j]6ddlmZddlmZddlmZddlmZmZm Z m Z ddl m Z m Z ddlmZmZmZGdd ZGd d Ze eefZe eZGd d ZdS)) annotations)aliases)sub)AnyIteratorListTuple)RE_POSSIBLE_ENCODING_INDICATIONTOO_BIG_SEQUENCE) iana_nameis_multi_byte_encoding unicode_rangeceZdZ d1d2dZd3dZd3dZed4dZd5dZd5dZ d6dZ ed5dZ ed7dZ ed8dZ ed8dZed7dZed5d Zed4d!Zed4d"Zed4d#Zed4d$Zed9d%Zed:d'Zed8d(Zed7d)Zed7d*Zd;d CharsetMatchNpayloadbytes | bytearrayguessed_encodingstrmean_mess_ratiofloathas_sig_or_bombool languagesCoherenceMatchesdecoded_payload str | Nonepreemptive_declarationc||_||_||_||_||_d|_g|_d|_d|_d|_ ||_ ||_ dS)N) _payload _encoding_mean_mess_ratio _languages_has_sig_or_bom_unicode_ranges_leaves_mean_coherence_ratio_output_payload_output_encoding_string_preemptive_declaration)selfrrrrrrrs GC:\Projects\DSCons\.venv\Lib\site-packages\charset_normalizer\models.py__init__zCharsetMatch.__init__ se,3 .'6,5%315+- ,/"-1,0#2 3I$$$otherobjectreturnct|ts0t|trt|d|jkSdS|j|jko|j|jkS)NF) isinstancerrr encoding fingerprintr-r1s r.__eq__zCharsetMatch.__eq__(sd%.. %%% @ !..$-??5}.X43CuGX3XXr0ct|tstt|j|jz }t|j|jz }|dkr|dkr|j|jkS|dkrC|dkr=t |jtkr|j|jkS|j |j kS|j|jkS)zQ Implemented to make sorted available upon CharsetMatches items. g{Gzt?g{Gz?) r5r ValueErrorabschaos coherencelenr!r multi_byte_usage)r-r1chaos_differencecoherence_differences r.__lt__zCharsetMatch.__lt__3s%..  "%dj5;&>"?"?&)$.5?*J&K&K e # #(EO3 3  % %*>$*F*F4=!!%555zEK//(5+AA AzEK''r0c~t|j}|dkrdSdtt||z z S)Nrr g?)r?rawr)r-raw_lens r.r@zCharsetMatch.multi_byte_usageIs;dh-- a<<3c#d))nnw.//r0c|j^t|j|jd|_|jr7|jdkr,|jr%|jddkr|jdd|_|jS)Nstrictutf_7rur )r+rr!r"r%r-s r.__str__zCharsetMatch.__str__Qsp < t}dnhGGDL $ 0Ng--L.LOx//#|ABB/ |r0c(d|jd|jdS)Nz)r6r7rJs r.__repr__zCharsetMatch.__repr__as IIIT5EIIIIr0Nonect|tr||kr'td|jd|_|j|dS)Nz;Unable to add instance <{}> as a submatch of a CharsetMatch)r5rr;format __class__r+r'appendr8s r. add_submatchzCharsetMatch.add_submatchdsk%.. %4--MTTO    E"""""r0c|jSN)r"rJs r.r6zCharsetMatch.encodingos ~r0 list[str]cg}tjD]F\}}|j|kr||&|j|kr||G|S)z Encoding name are known by many name, using this could help when searching for IBM855 when it's listed as CP855. )ritemsr6rR)r- also_known_asups r.encoding_aliaseszCharsetMatch.encoding_aliasesssn $& MOO ( (DAq}!!$$Q''''!##$$Q'''r0c|jSrUr%rJs r.bomzCharsetMatch.bom ##r0c|jSrUr^rJs r.byte_order_markzCharsetMatch.byte_order_markr`r0c$d|jDS)z Return the complete list of possible languages found in decoded sequence. Usually not really useful. Returned list may be empty even if 'language' property return something != 'Unknown'. cg|] }|d S)r).0es r. z*CharsetMatch.languages..s...!...r0r$rJs r.rzCharsetMatch.languagess /.do....r0c|jshd|jvrdSddlm}m}t |jr||jn||j}t|dksd|vrdS|dS|jddS)z Most probable language found in decoded sequence. If none were detected or inferred, the property will return "Unknown". asciiEnglishr)encoding_languagesmb_encoding_languagesz Latin BasedUnknown)r$could_be_from_charsetcharset_normalizer.cdrmrnrr6r?)r-rmrnrs r.languagezCharsetMatch.languages  $444 y X W W W W W W W*$-887%%dm444'' 66  9~~""my&@&@ yQ< q!!$$r0c|jSrU)r#rJs r.r=zCharsetMatch.chaoss $$r0c:|jsdS|jddS)Nr rr rirJs r.r>zCharsetMatch.coherences# 3q!!$$r0c4t|jdzdSNd)ndigits)roundr=rJs r. percent_chaoszCharsetMatch.percent_chaossTZ#%q1111r0c4t|jdzdSrv)rzr>rJs r.percent_coherencezCharsetMatch.percent_coherencesT^c)15555r0c|jS)z+ Original untouched bytes. )r!rJs r.rEzCharsetMatch.raws }r0list[CharsetMatch]c|jSrU)r'rJs r.submatchzCharsetMatch.submatchs |r0c2t|jdkSNr)r?r'rJs r. has_submatchzCharsetMatch.has_submatchs4<  1$$r0c|j|jSdt|D}ttd|D|_|jS)Nc,g|]}t|Sre)r)rfchars r.rhz*CharsetMatch.alphabets..s ,W,W,WT]4-@-@,W,W,Wr0ch|]}||Srere)rfrs r. z)CharsetMatch.alphabets..s+L+L+L!!+LA+L+L+Lr0)r&rsortedlist)r-detected_rangess r. alphabetszCharsetMatch.alphabetss^   +' ',W,WSQUYY,W,W,W%d+L+L+L+L+L&M&MNN##r0c6|jgd|jDzS)z The complete list of encoding that output the exact SAME str result and therefore could be the originating encoding. This list does include the encoding available in property 'encoding'. cg|] }|j Sre)r6)rfms r.rhz6CharsetMatch.could_be_from_charset..s"D"D"D!1:"D"D"Dr0)r"r'rJs r.rpz"CharsetMatch.could_be_from_charsets%"D"Dt|"D"D"DDDr0utf_8r6bytesc<j j|kr|_t}jKjdvr0t t fd|ddd}||ddz}||d_jS)z Method to get re-encoded bytes payload using given target encoding. Default to UTF-8. Any errors will be simply ignored by the encoder NOT replaced. N)zutf-8utf8rc&|j|d|d|dt jddS)Nrr _-)stringspanreplacegroupsr r*)rr-s r.z%CharsetMatch.output..shahqvvxx{QVVXXa['@AII 1 !$"788@@cJJr0i r )countr)r*rr,lowerrr encoder))r-r6decoded_stringpatched_headers` r.outputzCharsetMatch.outputs  (D,AX,M,M$,D ! YYN,806688122"%3#5D5)""""0.2G!G#1#8#89#M#MD ##r0intc:tt|S)z] Retrieve a hash fingerprint of the decoded payload, used for deduplication. )hashrrJs r.r7zCharsetMatch.fingerprints CIIr0)NN)rrrrrrrrrrrrrr)r1r2r3r)r3rr3r)r1rr3rN)r3rVr3r)r3r)r3r)r)r6rr3rr3r)__name__ __module__ __qualname__r/r9rCpropertyr@rKrMrSr6r\r_rbrrrr=r>r{r}rErrrrprr7rer0r.rr s'+-1JJJJJ8 Y Y Y Y((((,000X0 JJJJ # # # #X   X $$$X$$$$X$///X/%%%X%6%%%X%%%%X% 222X2666X6X X%%%X%$$$X$EEEXE$$$$$:Xr0rc\eZdZdZdddZddZdd ZddZddZddZ ddZ ddZ ddZ dS) CharsetMatchesz Container with every CharsetMatch items ordered by default from most probable to the less one. Act like a list(iterable) but does not implements all related methods. Nresultslist[CharsetMatch] | NonecD|rt|ng|_d|_dSNT)r_results _is_sorted)r-rs r.r/zCharsetMatches.__init__ s#?F,NF7OOOB $r0r3rNcX|js"|jd|_dSdSr)rrsortrJs r._ensure_sortedzCharsetMatches._ensure_sorteds5 # M   "DOOO # #r0Iterator[CharsetMatch]c#LK||jEd{VdSrU)rrrJs r.__iter__zCharsetMatches.__iter__s8 =         r0item int | strrct|tr!||j|St|tr't |d}|jD]}||jvr|cSt)z Retrieve a single item either by its position or encoding name (alias may be used here). Raise KeyError upon invalid index or encoding not present in results. F)r5rrrrr rpKeyError)r-rresults r. __getitem__zCharsetMatches.__getitem__s dC  '    ! ! !=& & dC  "T5))D- " "6777!MMM8r0rc*t|jSrUr?rrJs r.__len__zCharsetMatches.__len__'s4=!!!r0rc2t|jdkSrrrJs r.__bool__zCharsetMatches.__bool__*s4=!!A%%r0ct|ts4tdt |jt |jtkrB|j D]:}|j |j kr(|j |j kr| |dS;|j |d|_dS)z~ Insert a single match. Will be inserted accordingly to preserve sort. Can be inserted as a submatch. z-Cannot append instance '{}' to CharsetMatchesNF)r5rr;rPrrQr?rEr rr7r=rSrRr)r-rmatchs r.rRzCharsetMatches.append-s $ -- ?FF''  tx==+ + +  $(888U[DJ=V=V&&t,,,FF T"""r0CharsetMatch | NonecV|jsdS||jdS)zQ Simply return the first match. Strict equivalent to matches[0]. Nr)rrrJs r.bestzCharsetMatches.bestAs2} 4 }Qr0c*|S)zP Redundant method, call the method best(). Kept for BC reasons. )rrJs r.firstzCharsetMatches.firstJsyy{{r0rU)rr)r3rN)r3r)rrr3rrr)rrr3rN)r3r) rrr__doc__r/rrrrrrRrrrer0r.rrs %%%%%#### !!!!    """"&&&&    (    r0rc6eZdZddZeddZddZdS)CliDetectionResultpathrr6rr\rValternative_encodingsrrrrrr=rr> unicode_path is_preferredc ||_| |_||_||_||_||_||_||_||_| |_ | |_ dSrU) rrr6r\rrrrrr=r>r) r-rr6r\rrrrrr=r>rrs r.r/zCliDetectionResult.__init__Vs\ (4$, +;0E"% $-$2! )".r0r3dict[str, Any]c |j|j|j|j|j|j|j|j|j|j |j d S)N rr6r\rrrrrr=r>rrrrJs r.__dict__zCliDetectionResult.__dict__psOI $ 5%)%? "1Z - -   r0c4ddlm}||jddS)Nr)dumpsT) ensure_asciiindent)jsonrr)r-rs r.to_jsonzCliDetectionResult.to_jsons-uT]a@@@@r0N)rrr6rr\rVrrVrrrrrVrrr=rr>rrrrr)r3rr)rrrr/rrrrer0r.rrUsb////4    X  AAAAAAr0rN) __future__rencodings.aliasesrrertypingrrrr constantr r utilsr rrrrrrCoherenceMatchrrrer0r.rsN""""""%%%%%%------------GGGGGGGGCCCCCCCCCCwwwwwwwwtIIIIIIIIXsEz"'.A.A.A.A.A.A.A.A.A.Ar0