a [XhW&@s(dZddlZddlmZddlmZdZZdZiZ e dj Z e dj Zd ZiZzed ZWn eyeejd ZYn0eeD]BZejd krejd dZeee<eddZeevreee<qddeDZddZGdddZGddde Z!GdddZ"GdddZ#dS)zY Pyphen ====== Pure Python module to hyphenate text, inspired by Ruby's Text::Hyphen. N) resources)Pathz0.17.2) LANGUAGESPyphenlanguage_fallbackz\^{2}([0-9a-f]{2})z (\d?)(\D?))%#Z LEFTHYPHENMINZRIGHTHYPHENMINZCOMPOUNDLEFTHYPHENMINZCOMPOUNDRIGHTHYPHENMINzpyphen.dictionaries dictionariesz.dic_cCsi|]}||qS)lower).0namer r T/var/www/viveiro_mudafortebrasil/venv/lib/python3.9/site-packages/pyphen/__init__.py ,rcCsB|ddd}|r>d|}|tvr4t|S|qdS)a Get a fallback language available in our dictionaries. http://www.unicode.org/reports/tr35/#Locale_Inheritance We use the normal truncation inheritance. This function needs aliases including scripts for languages with multiple regions available. -r N)replacersplitjoinLANGUAGES_LOWERCASEpop)languagepartsr r rr/s  rc@s eZdZdZddZddZdS)AlternativeParserzParser of nonstandard hyphen pattern alternative. The instance returns a special int with data about the current position in the pattern when called with an odd value. cCsL|d}|d|_t|d|_t|d|_|drH|jd7_dS)N,r.)rchangeintindexcut startswith)selfpattern alternativer r r__init__Gs    zAlternativeParser.__init__cCs<|jd8_t|}|d@r4t||j|j|jfS|SdS)Nr)r#r"DataIntr!r$)r&valuer r r__call__Os zAlternativeParser.__call__N)__name__ __module__ __qualname____doc__r)r,r r r rr@src@seZdZdZdddZdS)r*zE``int`` with some other data can be stuck to in a ``data`` attribute.NcCs.t||}|r$t|tr$|j|_n||_|S)zCreate a new ``DataInt``. Call with ``reference=dataint_object`` to use the data from another ``DataInt``. )r"__new__ isinstancer*data)clsr+r3 referenceobjr r rr1Zs   zDataInt.__new__)NN)r-r.r/r0r1r r r rr*Xsr*c@s eZdZdZddZddZdS)HyphDictzHyphenation patterns.c sdi|_|d}|}Wdn1s20Y|dkrLd}||dddD]}|}|rd|t rqdt dd|}d |vrd |vr|d d\}}t ||nt t fd d t|D\}}t|d krqdd t|}} ||s |d7}q|| ds&| d8} q |||| f|jd|<qdi|_tdd|jD|_dS)zhRead a ``hyph_*.dic`` and parse its patterns. :param path: Path of hyph_*.dic to read rbNzmicrosoft-cp1251cp1251 rcSstt|ddS)Nr)chrr"group)matchr r rrz#HyphDict.__init__../=cs g|]\}}||pdfqS)0r )ristringfactoryr r sz%HyphDict.__init__..rcss|]}t|VqdS)N)len)rkeyr r r rz$HyphDict.__init__..)patternsopenreadlinedecoder read_textrstripr%ignored parse_hexrr"zipparsemaxrIrcachemaxlen) r&pathfdencodingr'r(tagsvaluesstartendr rErr)ls: *        zHyphDict.__init__c Cs|}|j|}|durd|d}dgt|d}tt|dD]}t||jt|d}t|d|D]T}|j|||}|sqt|\} } t|| || t| } t t | || || <qtqJddt |D|j|<}|S)aGet a list of positions where the word can be hyphenated. :param word: unicode string of the word to hyphenate E.g. for the dutch word 'lettergrepen' this method returns ``[3, 6, 9]``. Each position is a ``DataInt`` with a data attribute. If the data attribute is not ``None``, it contains a tuple with information about nonstandard hyphenation at that point: ``(change, index, cut)``. change a string like ``'ff=f'``, that describes how hyphenation should take place. index where to substitute the change, counting from the current point cut how many characters to remove while substituting the nonstandard hyphenation Nr rrcSs(g|] \}}|drt|d|dqS)rr)r5)r*)rrCr5r r rrGsz&HyphDict.positions..) rrWgetrIrangeminrXrLslicemaprV enumerate) r&wordZpointsZ pointed_wordZ referencesrCstopjr'offsetr]Zslice_r r r positionss$  zHyphDict.positionsN)r-r.r/r0r)rjr r r rr7is1r7c@sBeZdZdZdddZddZd d Zdd d ZdddZeZ dS)rzEHyphenation class, with methods to hyphenate strings in various ways.NrTcCsL|||_|_|rt|n tt|}|r2|tvr>t|t|<t||_dS)aCreate an hyphenation instance for given lang or filename. :param filename: filename or Path of hyph_*.dic to read :param lang: lang of the included dict to use if no filename is given :param left: minimum number of characters of the first syllabe :param right: minimum number of characters of the last syllabe :param cache: if ``True``, use cached copy of the hyphenation patterns N)leftrightrrrhdcacher7hd)r&filenamelangrkrlrWrYr r rr)s   zPyphen.__init__cs*t|jfddj|DS)zGet a list of positions where the word can be hyphenated. :param word: unicode string of the word to hyphenate See also ``HyphDict.positions``. The points that are too far to the left or right are removed. cs*g|]"}j|krkrnq|qSr )rk)rrCrlr&r rrGrz$Pyphen.positions..)rIrlrnrj)r&rfr rqrrjs zPyphen.positionsccst||D]~}|jrr|j\}}}||7}|r<|}|d\}}|d||||||dfVq|d|||dfVqdS)zIterate over all hyphenation possibilities, the longest first. :param word: unicode string of the word to hyphenate rAN)reversedrjr3isupperupperr)r&rfpositionr!r#r$Zc1c2r r riterates (zPyphen.iteratercCs@|t|8}||D]$\}}t||kr|||fSqdS)aGet the longest possible first part and the last part of a word. :param word: unicode string of the word to hyphenate :param width: maximum length of the first part :param hyphen: unicode string used as hyphen character The first part has the hyphen already attached. Returns ``None`` if there is no hyphenation point before ``width``, or if the word could not be hyphenated. N)rIrw)r&rfwidthhyphenZw1Zw2r r rwraps  z Pyphen.wrapcCsvt|}t||D]T}|jr^|j\}}}||7}|rD|}|d|||||<q|||qd|S)aGet the word as a string with all the possible hyphens inserted. :param word: unicode string of the word to hyphenate :param hyphen: unicode string used as hyphen character E.g. for the dutch word ``'lettergrepen'``, this method returns the unicode string ``'let-ter-gre-pen'``. The hyphen string to use can be given as the second parameter, that defaults to ``'-'``. rArH) listrrrjr3rsrtrinsertr)r&rfrylettersrur!r#r$r r rinserteds  zPyphen.inserted)NNrrT)r)r) r-r.r/r0r)rjrwrzr~r,r r r rrs    r)$r0re importlibrpathlibrVERSION __version____all__rmcompilesubrSfindallrUrRrfilesr TypeError__file__parentsortediterdirrYsuffixrrZ short_namerrrr"r*r7rr r r rs6        d