additional properties for defined MSG file fields.
• Added zip file support for the Attachment.save and Message.save. Simply pass a path for the zip keyword argument and it will create a new ZipFile instance and save all of it's data inside there. Alternatively, you can pass an instance of a class that is either a ZipFile or ZipFile-like and it will simply use that. When this argument is defined, the customPath argument refers to the path inside the zip file.
• Added the html and rtf keywords to Message.save. These will attempt to save the body in the html or rtf format, respectively. If the program cannot save in those formats, it will raise an exception unless the allowFallback keyword argument is True.
• Changed utils.hasLen to use hasattr instead of the try-except method it was using.
• Added new option recipientSeparator to MessageBase allowing you to specify a custom recipient separator (default is ";" to match Microsoft Outlook).
• Changed the openMsg function in Attachment to not be strict. This allows you to actually open the MSG file even if we don't recognize the type of embedded MSG that is being used.
• Attempted to normalize encoding names throughout the module so that a certain encoding will only show up using one name and not multiple.
• Finally figured out what CRC32 algorithm is used in named properties after directly asking in a Microsoft forum (see the thread here). Fortunately the is already defined in the compressed-rtf module so we can take advantage of that.
• Reworked MessageBase._genRecipient to improve it (because what on earth was that code it was using before?). Variables in the function are now more descriptive. Added comments in several places.
• Many renames to better fit naming convention:
  • dev.setup_dev_logger to dev.setupDevLogger.
  • MSGFile.fix_path to MSGFile.fixPath.
  • MessageBase.save_attachments to MessageBase.saveAttachments.
  • *.Exists to exists.
  • *.ExistsTypedProperty to *.existsTypedProperty.
  • prop.create_prop to prop.createProp.
  • Properties.attachment_count to Properties.attachmentCount.
  • Properties.next_attachment_id to Properties.nextAttachmentId.
  • Properties.next_recipient_id to Properties.nextRecipientId.
  • Properties.recipient_count to Properties.recipientCount.
  • utils.get_command_args to utils.getCommandArgs.
  • utils.get_full_class_name to utils.getFullClassName.
  • utils.get_input to utils.getInput.
  • utils.has_len to utils.hasLen.
  • utils.setup_logging to utils.setupLogging.
  • constants.int_to_data_type to constants.intToDataType.
  • constants.int_to_intelligence to constants.intToIntelligence.
  • constants.int_to_recipient_type to constants.intToRecipientType.
  • Misc internal function variables.

v0.28.5

• [TeamMsgExtractor #189] Forgot to import prepareFilename in attachment.py.
• Fixed bad link in the changelog.

v0.28.4

• [TeamMsgExtractor #184] Added code to Message to ensure subjects with null characters get stripped of them.
• Moved code for stripping subjects of bad characters to prepareFilename in utils.

v0.28.3

• Fixed minor typo in an exception description.
• Updated the README this time. Forgot to do it for at least 1 update.

v0.28.2

• Started preparing more of the code for when HTML and RTF saving are fully implemented. Please note that they do not work at all right now. Commented out the code for this because it wasn't meant to be uncommented.
• [TeamMsgExtractor #184] Added code to ensure file names don't have null characters when saving an attachment.
• Minor improvement to the section of the save code that checks if you have provided incompatible options.
• [TeamMsgExtractor #185] Added the IncompatibleOptionsError. It was supposed to be added a few updates ago, but was accidentally left out.
• Modified Message.save to return the current Message instance to allow for chained commands. This allows you to do something like extract_msg.openMsg("path/to/message.msg").save().close() where you could not before.

v0.28.1

• [TeamMsgExtractor #181] Fixed issue in Attachment that arose when moving some of the code to a base class.
• Fixed small error in utils.parse_type that caused it to incorrectly compare expected and actual length. Fortunately, this had no actual effect aside from a warning.
• Added the ebcdic module to the requirements to add more supported encodings.

v0.28.0

• [TeamMsgExtractor #87] Added a new system to handle NotImplementedError and other exceptions. All msg classes now have an option called attachmentErrorBehavior that tells the class what to do if it has an error. The value should be one of three constants: ATTACHMENT_ERROR_THROW, ATTACHMENT_ERROR_NOT_IMPLEMENTED, or ATTACHMENT_ERROR_BROKEN. ATTACHMENT_ERROR_THROW tells the class to not catch and exceptions and just let the user handle them. ATTACHMENT_ERROR_NOT_IMPLEMENTED tells the class to catch NotImplementedError exceptions and put an instance of UnsupportedAttachment in place of a regular attachment. ATTACHMENT_ERROR_BROKEN tells the class to catch all exceptions and either replace the attachment with UnsupportedAttachment if it is a NotImplementedError or BrokenAttachment for all other exceptions. With both of those options, caught exceptions will be logged.
• In making the previous point work, much code from Attachment has been moved to a new class called AttachmentBase. Both BrokenAttachment and UnsupportedAttachment are subclasses of AttachmentBase meaning data can be extracted from their streams in the same way as a functioning attachment.
• [TeamMsgExtractor #162] Pretty sure I actually got it this time. The execution flag should be applied by pip now.
• Fixed typos in some exceptions

v0.27.16

• [TeamMsgExtractor #177] Fixed incorrect struct being used. It should be the correct one now, but further testing will be required to confirm this.
• Fixed log error message in extract_msg.prop to actually format a value into the message.

v0.27.15

• [TeamMsgExtractor #177] Fixed missing import.

v0.27.14

• [TeamMsgExtractor #173] Fixed typo that I made in the last version that broke things. I didn't have the resources to test this one myself, unfortunately.
• Fixed a typo in an exception message.

v0.27.12

• [TeamMsgExtractor #173] Made an assumption about where an exception was thrown from and was wrong. While that location would have throw an exception, the function that called that code was the one to actually throw the exception in question. This issue should be fixed...
• [TeamMsgExtractor #162] Made another attempt to fix the execution flag on the wrapper script.

v0.27.11

• [TeamMsgExtractor #173] Tentatively implemented type 0x1014 (PtypMultipleInteger64). Apparently I forgot to do it earlier.

v0.27.9

• [TeamMsgExtractor #161] Added commands to the command line that will allow the user to specify that they want the message data to be output to stdout rather than to a file.
• [TeamMsgExtractor #162] Added a wrapper for extract_msg that will be installed.
• Fixed some of the encoding names to allow them to actually be used in Python. The names they previously held were not aliases that currently exist.
• Added more documentation to constants.CODE_PAGES to give more information about what it is. As it is a list of the possible encodings an msg file can use, I also specified which ones were supported by Python 3.
• Moved the main code into a function so it is now callable from outside of the file.

v0.27.8

• [TeamMsgExtractor #158] Fixed a spelling error in a function name that was causing it to not be seen. The function was called ceilDiv but was accidentally called as cielDiv.

v0.27.6

• Fixed a few lines in data.py.

v0.27.5

• Fixed an error in utils.divide that would cause it to drop the extra data if there was not enough to create a full division. For example, if you had a string that was 10 characters, and divided by 3, you would only receive a total of 9 characters back.
• Added some useful functions that will be used in the future.
• [TeamMsgExtractor] Updated to use new version of tzlocal.
• Updated changelog to fit new repository.

v0.27.4

• [mattgwwalker #152] Fixed an issue where the name of an exception was put as the wrong thing.

v0.27.3

• [mattgwwalker #105] Added code to fix an internal msg issue that had recipient lists being split up in the header after a certain amount of characters. This was now a bug on our part, but an issue with the generation of the msg file itself.
• Exposed the MessageBase class directly from extract_msg. I forgot to do this when I created it.
• Added MessageBase.bcc. I think it used to exist but got erased somehow on accident. Either way, it exists now.

v0.27.1

• [mattgwwalker #147]Fixed an issue in Message.save caused by it attempting to directly access a private variable that was moved to the base class MessageBase.

v0.27.0

• [mattgwwalker #143] Added new class Appointment that can handle outlook appointments or meetings.
• [mattgwwalker #143] Added new class MessageBase for classes that end up being mostly like the Message class. Currently subclassed by Message and Appointment. This should not have a direct effect on any code that uses this module.
• Added support for PtypFloatingTime in utils.parseType.
• Added proper support for PtypTime in FixedLengthProp.parseType
• Added pretty print functions to the Properties class and the Named class.
• Added new function MSGFile._ensureSetProperty that acts like MSGFile._ensureSet except that it works with properties from the Properties instance.
• Added equivalent of the previously mentioned function to the Attachment class and the Recipient class.

v0.26.4

• Added new function MSGFile._ensureSetNamed which acts like MSGFile._ensureSet except that it works with named properties.
• Added a version of the previous function to the Attachment and Recipient classes.
• Added new file data.py that contains various data structures that are used for specific properties.
• Expanded the functionality of the Recipient class by adding more properties.
• Added new functions Named.getNamed and Named.getNamedValue which retrieves a named property or the value of a named property, respectively, based on its name.

v0.26.3

• Added new function MSGFile.save that causes it and subclasses to raise a NotImplementedError if they do not override it.
• Fixed some issues in the changelog.
• Added some additional constants for future use.

v0.26.1

• Fixed an issue in openMsg that would leave the basic MSGFile instance open with the function returned, even if the function was returning a specific msg instance.

v0.24.4

• Added new property Message.isRead to show whether the email has been marked as read.
• Renamed Message.header_dict to Message.headerDict to better match naming conventions.
• Renamed Message.message_id to Message.messageId to better match naming conventions.