o ѹg.@srddlmZddlZddlZddlZddlmZddlmZddl m Z ddl m Z ddl mZddlmZd d lmZmZmZmZmZmZe ed dgddZe ed dhddZe ed diddZe ed dgddZe ed dgddZe ed dgddZe ed dgddZe ed dgdd Ze ed dgd!d"Z e ed dgd#d$Z!e ed dgd%d&Z"e ed dgd'd(Z#e ed dgd)d*Z$e ed dgd+d,Z%e ed dgd-d.Z&e ed dgd/d0Z'e e(ed djd2d3Z)e ed dgd4d5Z*dkdld;d<Z+e d=d dmd?d@Z,dndBdCZ-dodEdFZ.dpdqdJdKZ/drdOdPZ0dsdQdRZ1dSej2dTfdtdXdYZ3 dudvdedfZ4dS)w) annotationsN)IncrementalDecoder)aliases) lru_cache)findall) Generator)MultibyteIncrementalDecoder)ENCODING_MARKSIANA_SUPPORTED_SIMILARRE_POSSIBLE_ENCODING_INDICATIONUNICODE_RANGES_COMBINEDUNICODE_SECONDARY_RANGE_KEYWORDUTF8_MAXIMAL_ALLOCATION)maxsize characterstrreturnboolcCsdzt|}Wn tyYdSwd|vp1d|vp1d|vp1d|vp1d|vp1d|vp1d|vp1d |vS) NFz WITH GRAVEz WITH ACUTEz WITH CEDILLAzWITH DIAERESISzWITH CIRCUMFLEXz WITH TILDEz WITH MACRONzWITH RING ABOVE unicodedataname ValueErrorr descriptionr\/Users/admin/.pyenv/versions/3.10.0/lib/python3.10/site-packages/charset_normalizer/utils.pyis_accentuateds( rcCs.t|}|s |S|d}tt|ddS)N r)r decompositionsplitchrint)rZ decomposedcodesrrr remove_accent,s  r% str | NonecCs.t|}tD] \}}||vr|SqdS)zK Retrieve the Unicode range official name from a single character. N)ordr items)rZ character_ord range_nameZ ord_rangerrr unicode_range7s r*cC*z t|}Wd|vStyYdSw)NFZLATINrrrrris_latinEs   r,cCs2t|}d|vr dSt|}|durdSd|vS)NPTFZ Punctuationrcategoryr*rcharacter_categorycharacter_rangerrris_punctuationNs r3cCsBt|}d|vs d|vrdSt|}|durdSd|vo |dkS)NSNTFZFormsZLor.r0rrr is_symbol]s r6cCs$t|}|dur dSd|vpd|vS)NFZ EmoticonsZ Pictographs)r*)rr2rrr is_emoticonlsr7cCs.|s|dvr dSt|}d|vp|dvS)N>>u｜+ZPoZPdZPc)isspacerr/)rr1rrr is_separatorvs r=cCs||kSN)islowerisupperrrrris_case_variablesrBcCr+)NFCJKrrZcharacter_namerrris_cjk   rEcCr+)NFZHIRAGANArrDrrr is_hiraganarFrGcCr+)NFZKATAKANArrDrrr is_katakanarFrHcCr+)NFZHANGULrrDrrr is_hangulrFrIcCr+)NFZTHAIrrDrrris_thairFrJcCr+)NFARABICrrDrrr is_arabicrFrLcCs4zt|}Wn tyYdSwd|vod|vS)NFrKz ISOLATED FORMrrDrrris_arabic_isolated_forms  rMr)cstfddtDS)Nc3s|]}|vVqdSr>r).0keywordr)rr sz-is_unicode_range_secondary..)anyrrPrrPris_unicode_range_secondarysrScCs(|duo|duo|dko|dkS)NFu)r< isprintablerArrris_unprintables  rV sequencebytes search_zoner#cCst|tstt|}tt|dt||jddd}t|dkr$dS|D]'}| dd}t D]\}}||krB|S||krL|Sq4q&dS)zW Extract using ASCII-only decoder any specified encoding in the first n-bytes. Nasciiignoreerrorsr-_) isinstancerY TypeErrorlenrr mindecodelowerreplacerr()rXrZZseq_lenresultsZspecified_encodingencoding_alias encoding_ianarrrany_specified_encodings&    rkrcCs |dvpttd|jtS)zQ Verify is a specific encoding is a multi byte one based on it IANA name > utf_16 utf_32_beutf_7utf_32utf_8 utf_32_le utf_16_be utf_16_leZ utf_8_sig encodings.N) issubclass importlib import_modulerr)rrrris_multi_byte_encodings  rytuple[str | None, bytes]cCsJtD] }t|}t|tr|g}|D]}||r!||fSqqdS)z9 Identify and extract SIG/BOM in given sequence. )NN)r rarY startswith)rX iana_encodingZmarksmarkrrridentify_sig_or_boms  rr}cCs|dvS)N>rmrpr)r}rrrshould_strip_sig_or_bom"srTcp_namestrictcCsN|dd}tD]\}}|||fvr|Sq |r%td|d|S)zIReturns the Python normalized encoding name (Not the IANA official name).r_r`zUnable to retrieve IANA for ''N)rfrgrr(r)rrrirjrrr iana_name&s r iana_name_a iana_name_bfloatc Cst|st|r dStd|j}td|j}|dd}|dd}d}tdD]}t|g}||||krA|d7}q,|dS) Ngrur\r]rr )ryrwrxrrangerYre) rrZ decoder_aZ decoder_bZid_aZid_bZcharacter_match_countiZ to_be_decodedrrr cp_similarity7s    rcCs|tvo |t|vS)z Determine if two code page are at least 80% similar. IANA_SUPPORTED_SIMILAR dict was generated using the function cp_similarity. N)r )rrrrr is_cp_similarKs rZcharset_normalizerz)%(asctime)s | %(levelname)s | %(message)slevel format_stringNonecCs:t|}||t}|t|||dSr>)logging getLoggersetLevel StreamHandler setFormatter Formatter addHandler)rrrloggerhandlerrrrset_logging_handlerVs  r sequencesrjoffsetsr chunk_sizebom_or_sig_availablestrip_sig_or_bom sig_payloadis_multi_byte_decoderdecoded_payloadGenerator[str, None, None]c cs&|r|dur|D]} || | |} | sdS| Vq dS|D]p} | |} | t|dkr/q || | |} |rA|durA|| } | j||rHdndd} |r| dkrt|d} |r| d| |vrt| | dd D]#}||| } |r{|dur{|| } | j|dd} | d| |vrqqi| Vq dS) NFr\rr]rr)rcrerdr)rrjrrrrrrrrchunkZ chunk_endZ cut_sequenceZchunk_partial_size_chkjrrrcut_sequence_chunkscsF       r)rrrr)rrrr)rrrr&)r)rrr)rW)rXrYrZr#rr&)rrrr)rXrYrrz)r}rrr)T)rrrrrr)rrrrrr)rrrrrr)rrrr#rrrrr>)rrYrjrrrrr#rrrrrrYrrrr&rr)5 __future__rrwrrcodecsrZencodings.aliasesr functoolsrrertypingrZ_multibytecodecrZconstantr r r r rrrr%r*r,r3r6r7r=rBrErGrHrIrJrLrMrcrSrVrkryrrrrrINFOrrrrrrsz