I can connect to the IMAP mail server easy enough:
use Mail::IMAPClient;
use MIME::Base64;
use MIME::Parser;
my $imap = Mail::IMAPClient->new(
Server => '192.168.2.2',
User => 'xxxxxx',
Password => 'yyyyyy',
Ssl => 1,
Uid => 1,
);
my $folders = $imap->folders
or die "List folders error: ", $imap->LastError, "\n";
print "Folders: @$folders\n";
$sfolder="INBOX.2012";
$imap->select( $sfolder )
or die "Select '$Opt{sfolder}' error: ", $imap->LastError, "\n";
my @msgs = $imap->messages or die "Could not messages: $@\n";
However, the text and html I want is not easily parsed due to codes like this:
Content-Transfer-Encoding:base64
Content-Type:text/html; charset=utf-8
Content-Transfer-Encoding:base64
Content-Type:text/html; charset=utf-8
Content-Transfer-Encoding:
Content-Type:multipart/mixed; boundary="----------=_4F0F4830.7079357A"
Multipart
Content-Transfer-Encoding:
Content-Type:multipart/mixed; boundary="----=_Part_4487195_1184536749.1326753403034"
Multipart
Content-Transfer-Encoding:
Content-Type:multipart/alternative; boundary=--boundary_164442_d184e417-739f-
46d6-824a-6ea1846e79de
Multipart
Content-Transfer-Encoding:
Content-Type:multipart/mixed; boundary="----=_Part_3882878_23916831.1326509484032"
Multipart
Content-Transfer-Encoding:
I tried this but it only works on a tiny number of different encodings.
if ($imap->get_header($msg,"Content-Transfer-Encoding")=~ /base64/i) {
print "\nMatch base64";
if ($imap->get_header($msg,"Content-Type")=~m/text/i ) {
push(@mail,decode_base64($imap->body_string($msg)));
}
elsif ($imap->get_header($msg,"Content-Type")=~m/image/i )
{ print "\nImage detected"; }
elsif ($imap->get_header($msg,"Content-Type")=~m/application/i )
{ print "\nApplication detected"; }
There are 7bit and 8bit variants and other encoding methods that contain the html or text I want for later use. I successfully use decode_base64() to decode base64. The worse ones to decode are the ones that contain multi-part codes. I feel like I am re-inventing the wheel and there must be a library or module that can do all the heavy lifting for me.
Other content types such as .jpg,.gif, and .pdf should simply be ignored. The multi-part emails contain at least 1 part that I an interested but many that are useless to me.
After further research this structure has some of the information I need but don’t know how to get it out efficiently is another matter.
Dumping:$VAR1 = bless( {
'bodyparms' => {
'boundary' => '----=_NextPart_002_BC64_7D688C1F.A2FF9BE0'
},
'bodyextra' => undef,
'_top' => 1,
'bodydisp' => 'NIL',
'_id' => 'HEAD',
'bodysubtype' => 'mixed',
'PartsIndex' => {
'1.3' => bless( {
'bodyparms' => 'NIL',
'bodyid' => '<d9e26cc0-019c-4ac0-9b1e-9c9ac8424f52>',
'bodyextra' => 'NIL',
'bodydisp' => 'NIL',
'_id' => '1.3',
'bodysubtype' => 'jpeg',
'_prefix' => '1.3',
'bodysize' => '4808',
'bodytype' => 'image',
'bodyMD5' => 'NIL',
'bodylang' => 'NIL',
'bodydesc' => 'NIL',
'bodyenc' => 'base64'
}, 'Mail::IMAPClient::BodyStructure' ),
'1.1' => bless( {
'bodyparms' => {
'boundary' => '----=_NextPart_000_36AE_880DDD08.0A776E35'
},
'bodyextra' => undef,
'bodydisp' => 'NIL',
'_id' => '1.1',
'bodysubtype' => 'alternative',
'_prefix' => '1.1',
'bodytype' => 'MULTIPART',
'bodystructure' => [
bless( {
'bodyparms' => {
'charset' => 'utf-8'
},
'bodyextra' => 'NIL',
'bodyid' => 'NIL',
'bodydisp' => 'NIL',
'_id' => '1.1.1',
'bodysubtype' => 'PLAIN',
'_prefix' => '1.1.1',
'bodysize' => '1971',
'bodytype' => 'TEXT',
'bodyMD5' => 'NIL',
'textlines' => '74',
'bodylang' => 'NIL',
'bodydesc' => 'NIL',
'bodyenc' => 'quoted-printable'
}, 'Mail::IMAPClient::BodyStructure' ),
bless( {
'bodyparms' => {
'charset' => 'utf-8'
},
'bodyextra' => 'NIL',
'bodyid' => 'NIL',
'bodydisp' => 'NIL',
'_id' => '1.1.2',
'bodysubtype' => 'HTML',
'_prefix' => '1.1.2',
'bodysize' => '23364',
'bodytype' => 'TEXT',
'bodyMD5' => 'NIL',
'textlines' => '331',
'bodylang' => 'NIL',
'bodydesc' => 'NIL',
'bodyenc' => 'quoted-printable'
}, 'Mail::IMAPClient::BodyStructure' )
],
'bodyloc' => 'NIL',
'bodylang' => 'NIL'
}, 'Mail::IMAPClient::BodyStructure' ),
'1' => bless( {
'bodyparms' => {
'boundary' => '----=_NextPart_001_EA96_2BF8DEDE.32622D51'
},
'bodyextra' => undef,
'bodydisp' => 'NIL',
'_id' => 1,
'bodysubtype' => 'related',
'_prefix' => 1,
'bodytype' => 'MULTIPART',
'bodystructure' => [
$VAR1->{'PartsIndex'}{'1.1'},
bless( {
'bodyparms' => 'NIL',
'bodyid' => '<5dff39db-e81c-4410-be75-8662564fd328>',
'bodyextra' => 'NIL',
'bodydisp' => 'NIL',
'_id' => '1.2',
'bodysubtype' => 'jpeg',
'_prefix' => '1.2',
'bodysize' => '14406',
'bodytype' => 'image',
'bodyMD5' => 'NIL',
'bodylang' => 'NIL',
'bodydesc' => 'NIL',
'bodyenc' => 'base64'
}, 'Mail::IMAPClient::BodyStructure' ),
$VAR1->{'PartsIndex'}{'1.3'},
bless( {
'bodyparms' => 'NIL',
'bodyid' => '<717f2ef4-f795-4d1c-87cc-283c9b0a59b0>',
'bodyextra' => 'NIL',
'bodydisp' => 'NIL',
'_id' => '1.4',
'bodysubtype' => 'gif',
'_prefix' => '1.4',
'bodysize' => '2912',
'bodytype' => 'image',
'bodyMD5' => 'NIL',
'bodylang' => 'NIL',
'bodydesc' => 'NIL',
'bodyenc' => 'base64'
}, 'Mail::IMAPClient::BodyStructure' )
],
'bodyloc' => 'NIL',
'bodylang' => 'NIL'
}, 'Mail::IMAPClient::BodyStructure' ),
'1.2' => $VAR1->{'PartsIndex'}{'1'}{'bodystructure'}[1],
'1.1.2' => $VAR1->{'PartsIndex'}{'1.1'}{'bodystructure'}[1],
'2' => bless( {
'bodyparms' => {
'name' => 'BKD-7361945220.pdf'
},
'bodyid' => 'NIL',
'bodyextra' => 'NIL',
'bodydisp' => {
'attachment' => {
'filename' => 'BKD-7361945220.pdf'
}
},
'_id' => 2,
'bodysubtype' => 'octetstream',
'_prefix' => 2,
'bodysize' => '47540',
'bodytype' => 'application',
'bodyMD5' => 'NIL',
'bodystructure' => [],
'bodylang' => 'NIL',
'bodydesc' => 'NIL',
'bodyenc' => 'base64'
}, 'Mail::IMAPClient::BodyStructure' ),
'1.4' => $VAR1->{'PartsIndex'}{'1'}{'bodystructure'}[3],
'1.1.1' => $VAR1->{'PartsIndex'}{'1.1'}{'bodystructure'}[0]
},
'_prefix' => 'HEAD',
'PartsList' => [
1,
'1.1',
'1.1.1',
'1.1.2',
'1.2',
'1.3',
'1.4',
2
],
'bodytype' => 'MULTIPART',
'bodystructure' => [
$VAR1->{'PartsIndex'}{'1'},
$VAR1->{'PartsIndex'}{'2'}
],
'bodyloc' => 'NIL',
'bodylang' => 'NIL'
}, 'Mail::IMAPClient::BodyStructure' );
As you can see none of the values are guaranteed to be part of every part on the PartsIndex and some them are nested.
variable of interest for each PartsIndex item:
bodytype
bodysubtype
bodyenc
I tried using a couple of prebuilt modules but they had too many dependencies and was hard to work with. This solution adds no dependencies beyond the original. I also had issues with the dependencies for libMagic, see above, and I did not want anyone who uses my program to have to deal with that issue either.
You have to call decode twice once for the main parent, and again for each child. Since this $imap->get_bodystructure($msg); contains all the information you need why add dependencies where none are needed. It took many many hours to figure out how to decode it manually, but it was worth it.
You can add whatever decoders you want to the decode() subroutine. I only need to decode the text/html and base64 encoded versions there of. The IMAPClient functions give you a list of all parents and children so you don’t have to go making a list by yourself. The tricky part is you can have any number of parent each with any number of children, but only the children contain useful data. The parents can be ignored, since many of their values are blank,undef, or ‘NIL’ (literally). In fact a vast number of variables have the value of ‘NIL’. Even ones that the email client could have answered for the user like bodyMD5 and bodylang are USUALLY equal ‘NIL’. Due to the overwhelming use of ‘NIL’ parsing and using other fields may prove futile. Depend on your imap server and the people you recieve email from you mileage may vary.
If you have further questions leave a comment.