,h9ddlZddlZddlZddlZddlZddlmZddlm Z ddl m Z ddl mZmZmZddlmZddlmZddlmZdd lmZmZmZej8eZej?ej@Gd d eZ!y) N) decode_utf7)Parser) constants) AttachmentBrokenAttachmentUnsupportedAttachment)os_)MSGFile) Recipient) addNumToDir inputToBytes inputToStringceZdZdZdedddej fdZdZd dZ dZ d Z d!d Z e d Ze d Ze d Ze dZe dZe dZe dZe dZe dZe dZe dZe dZe dZe dZe dZe dZe dZe dZe dZ e dZ!e dZ"y)" MessageBasez0 Base class for Message like msg files. NFc Xtj|||||||||_d|_|j|j |j |s |j|j|j|j|jd|_ |j|jy)a :param path: path to the msg file in the system or is the raw msg file. :param prefix: used for extracting embeded msg files inside the main one. Do not set manually unless you know what you are doing. :param attachmentClass: optional, the class the Message object will use for attachments. You probably should not change this value unless you know what you are doing. :param filename: optional, the filename to be used by default when saving. :param delayAttachments: optional, delays the initialization of attachments until the user attempts to retrieve them. Allows MSG files with bad attachments to be initialized so the other data can be retrieved. :param overrideEncoding: optional, an encoding to use instead of the one specified by the msg file. Do not report encoding errors caused by this. F N)r __init__ _MessageBase__attachmentsDelayed_MessageBase__attachmentsReadymainPropertiesheader recipients attachmentstoccsenderdate_MessageBase__crlfbodynamed)selfpathprefixattachmentClassfilenamedelayAttachmentsoverrideEncodingattachmentErrorBehaviors Z/var/www/html/Resume-Scraper/venv/lib/python3.12/site-packages/extract_msg/message_base.pyrzMessageBase.__init__s& tV_hHXZqr$4!"'             c~d|z} t||S#t$rd}|jr|j|}||jr$tj dj |g}|jD]0}|jdz|k(s|j|j2t|dkDrF|d}t|dkDr1tdt|D]}|dj ||z }|}||jdd jd d jd d }|jd d jd d jdd }|jddk7r'|jdd }|jddk7r't||||cYSwxYw)z7 Returns the specified recipient field _NzMHeader found, but "{}" is not included. Will be generated from other streams.rz, {0}z  z z   rz )getattrAttributeError headerInitrloggerinfoformatrtypeappend formattedlenrangereplacefindsetattr)r" recipientType recipientIntprivatevaluefxsts r* _genRecipientzMessageBase._genRecipientAs % 4) ) E  M2}??$KK o v vxE!FG.Avv *l:-.q6A:1B1vz!&q#a&!17A'..1"66B7E  i5==iMUUV^`cd fc2::4EMMdTWXjj&",!MM$4Ejj&", D'5 )L1 s BF<DF<*F<;F<c|jr8|js, |j|jj |||fy|j D]}|j |||y#t$r g|_YVwxYwN)attachmentsDelayedattachmentsReady_MessageBase__waitingPropertiesr5r;r_registerNamedProperty)r"entry_typename attachments r*rOz"MessageBase._registerNamedPropertybs  " "4+@+@ .((  $ $ + +UE4,@ A".. F 11%E F " .+-( .s A**A=<A=c |j|jD],}|jdk(s|jj . t j|y#t $rY!wxYw)Nmsg) _attachmentsrr:datacloser5r )r"rSs r*rXzMessageBase.closemsc    ".. , ??e+OO))+ ,  d   s*A A A,+A,c< |jy#t$rYywxYw)zA Checks whether the header has been initialized. TF)_headerr5r"s r*r6zMessageBase.headerInitws#  LL  s  cP|jD]}|j|||||y)z< Saves only attachments in the same folder. N)rsave)r" contentIdjson useFileNameraw customPathrSs r*save_attachmentszMessageBase.save_attachmentss/** KJ OOIt[#z J Kr+c  |jS#t$rHg}|jddD]l}|t|jj ds+|t|j|vsF|j |t|jng|_|D]P} |jj |j||1#t$r}|jtjkDr_tjdj|tj||jj t!||nYd}~d}~wt"$r}|jtj$k(r_tjdj|tj||jj t'||nYd}~Kd}~wwxYwd|_ |j*|j,r1|jD]"}|j*D]}|j.|$n#YnxYw|jcYSwxYw)z4 Returns a list of all attachments. FT__attachz!Error processing attachment at {}N)rVr5listDirr= prefixList startswithr;r%NotImplementedErrorr)rATTACHMENT_ERROR_THROWr7errorr9 exceptionr ExceptionATTACHMENT_ERROR_BROKENrrrNrrO)r"attachmentDirsdir_ attachmentDirerSprops r*rzMessageBase.attachmentss! ) %$$ $' %N UD1 FDOO,-88DS12.H"))$s4??/C*DE F !#D !/  %%,,T-A-A$ -VW*33i6V6VV %H%O%OP]%^_((+))001Ft]1[\]!33y7X7XX %H%O%OP]%^_((+))001A$ 1VWX  $'+D # ((,,&*&7&7E $($<$<ED=J==tDEE $$ $O' %ss AI I 26I )+CI  G4A=E I  G4,A=G/)I /G44 I A I  I  I I I c|jS)zL Returns True if the attachment initialization was delayed. )rr[s r*rLzMessageBase.attachmentsDelayeds (((r+c|jS)zG Returns True if the attachments are ready to be used. )rr[s r*rMzMessageBase.attachmentsReadys &&&r+c&|jddS)z6 Returns the bcc field, if it exists. bccrIr[s r*rwzMessageBase.bccs !!%++r+c^ |jS#t$r|jd|_|jrdt|jd|_t j d|j}|'t j d|jd|_|jcYSwxYw)z9 Returns the message body, if it exists. z__substg1.0_1000zutf-8rr1)_bodyr5_getStringStreamrresearchr)r"as r*r zMessageBase.bodys ::  ../ABDJzz*4::w? IIdDJJ/=yy4@&, ::  s BB,+B,c&|jddS)z5 Returns the cc field, if it exists. rryr[s r*rzMessageBase.cc !!$**r+c(|jdddS)zB Returns the compressed RTF stream, if it exists. _compressedRtfz__substg1.0_10090102F _ensureSetr[s r* compressedRtfzMessageBase.compressedRtfs /1GOOr+c2|j|jS)z[ Returns the value of self.__crlf, should you need it for whatever reason. )r rr[s r*crlfzMessageBase.crlfs {{r+c |jS#t$r*|jj|_|jcYSwxYw)z6 Returns the send date, if it exists. )_dater5_proprr[s r*rzMessageBase.dates9  ::  DJ::  s 0AAc |jS#t$r-|jd}|8tj ||_|j |jd<nt jdtj d}|jd|j |jd|j|jd|j|jd |j|jd |j|jd d||_|jcYSwxYw) z[ Returns the message header, if it exists. Otherwise it will generate one. z__substg1.0_007DNrzNHeader is empty or was not found. Header will be generated from other streams.rDateFromToCcz Message-IdzAuthentication-Results) rZr5r| EmailParserparsestrrr7r8 add_headerrrr messageId)r" headerTextrs r*rzMessageBase.headers  <<  ../ABJ%*}55jA '+yy V$ lm$//3!!&$))4!!&$++6!!$0!!$0!!,?!!":DA% << ! s D4EEc |jS#t$r_t|jj|_ |jj dn#t $rYnwxYw|jcYSwxYw)zC Returns a dictionary of the entries in the header Received) _headerDictr5dictr_headerspopKeyErrorr[s r* headerDictzMessageBase.headerDictsp  $## # $#DKK$8$89D    $$Z0  ## #  $s2 -A6AA6 A$!A6#A$$A65A6c(|jdddS)z6 Returns the html body, if it exists. _htmlBodyz__substg1.0_10130102Frr[s r*htmlBodyzMessageBase.htmlBody's {,BEJJr+c&|jddS)zJ Returns the message id that this message is in reply to. _in_reply_toz__substg1.0_1042rr[s r* inReplyTozMessageBase.inReplyTo.s ~/ABBr+cLt|jdjdzS)z@ Returns if this email has been marked as read. 0E070003r/)boolrrEr[s r*isReadzMessageBase.isRead5s% D'' 399A=>>r+c |jS#t$rud}|jr|jd}|||_n;|jrtj d|j d|_|jcYSwxYw)Nz message-idzUHeader found, but "Message-Id" is not included. Will be generated from other streams.z__substg1.0_1035) _messageIdr5r6rZr7r8r|)r" headerResults r*rzMessageBase.messageId<s~ #?? " #L #||L9 '".??$KK wx"&"7"78J"K?? " #s A;B  B cTtjj|jSrK)emailutils parsedaterr[s r* parsedDatezMessageBase.parsedDateLs{{$$TYY//r+c |jS#t$rg}|jD]l}|t|jj ds+|t|j|vsF|j |t|jng|_|D]'}|jj t||)|jcYSwxYw)z3 Returns a list of all recipients. __recip) _recipientsr5rfr=rgrhr;r )r" recipientDirsrp recipientDirs r*rzMessageBase.recipientsPs  $## # $M  EDOO,-88CS12-G!((c$//.B)CD E "D  - G   '' ,(EF G## # $s ACC/A(CCc@tj|jS)zE Returns the decompressed Rtf body from the message. )compressed_rtf decompressrr[s r*rtfBodyzMessageBase.rtfBodygs ((););< |jS#t$r|jr1|jd}| ||_|cYStj d|j d}|j d}d}||}n|}| |d|zdzz }||_|cYSwxYw)z; Returns the message sender, if it exists. fromNzQHeader found, but "sender" is not included. Will be generated from other streams.z__substg1.0_0C1Az__substg1.0_5D01z <>)_senderr5r6rr7r8r|)r"rtextrresults r*rzMessageBase.senderns  <<   #{{62 +#/DL'' op(();>r+c&|jddS)z5 Returns the to field, if it exists. rr/ryr[s r*rzMessageBase.torr+rK)FFFFN)#__name__ __module__ __qualname____doc__rrrjrrIrOrXr6rcpropertyrrLrMrwr rrrrrrrrrrrrrrrrr+r*rrs')JSW$)d+4+K+K$LB FK-%-%^)) '' ,,  ++ PP   0 $ $KK CC ??  # #00$$,== :?? ++r+r)" email.utilsrr_loggingr}rimapclient.imapclientr email.parserrr extract_msgrextract_msg.attachmentrrrextract_msg.compatr osextract_msg.msgr extract_msg.recipientr extract_msg.utilsr r r getLoggerrr7 addHandler NullHandlerrrr+r*rsn  -.!VV(#+FF   8 $%'%%'(B+'B+r+