jE'UddlZddlZddlZddlZddlZddlZddlmZddlm Z ddl m Z m Z m Z ddlmZmZmZmZmZmZmZddlmZmZmZmZddlmZddlmZmZdd l m!Z!ej"d krdd l#m$Z$ndd lm$Z$dd l%m&Z&dd l'm(Z(m)Z)ddl*m+Z+m,Z,m-Z-m.Z.m/Z/ddl0m1Z1m2Z2m3Z3m4Z4m5Z5m6Z6m7Z7ddl8m9Z9m:Z:m;Z;mm?Z?m@Z@ddlAmBZBddlCmDZDddlEmFZFddgZGe$eeHeId<dZJeHeId<deHdee1fdZKdeHdeeHeHffdZLdedee1fdZM d@ded eeHd!eeHde2fd"ZNd#eHdeeHfd$ZO d@d%ed&eeHdeeeHeHffd'ZPd(ZQd)ee4e6fd*ejRdee1e1ffd+ZS dAd eeHd,eeeeTe fdeeeHeffd-ZUe7e@e?jVe&dddddd.d/d0ddd.dddd1gd.fd eeHd,eeeeTe fd2eeHd3eHd4eeHd5eWd6eeXd7eWd8eeHd!eeHd9eWd:eed;eeXdeWdee1f"d?ZYdS)BN)Message)partial)NamedTemporaryFileSpooledTemporaryFileTemporaryDirectory)IOCallableDictListOptionalTupleUnion)COMMON_ENCODINGSformat_encoding_str read_txt_filevalidate_encoding)logger)convert_to_bytes exactly_one)apply_lang_metadata))Final)add_chunking_strategy)clean_extra_whitespacereplace_mime_encodings)extract_datetimetzextract_email_addressextract_ip_addressextract_ip_address_nameextract_mapi_id)ElementElementMetadataImage NarrativeTextTextTitleprocess_metadata)MetaData ReceivedInfo RecipientSenderSubject)FileTypeadd_metadata_with_filetype)EMAIL_DATETIMETZ_PATTERN_RE)partition_html)partition_text text/html text/plainVALID_CONTENT_SOURCESemailDETECTION_ORIGINdatareturnct|}t|}t|}t|}g}|r<|r:t ||D])\}}|t ||*|r*|t d|d|r2|t dt|||S)Nnametextmapi_idrreceived_datetimetz)r<r= datestamp)r rr!rzipappendr*str)r8ip_address_names ip_addressesr> datetimetzelementsr<ips f/lsinfo/ai/hellotax_ai/base_platform/venv/lib/python3.11/site-packages/unstructured/partition/email.py_parse_received_datarJ@s.t44%d++Ld##G#D))J H>L>,l;; > >HD" OOLd<<< = = = =G )'!*EEEFFF  *__$       Oct|}d}tj||d}||dfS)Nz)<[a-z0-9\.\-+_]+@[a-z0-9\.\-+_]+\.[a-z]+>r)rresplitlowertitlestrip)r8 email_addressPATTERNr<s rI_parse_email_addressrTWsZ)$//M9G 8GTZZ\\ * *1 - 3 3 5 5 ; ; = =D q! !!rKmsgcg}|D]3}|ddkrFt|d}|t|d|dU|ddkrFt|d}|t |d|d|ddkr*|t |d|ddkr|t |dz }|t|d|d5|S) NrTor;Fromr-)r=Received) raw_itemsrTrBr+r,r-rJr))rUrGitemr=s rIpartition_email_headerr]`s< H  B B 7d??'Q00D OOI47aAAA B B B B !W  'Q00D OOFQd1g>>> ? ? ? ? !W ! ! OOGa111 2 2 2 2 !W " " ,T!W55 5HH OOH$q'Q@@@ A A A A OrKfilenamemetadata_last_modifiedct|}|d}|t|}|d}|d|dD}|d}|d|dD}t |||d|p|| }t |_|S) zKCreates an ElementMetadata object from the header information in the email.DateNrYc6g|]}|SrQ).0senders rI z(build_email_metadata..s GGGV\\^^GGGrK,rWc6g|]}|Srcrd)re recipients rIrgz(build_email_metadata..s"III9??$$IIIrKr-)sent_to sent_fromsubject last_modifiedr^)dictr[getconvert_to_iso_8601rNr#r7detection_origin)rUr^r_ header_dict email_daterlrkelement_metadatas rIbuild_email_metadatarvss s}}''K((J(44 ''IGG)//#2F2FGGG ood##GIIgmmC6H6HIII& **,:  )9% rKtimec(t|}tj|}|tj|ddS|\}}|||}t j|d}|S)z?Converts the datetime from the email output to ISO-8601 format.Nz; did not match RFC-2822 format. Unable to extract the time.z%a, %d %b %Y %H:%M:%S %z) rr0searchrwarningspandatetimestrptime isoformat)rw cleaned_time regex_matchstartend dt_stringdatetime_objects rIrqrqs)$//L-4\BBK P P P   t!!##JE3U3Y'I'00.s!DDDd+D11DDDrK) attachmentinline=rX"Tdecodepayloadr^/wbF)modedirdelete)walkrNrOrreplace get_payloadrB enumerateopenwriterospathbasenamer<) rrlist_attachmentspartcdispattachment_infor\keyvalueidxrr^fs rIextract_attachment_infors !;!; D ( (./55c::EDDeDDDE O  ::<<#;;;!ZZQ// U,S[[b-A-ABB.u}}S"/E/EFF?U@@ 6s ; ;<<*.)9)9)9)F)FOI &  # #O 4 4 4#,-=#>#> ; ;Z;!Z//#-#3j6L#L!(D11;QGGJy$9:::;;;;;;;;;;;;;;; 0!% *#(;@B@P@PQRQW@X@X,S1*=GGJy$9::: ;;;;;;;;;;;;;;; s%E00E4 7E4 A G##G' *G' c^tjd}||jS)Nz \[image: .+\])rMcompileryr=)elementrSs rIhas_embedded_imagers%j))G >>', ' ''rKrindicescR||}}|j||}t|dd}|jd|ddzdzd|_t |ddd|fS) N:rXz[image: ]rr6)r=rr)rrr=rrNrr$)rrrrimage_raw_info image_infos rIfind_embedded_imagers'++--3E\%),N'(<( dddn #1swxYwYtitk|(||}|S)a0Partitions an .eml documents into its constituent elements. Parameters ---------- filename A string defining the target filename path. file A file-like object using "r" mode --> open(filename, "r"). text The string representation of the .eml document. content_source default: "text/html" other: "text/plain" encoding The encoding method used to decode the text input. If None, utf-8 will be used. max_partition The maximum number of characters to include in a partition. If None is passed, no maximum is applied. Only applies if processing the text/plain content. metadata_filename The filename to use for the metadata. metadata_last_modified The last modified date for the document. process_attachments If True, partition_email will process email attachments in addition to processing the content of the email itself. attachment_partitioner The partitioning function to use to process attachments. min_partition The minimum number of characters to include in a partition. Only applies if processing the text/plain content. languages User defined value for `metadata.languages` if provided. Otherwise language is detected using naive Bayesian filter via `langdetect`. Multiple languages indicates text could be in either language. Additional Parameters: detect_language_per_element Detect language per element instead of at the document level. zE is not a valid value for content_source. Valid content sources are: Nr)r^rr=zutf-8)r^)r^r)r)rrF encryptedTr=zcontent-transfer-encodingbase64rzGEncrypted email detected. Partition function will return an empty list.r3z= r6)r=rrrrr)rr4)r=rrrrrrrrX)r^r_z@Specify the attachment_partitioner kwarg to process attachments.)r^r_rr)rGrr)6r5rrQrrrr6message_from_stringrCrget_content_dispositionget_content_typeendswithget_content_maintyperprrUnicodeDecodeError UnicodeErrorrrzrNjoinr1 isinstancer&rrapplyrrrBr2rrr%r'rinsertr]rvcopydeepcopymetadatarrrlistdirrr^file_directoryattached_to_filenamelistr)/r^rr=rrrrrrr_rrrrrrkwargsdetected_encodingextracted_encodingrU file_text_text is_encrypted content_mapr content_typecontentrG list_contentr_replace_mime_encodingscommon_encodingsx_xencrrr clean_elementheader all_elementsrtmpdirattached_files attached_fileattached_filenameattached_elementss/ rIpartition_emailrsv222 B B*? B B    DJJLLB..t.H. 48888"-x"@"@"@C  7 2  +8!!,,, ( y+I66CC "-4"8"8"8C  7 2  +8dX+V+V+V ( y+I66CC YY'.. %$L"$K ;;  ' ' ) ) 5 ,,..   - - L  % % ' '6 1 14d;;xGG ?,0,<,,> L))) ?)-(8(8(:(:K % %oonb11G H=  U    8  ; & &}}U++ '',''!"/d$      % %G'4(( %*1*%+++'%MM"9::::*L9%%%')$-88033>>,33B777/ % %%6= 6),7773$MM*ABBB!E 2LA%%%$H%!% % %6 < ' '!''d"$   "(++11 W$W-- w 6 7 7 1W 1(;GW(M(M %J )HSM OOC!GZ 0 0 0F-',,H$L# ".h5H  33=221  ! ! 1V #C 0 0 0Z//N!/ 1 1 $&GLL$G$G!)1$Z%;$:.+A"/"/ %%%! 111G0=G$-6:G$3rs  !!!!!!QQQQQQQQQQCCCCCCCCCCCCCCCCCC '&&&&&<;;;;;f'''''''777777UUUUUUUURQQQQQQQAAAAAA666666666666+6 *EuT#Y'EEE#stG}."s"uS#X""""DM,-1 sm%SM < 'c 'hsm ' ' ' '$!%)) ) ) $sCx.))))X((( J =%' ( J X J 7G  J J J J#=A##sm# 5E$889 :# 8C=' !"####2HL))"=A%"!#'!'+,0 %15#$'+&,X(-!mmsmm 5E$889 :m 3-m m sm m  mC=mm }m%SMmm%X.mC=m }mS "m "&!m$ ']%mmm*)mmmrK