ó Nä²Wc@@s\ddlmZddlZyddlmZWn!ek rSddlmZnXddlmZdZ dZ d Z e d ƒZ d „Zd „Zd „Zd„Zd„Zdefd„ƒYZd„Zdefd„ƒYZdefd„ƒYZeƒZdefd„ƒYZdefd„ƒYZdefd„ƒYZdefd„ƒYZdefd „ƒYZeƒZd!e_d"e_ d#„Z!d$„Z"d%„Z#d&„Z$e$d'ƒZ%d(e%_d)e%_ dd*„Z'd+„Z(d,„Z)d-„Z*d.„Z+ee ƒZ,d/e,_d0e,_ ee ƒZ-d1e-_d2e-_ ee ƒZ.d3e._d4e._ dS(5i(tabsolute_importN(tmaxsize(tmaxinti(tErrorstbolteolteofs cC@sÀt|ƒ}|jƒd}t|ƒ}g}x‹||kr»t||ƒ}|d}|d7}x:||kr|t||ƒkr|d7}|d7}qdW|j|ƒ|j|ƒq1W|S(s’ Return a list of character codes consisting of pairs [code1a, code1b, code2a, code2b,...] which cover all the characters in |s|. ii(tlisttsorttlentordtappend(tst char_listtitntresulttcode1tcode2((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pytchars_to_ranges"s     %  cC@sjt|tdƒƒ}t|tdƒdƒ}||krbtdƒtdƒ}||||fSdSdS(sŠ If the range of characters from code1 to code2-1 includes any lower case letters, return the corresponding upper case range. tatzitAN(tmaxR tmintNone(RRtcode3tcode4td((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pytuppercase_range9s  cC@sjt|tdƒƒ}t|tdƒdƒ}||krbtdƒtdƒ}||||fSdSdS(sŠ If the range of characters from code1 to code2-1 includes any upper case letters, return the corresponding lower case range. RtZiRN(RR RR(RRRRR((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pytlowercase_rangeGs  cC@sJgtdt|ƒdƒD]!}t||||dƒ^q}t|ŒS(sƒ Given a list of codes as returned by chars_to_ranges, return an RE which will match a character in any of the ranges. iii(trangeR t CodeRangetAlt(t code_listRtre_list((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyt CodeRangesUs@cC@sS|tko|knrBtt|tƒtttd|ƒƒSt||ƒSdS(s CodeRange(code1, code2) is an RE which matches any character with a code |c| in the range |code1| <= |c| < |code2|. iN(tnl_codeR"t RawCodeRanget RawNewline(RR((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR!^s tREcB@sqeZdZdZdZd Zd„Zd„Zd„Z d„Z d„Z d„Z d„Z d „Zd „ZRS( sRE is the base class for regular expression constructors. The following operators are defined on REs: re1 + re2 is an RE which matches |re1| followed by |re2| re1 | re2 is an RE which matches either |re1| or |re2| icC@std|jjƒ‚dS(sM This method should add states to |machine| to implement this RE, starting at |initial_state| and ending at |final_state|. If |match_bol| is true, the RE must be able to match at the beginning of a line. If nocase is true, upper and lower case letters should be treated as equivalent. s %s.build_machine not implementedN(tNotImplementedErrort __class__t__name__(tselftmachinet initial_statet final_statet match_boltnocase((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyt build_machine{s cC@s-|jƒ}|j|ƒ|j||ƒ|S(s~ Given a state |s| of machine |m|, return a new state reachable from |s| on character |c| or epsilon. (t new_statetlink_totadd_transition(R-tmR/tcR ((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyt build_opt‡s  cC@s t||ƒS(N(tSeq(R-tother((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyt__add__‘scC@s t||ƒS(N(R"(R-R;((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyt__or__”scC@s|jr|jS|jƒSdS(N(tstrtcalc_str(R-((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyt__str__—s cC@s)t|tƒs%|j||dƒndS(NsPlex.RE instance(t isinstanceR)t wrong_type(R-tnumtvalue((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pytcheck_rescC@s2t|ƒtdƒkr.|j||dƒndS(Nttstring(ttypeRB(R-RCRD((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyt check_string¡scC@sQ|j||ƒt|ƒdkrMtjd||jjt|ƒfƒ‚ndS(NisOInvalid value for argument %d of Plex.%s.Expected a string of length 1, got: %s(RIR RtPlexValueErrorR+R,trepr(R-RCRD((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyt check_char¥s cC@slt|ƒtjkr4d|jj|jjf}nt|ƒj}tjd||jj||fƒ‚dS(Ns%s.%s instances<Invalid type for argument %d of Plex.%s (expected %s, got %s(RHttypest InstanceTypeR+t __module__R,Rt PlexTypeError(R-RCRDtexpectedtgot((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyRB¬s  N(R,ROt__doc__tnullabletmatch_nlRR>R3R9R<R=R@RERIRLRB(((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR)os      cC@sWt|ƒdkr4tt|ƒt|ƒdƒ}n t|ƒ}dt|ƒ|_|S(s; Char(c) is an RE which matches the character |c|. isChar(%s)(R R!R t SpecialSymbolRKR>(R8R((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pytCharÝs " R'cB@sGeZdZdZdZdZdZdZd„Z d„Z d„Z RS(sÒ RawCodeRange(code1, code2) is a low-level RE which matches any character with a code |c| in the range |code1| <= |c| < |code2|, where the range does not include newline. For internal use only. icC@s7||f|_t||ƒ|_t||ƒ|_dS(N(R RR(R-RR((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyt__init__õscC@s||r|j||tƒ}n|j|j|ƒ|rx|jrV|j|j|ƒn|jrx|j|j|ƒqxndS(N(R9tBOLR6R RR(R-R7R/R0R1R2((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR3ús  cC@sd|j|jfS(NsCodeRange(%d,%d)(RR(R-((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR?sN( R,RORSRTRURR RRRXR3R?(((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR'és  t _RawNewlinecB@s#eZdZdZdZd„ZRS(sd RawNewline is a low-level RE which matches a newline character. For internal use only. iicC@sQ|r|j||tƒ}n|j||tƒ}|jttdf|ƒdS(Ni(R9RYtEOLR6R&(R-R7R/R0R1R2R ((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR3s(R,RORSRTRUR3(((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyRZsRVcB@s2eZdZdZdZdZd„Zd„ZRS(sx SpecialSymbol(sym) is an RE which matches the special input symbol |sym|, which is one of BOL, EOL or EOF. icC@s ||_dS(N(tsym(R-R\((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyRX#scC@sD|r-|jtkr-|j||tƒ}n|j|j|ƒdS(N(R\R[R9RYR6(R-R7R/R0R1R2((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR3&sN( R,RORSRTRURR\RXR3(((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyRVs  R:cB@s)eZdZd„Zd„Zd„ZRS(s]Seq(re1, re2, re3...) is an RE which matches |re1| followed by |re2| followed by |re3|...cG@s´d}x9t|ƒD]+\}}|j||ƒ|o;|j}qW||_||_t|ƒ}d}x>|r¦|d8}||}|jr–d}Pn|jsiPqiqiW||_dS(Nii(t enumerateRERTR$R RU(R-R$RTRtreRU((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyRX2s"        c C@sº|j}t|ƒdkr+|j|ƒn‹|}t|ƒ}xvt|ƒD]h\} } | |dkru|jƒ} n|} | j||| ||ƒ| }| jp¯|o¯| j}qJWdS(Nii(R$R R5R]R4R3RURT( R-R7R/R0R1R2R$ts1RRR^ts2((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR3Es  cC@sddjtt|jƒƒS(NsSeq(%s)t,(tjointmapR>R$(R-((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR?Us(R,RORSRXR3R?(((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR:.s  R"cB@s)eZdZd„Zd„Zd„ZRS(sRAlt(re1, re2, re3...) is an RE which matches either |re1| or |re2| or |re3|...cG@s¸||_d}d}g}g}d}xf|D]^}|j||ƒ|jrc|j|ƒd}n |j|ƒ|jr‚d}n|d7}q.W||_||_||_||_dS(Nii(R$RERTR RUt nullable_restnon_nullable_res(R-R$RTRURdReRR^((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyRX]s&           cC@sˆx*|jD]}|j|||||ƒq W|jr„|rT|j||tƒ}nx-|jD]}|j|||d|ƒq^WndS(Ni(RdR3ReR9RY(R-R7R/R0R1R2R^((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR3ss cC@sddjtt|jƒƒS(NsAlt(%s)Ra(RbRcR>R$(R-((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR?|s(R,RORSRXR3R?(((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR"Ys  tRep1cB@s)eZdZd„Zd„Zd„ZRS(s@Rep1(re) is an RE which matches one or more repetitions of |re|.cC@s5|jd|ƒ||_|j|_|j|_dS(Ni(RER^RTRU(R-R^((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyRXƒs  cC@sk|jƒ}|jƒ}|j|ƒ|jj||||pF|jj|ƒ|j|ƒ|j|ƒdS(N(R4R5R^R3RU(R-R7R/R0R1R2R_R`((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR3‰s    ( cC@s d|jS(NsRep1(%s)(R^(R-((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR?‘s(R,RORSRXR3R?(((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyRf€s  t SwitchCasecB@s5eZdZdZdZd„Zd„Zd„ZRS(sÓ SwitchCase(re, nocase) is an RE which matches the same strings as RE, but treating upper and lower case letters according to |nocase|. If |nocase| is true, case is ignored, otherwise it is not. cC@s.||_||_|j|_|j|_dS(N(R^R2RTRU(R-R^R2((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyRXžs   cC@s#|jj|||||jƒdS(N(R^R3R2(R-R7R/R0R1R2((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR3¤scC@s)|jrd}nd}d||jfS(NtNoCasetCases%s(%s)(R2R^(R-tname((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyR?¨s  N( R,RORSRR^R2RXR3R?(((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyRg•s   s8 Empty is an RE which matches the empty string. tEmptycC@s2tttt|ƒƒŒ}dt|ƒ|_|S(s@ Str1(s) is an RE which matches the literal string |s|. sStr(%s)(R:ttupleRcRWRKR>(R R((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pytStr1¾scG@sbt|ƒdkr t|dƒStttt|ƒƒŒ}ddjtt|ƒƒ|_|SdS(sŽ Str(s) is an RE which matches the literal string |s|. Str(s1, s2, s3, ...) is an RE which matches any of |s1| or |s2| or |s3|... iisStr(%s)RaN(R RmR"RlRcRbRKR>(tstrsR((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pytStrÇs cC@s)tt|ƒƒ}dt|ƒ|_|S(sH Any(s) is an RE which matches any character in the string |s|. sAny(%s)(R%RRKR>(R R((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pytAnyÔscC@sMt|ƒ}|jdt ƒ|jtƒt|ƒ}dt|ƒ|_|S(sp AnyBut(s) is an RE which matches any character (including newline) which is not in the string |s|. is AnyBut(%s)(RtinsertRR R%RKR>(R trangesR((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pytAnyButÞs    RFsT AnyChar is an RE which matches any single character (including a newline). tAnyCharcC@s¹|r;tt|ƒt|ƒdƒ}d||f|_nzg}xRtdt|ƒdƒD]8}|jtt||ƒt||dƒdƒƒqZWt|Œ}dt|ƒ|_|S(s Range(c1, c2) is an RE which matches any single character in the range |c1| to |c2| inclusive. Range(s) where |s| is a string of even length is an RE which matches any single character in the ranges |s[0]| to |s[1]|, |s[2]| to |s[3]|,... is Range(%s,%s)iis Range(%s)(R!R R>R R R R"RK(R_R`RRrR((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pytRangeós6 cC@s t|tƒ}d||_|S(sI Opt(re) is an RE which matches either |re| or the empty string. sOpt(%s)(R"RkR>(R^R((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pytOpts cC@s#tt|ƒƒ}d||_|S(sJ Rep(re) is an RE which matches zero or more repetitions of |re|. sRep(%s)(RvRfR>(R^R((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pytReps cC@st|ddƒS(s€ NoCase(re) is an RE which matches the same strings as RE, but treating upper and lower case letters as equivalent. R2i(Rg(R^((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyRhscC@st|ddƒS(s¶ Case(re) is an RE which matches the same strings as RE, but treating upper and lower case letters as distinct, i.e. it cancels the effect of any enclosing NoCase(). R2i(Rg(R^((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyRi ss= Bol is an RE which matches the beginning of a line. tBols7 Eol is an RE which matches the end of a line. tEols9 Eof is an RE which matches the end of the file. tEof(/t __future__RRMtsysRRt ImportErrorRFRRYR[tEOFR R&RRRR%R!tobjectR)RWR'RZR(RVR:R"RfRgRkRSR>RmRoRpRsRtRRuRvRwRhRiRxRyRz(((s6/bar/jli/Chip-seq/script/cython/Cython/Plex/Regexps.pyt sb       n  +'!