3 (h*@s@dZddlmZddlZddlZddlZyeWnek rHeZYnXdZ iZ ej dj Z ej djZzFy(ddlZejjejd d d d d fZWnek rfZYnXWdeejjejd d d ejjejjed f7ZXed deDZddZGdddeZGdddeZGdddeZ GdddeZ!dS)zY Pyphen ====== Pure Python module to hyphenate text, inspired by Ruby's Text::Hyphen. )unicode_literalsNPyphen LANGUAGESlanguage_fallbackz\^{2}([0-9a-f]{2})z (\d?)(\D?)ZpyphenZshareZ dictionariesccsN|]F}tjj|rtj|D]*}|jdr|ddtjj||fVqqdS)z.dicN)ospathisdirlistdirendswithjoin).0Zdictionaries_rootfilenamer*/tmp/pip-install-q3hcpn_q/Pyphen/pyphen.py @s rcCs>|jddjd}x&|r8dj|}|tkr.|S|jqWdS)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. -_N)replacesplitrrpop)languagepartsrrrrGs  c@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|jd}|d|_t|d|_t|d|_|jdrH|jd7_dS)N,r.)rchangeintindexcut startswith)selfpattern alternativerrr__init___s    zAlternativeParser.__init__cCs<|jd8_t|}|d@r4t||j|j|jfS|SdS)Nr)r#r"DataIntr!r$)r&valuerrr__call__gs zAlternativeParser.__call__N)__name__ __module__ __qualname____doc__r)r,rrrrrXsrc@seZdZdZdddZdS)r*zE``int`` with some other data can be stuck to in a ``data`` attribute.NcCs.tj||}|r$t|tr$|j|_n||_|S)zCreate a new ``DataInt``. Call with ``reference=dataint_object`` to use the data from another ``DataInt``. )r"__new__ isinstancer*data)clsr+r3 referenceobjrrrr1rs   zDataInt.__new__)NN)r-r.r/r0r1rrrrr*psr*c@s eZdZdZddZddZdS)HyphDictzHyphenation patterns.c sVi|_t|d}|jjjd}|jdkr6d}x|D]}|j|j}| s<|jds<|jdrjqsz#HyphDict.__init__../rcs g|]\}}||pdfqS)0r)ristring)factoryrr sz%HyphDict.__init__..rrNcss|]}t|VqdS)N)len)rkeyrrrrsz$HyphDict.__init__..)patternsopenreadlinestripdecodelowerr% parse_hexrrr"zipparsemaxrHrcachemaxlen) r&rstreamcharsetr'r(tagsvaluesstartendr)rFrr)s<           *zHyphDict.__init__c Cs|j}|jj|}|dkrd|}dgt|d}xtt|dD]~}xxt|dt||jt|dD]R}|jj|||}|rp|\}} t||||t| } t t | || || <qpWqHWddt |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 Nz.%s.rrcSs(g|] \}}|drt|d|dqS)rr)r5)r*)rrDr5rrrrGsz&HyphDict.positions..) rOrTgetrHrangeminrUrJslicemaprS enumerate) r&wordZpointsZ pointed_wordZ referencesrDjr'offsetrYZslice_rrr positionss$ $  zHyphDict.positionsN)r-r.r/r0r)rerrrrr7s3r7c@sBeZdZdZdddZddZd d Zdd d ZdddZeZ dS)rzEHyphenation class, with methods to hyphenate strings in various ways.NrTcCsD|stt|}||_||_| s*|tkr6t|t|<t||_dS)aCreate an hyphenation instance for given lang or filename. :param filename: filename 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)rrleftrighthdcacher7hd)r&rlangrfrgrTrrrr)s   zPyphen.__init__cs*t|jfddjj|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|koknr|qSr)rf)rrD)rgr&rrrGsz$Pyphen.positions..)rHrgrire)r&rbr)rgr&rres zPyphen.positionsccsxt|j|D]~}|jrt|j\}}}||7}|jr>|j}|jd\}}|d||||||dfVq|d|||dfVqWdS)zIterate over all hyphenation possibilities, the longest first. :param word: unicode string of the word to hyphenate =N)reversedrer3isupperupperr)r&rbpositionr!r#r$Zc1c2rrriterate s (zPyphen.iteratercCs@|t|8}x.|j|D] \}}t||kr|||fSqWdS)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)rHrq)r&rbwidthhyphenZw1Zw2rrrwraps  z Pyphen.wrapcCszt|}xft|j|D]T}|jr`|j\}}}||7}|jrF|j}|jd|||||<q|j||qWdj|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 ``'-'``. rkr) listrlrer3rmrnrinsertr)r&rbrsZ word_listror!r#r$rrrinserted-s  zPyphen.inserted)NNrrT)r)r) r-r.r/r0r)rerqrtrwr,rrrrrs    )rrr)"r0 __future__rr sysrer> NameErrorchr__all__rhcompilesubrPfindallrR pkg_resourcesr rresource_filenameZdictionaries_roots ImportErrorprefixdirname__file__dictrrobjectrr"r*r7rrrrrs<      g