ansaurus

Question

Using regex to separate a multipart email

Answer 1

+3 A:

This get's a little complicated, but you can google MIME email structures if you want to get into the nitty gritty. I'm going to try to explain the email structure a little bit rather than attempt to answer with a specific RegEx (mainly because I'm not sure a RegEx will do what you want).

If you look at your raw emails you will see:

--1b4078c9-04f5-4cca-a220-e5b30eddef46

This is the MIME boundary, it separates the individual pieces of a MIME email message. A MIME email message can contain many parts, including an HTML version of the email, a plain text version, as well as file or image attachments. If you look at the two lines following the boundary, they explain what the upcoming part is by using it's MIME type.

If you look at the top of the raw email, you will see the 'Content-Type' header, which in a multi-part MIME message should be followed by a 'boundary=' section. You can take that boundary (as shown above) and use that to break up the pieces of your email.

What I think makes it difficult to do via a RegEx is that the boundary will be different for every email, so it's something that is more applicable to some code. You might want to use a RegEx to find the boundary, and then some logic to break the message up, maybe something like:

myMessage.Split(myBoundary)

Coding Gorilla 2010-06-18 18:53:50

Answer 2

A:

It seems to to me that the HTML part of your email begins shortly after Content-Type: text/html;, so I'd say those few lines would be a good token to indicate that the HTML is begining. As for a regex, I think this would do (.+)Content-Type: text/html; charset="iso-8859-1"(.+). The text part of your input will be in capture group 1, and the HTML part in capture group 2. You have to be able to set an option so that . matches \n as well as other characters.

FrustratedWithFormsDesigner 2010-06-18 18:57:51

Answer 3

A:

New to this site, just polking around for a few minutes before going to back work.

This thing here: "--1b4078c9-04f5-4cca-a220-e5b30eddef46" is a string declared as a MIME part separator in the headers of the incoming email message. Look for it and save it.

Looks like each MIME part delimited by that string breaks down into two sections: a list of name-value values and then the "real content" of the section, separated by a blank line. I'm pretty sure there's some hairy MIME standard that says what I just said. :) In the first ("attributes") section, look for the content-type you want (text/html). Once you've found it, look for the next blank line, and once you've found THAT, suck in the content until the next MIME part delimiter. That'll be your html email message, which you can then process.

You will not be able to do all this with with one magnificent regex, I believe. You'll have to code a loop over each line of the incoming message, and do a state machine sorta thing. States will be something like:

(1) Mime part delimiter unknown (2) Mime part delimiter known, mime part unseen (3) Mime part seen, type not html (4) Mime part seen, type is html, blank line not seen (5) Blank line seen, mime part delimiter unseen (6) Mime part delimiter seen (end of processing)

I'm being very fast , loose, sloppy. Hope this helps.

John.

JohnL4 2010-06-18 19:00:12

Answer 4

+1 A:

There are a few open source C# MIME parsers available:

The last two are a bit old. If they don't easily compile, their source might point you in the right direction.

Remember, an email can contain an attachment that is an email that contains an attachment, etc, etc... At some point, Regex will let you down.

Corbin March 2010-06-18 19:19:48

I think you may have rendered a whole day of my programming useless :/

Immanu'el Smith 2010-06-18 20:04:32

ansaurus

tags:

views:

answers:

Using regex to separate a multipart email

related questions