o agx @sddlZddlZ edZiddddddddd d d d d dddddddddddddddddddd d!d!d"dd#d$d%d&d'd(d) Zed*ejZed+ejZd0d.d/Z dS)1N($aftncnaftnsepaftnsepc annotation atnauthoratndateatnicnatnid atnparentatnrefatntimeatrfend atrfstartauthor backgroundbkmkend bkmkstartblipuidbuptimcategorycolorschememappingcolortblcommentcompanycreatim datafield datastoredefchpdefpapdodoccommdocvar dptxbxtextebcendebcstart factoidnamefaltfchars ffdeftext ffentrymcr ffexitmcrffformat ffhelptextfflffname ffstattextfilefiletblfldinstfldtypefnamefontembfontfilefonttblfooterfooterffooterlfooterrfootnote formfieldftncnftnsepftnsepcg generatorgridtblheaderheaderfheaderlheaderrhlhlfr hlinkbasehllochlsrchsvhtmltaginfokeycodekeywords latentstyleslchars levelnumbers leveltextlfolevellinkvallist listlevellistname listoverridelistoverridetable listpicture liststylename listtablelisttextlsdlockedexceptmaccmaccPr mailmergemalnmalnScrmanagermargPrmbarmbarPrmbaseJcmbegChr mborderBox mborderBoxPrmboxmboxPrmchrmcountmctrlPrmdmdegmdegHidemdenmdiffmdPrmemendChrmeqArrmeqArrPrmfmfNamemfPrmfuncmfuncPr mgroupChr mgroupChrPrmgrowmhideBot mhideLeft mhideRightmhideTopmhtmltagmlimmlimlocmlimlow mlimlowPrmlimupp mlimuppPrmmmmaddfieldnamemmath mmathPictmmathPrmmaxdistmmcmmcJc mmconnectstrmmconnectstrdatammcPrmmcs mmdatasourcemmheadersource mmmailsubjectmmodso mmodsofiltermmodsofldmpdatammodsomappedname mmodsonamemmodsorecipdata mmodsosort mmodsosrc mmodsotable mmodsoudl mmodsoudldatammodsouniquetagmmPrmmquerymmrmnarymnaryPrmnoBreakmnummobjDistmoMath moMathPara moMathParaPrmopEmumphantmphantPrmplcHidemposmrmradmradPrmrPrmsepChrmshowmshpmsPremsPrePrmsSubmsSubPrmsSubSup msSubSupPrmsSupmsSupPr mstrikeBLTRmstrikeH mstrikeTLBRmstrikeVmsubmsubHidemsupmsupHidemtranspmtypemvertJcmvfmfmvfmlmvtofmvtolmzeroAsc mzeroDescmzeroWidnesttablepropsnextfile nonesttablesobjaliasobjclassobjdataobjectobjnameobjsectobjtime oldcprops oldpprops oldsprops oldtpropsoleclsidoperatorpanosepassword passwordhashpgppgptblpicproppictpnpnseclvlpntextpntxtapntxtbprintimprivatepropnameprotend protstart protusertblpxeresultrevtblrevtimrsidtblrxeshpshpgrpshpinstshppictshprsltshptxtsnsp staticval stylesheetsubjectsvsvbtctemplate themedatatitletxeudupr userprops wgrffmtfilter windowcaptionwritereservationwritereservhashxexform xmlattrname xmlattrvaluexmlclosexmlnamexmlnstblxmlopenpar sectz pagelinetab emdashu—endashu–emspaceu enspaceu qmspaceu bulletu•lquoteu‘rquoteu’ ldblquoteu“ rdblquoteu”rowcell|  {}\­u‑) nestcell~r'r;r<r=r>-_zO\\([a-z]{1,32})(-?\d{1,10})?[ ]?|\\'([0-9a-f]{2})|\\([^a-z])|([{}])|[\r\n]+|(.)zP(\{\\field\{\s*\\\*\\fldinst\{.*HYPERLINK\s(\".*\")\}{2}\s*\{.*?\s+(.*?)\}{2,3})cp1252strictc Csttd|}g}d}d}d}d}d}t|D]} | \} } } } }}|r6| s6|t|j||d7}d}|rXd}|dkrF| ||fq|d krW|rS| \}}qd}d }q| rpd}| t vri|sh|t | 7}q| d krod }q| rd}| t vr{d }n| d krd | }zt |Wn tyd}Ynw|rq| t vr|t | 7}q| dkrt| }q| dkr| dur|}qt| }|dkr|d7}|t|7}|}q| r|dkr|d8}q|st| d}|s| }q|| 7}q|r|dkr|d8}q|s||7}q|S)aW Converts the rtf text to plain text. Parameters ---------- text : str The rtf text encoding : str Input encoding which is ignored if the rtf file contains an explicit codepage directive, as it is typically the case. Defaults to `cp1252` encoding as it the most commonly used. errors : str How to handle encoding errors. Default is "strict", which throws an error. Another option is "ignore" which, as the name says, ignores encoding errors. Returns ------- str the converted rtf text as a python unicode string z\1(\2)FrN)encodingerrorsr<r=T*ansicpgcputf8ucui)resub HYPERLINKSPATTERNfinditergroupsbytesfromhexdecodeappendpop specialchars destinationscodecslookup LookupErrorintchr)textrHrIstack ignorableucskipcurskiphexesoutmatchwordarg_hexcharbracetcharcrrk/var/www/eduai.edurigo.com/doc_train/edurigo_ai/Puru/venv/lib/python3.10/site-packages/striprtf/striprtf.py rtf_to_textbs        rt)rDrE) rQr^ frozensetr]r\compile IGNORECASErTrSrtrrrrrrrsst.