Grammalecte  Check-in [391d4d820a]

    "set options to default values"
    global _dOptions
    _dOptions = getDefaultOptions()


#### Parsing

def parse (sText, sCountry="${country_default}", bDebug=False, dOptions=None, bContext=False):
    "init point to analyze a text"
    oText = TextParser(sText)
    return oText.parse(sCountry, bDebug, dOptions, bContext)


#### TEXT PARSER

class TextParser:
    "Text parser"

    def __init__ (self, sText):
        self.sText = sText
        self.sText0 = sText
        self.sSentence = ""
        self.sSentence0 = ""
        self.nOffsetWithinParagraph = 0
        self.lToken = []
        self.dTokenPos = {}
        self.dTags = {}
        self.dError = {}

        self.dErrorPriority = {}  # Key = position; value = priority

    def __str__ (self):
        s = "===== TEXT =====\n"
        s += "sentence: " + self.sSentence0 + "\n"
        s += "now:      " + self.sSentence  + "\n"
        for dToken in self.lToken:
            s += '#{i}\t{nStart}:{nEnd}\t{sValue}\t{sType}'.format(**dToken)
            if "lMorph" in dToken:
                s += "\t" + str(dToken["lMorph"])
            if "aTags" in dToken:
                s += "\t" + str(dToken["aTags"])
            s += "\n"
        #for nPos, dToken in self.dTokenPos.items():
        #    s += "{}\t{}\n".format(nPos, dToken)
        return s

    def parse (self, sCountry="${country_default}", bDebug=False, dOptions=None, bContext=False):
        "analyses sText and returns an iterable of errors"
        #sText = unicodedata.normalize("NFC", sText)
        dOpt = dOptions or _dOptions
        bShowRuleId = option('idrule')
        # parse paragraph
        try:
            self.parseText(self.sText, self.sText0, True, 0, sCountry, dOpt, bShowRuleId, bDebug, bContext)
        except:
            raise
        # parse sentences
        sText = self._getCleanText()

        for iStart, iEnd in text.getSentenceBoundaries(sText):
            if 4 < (iEnd - iStart) < 2000:
                try:
                    self.sSentence = sText[iStart:iEnd]
                    self.sSentence0 = self.sText0[iStart:iEnd]
                    self.nOffsetWithinParagraph = iStart
                    self.lToken = list(_oTokenizer.genTokens(self.sSentence, True))
                    self.dTokenPos = { dToken["nStart"]: dToken  for dToken in self.lToken  if dToken["sType"] != "INFO" }



                    self.parseText(self.sSentence, self.sSentence0, False, iStart, sCountry, dOpt, bShowRuleId, bDebug, bContext)
                except:
                    raise
        return self.dError.values() # this is a view (iterable)

    def parseAndGetSentences (self, sCountry="${country_default}", bDebug=False):
        "analyses sText and returns a list of sentences with their tokens"
        #sText = unicodedata.normalize("NFC", sText)
        # parse paragraph
        try:
            self.parseText(self.sText, self.sText0, True, 0, sCountry, dOptions, bShowRuleId, bDebug, bContext)
        except:
            raise
        # parse sentences
        sText = self._getCleanText()
        lSentence = []
        i = 0
        for iStart, iEnd in text.getSentenceBoundaries(sText):
            try:
                self.sSentence = sText[iStart:iEnd]
                self.sSentence0 = self.sText0[iStart:iEnd]
                self.nOffsetWithinParagraph = iStart
                self.lToken = list(_oTokenizer.genTokens(self.sSentence, True))
                self.dTokenPos = { dToken["nStart"]: dToken  for dToken in self.lToken  if dToken["sType"] != "INFO" }
                i += 1
                lSentence.append({
                    "i": i,
                    "iStart": iStart,
                    "iEnd": iEnd,
                    "sSentence": self.sSentence,
                    "sSentence0": self.sSentence0,
                    "lToken": list(lToken) # this is a copy
                })
                self.parseText(self.sSentence, self.sSentence0, False, iStart, sCountry, dOptions, False, False, False)
            except:
                raise


        return lSentence




    def _getCleanText (self):
        sText = self.sText
        if " " in sText:
            sText = sText.replace(" ", ' ') # nbsp
        if " " in sText:
            sText = sText.replace(" ", ' ') # nnbsp

                                            echo("RULE: " + sLineId)
                                        if cActionType == "-":
                                            # grammar error
                                            nErrorStart = nOffset + m.start(eAct[0])
                                            if nErrorStart not in self.dError or nPriority > self.dErrorPriority.get(nErrorStart, -1):
                                                self.dError[nErrorStart] = self._createErrorFromRegex(sText, sText0, sWhat, nOffset, m, eAct[0], sLineId, sRuleId, bUppercase, eAct[1], eAct[2], bShowRuleId, sOption, bContext)
                                                self.dErrorPriority[nErrorStart] = nPriority

                                        elif cActionType == "~":
                                            # text processor
                                            sText = self.rewriteText(sText, sWhat, eAct[0], m, bUppercase)
                                            bChange = True
                                            if bDebug:
                                                echo("~ " + sText + "  -- " + m.group(eAct[0]) + "  # " + sLineId)
                                        elif cActionType == "=":

                                if "bImmune" not in self.lToken[nTokenErrorStart]:
                                    nTokenErrorEnd = nTokenOffset + iTokenEnd  if iTokenEnd > 0  else nLastToken + iTokenEnd
                                    nErrorStart = self.nOffsetWithinParagraph + (self.lToken[nTokenErrorStart]["nStart"] if cStartLimit == "<"  else self.lToken[nTokenErrorStart]["nEnd"])
                                    nErrorEnd = self.nOffsetWithinParagraph + (self.lToken[nTokenErrorEnd]["nEnd"] if cEndLimit == ">"  else self.lToken[nTokenErrorEnd]["nStart"])
                                    if nErrorStart not in self.dError or nPriority > self.dErrorPriority.get(nErrorStart, -1):
                                        self.dError[nErrorStart] = self._createErrorFromTokens(sWhat, nTokenOffset, nLastToken, nTokenErrorStart, nErrorStart, nErrorEnd, sLineId, sRuleId, bCaseSvty, sMessage, sURL, bShowRuleId, sOption, bContext)
                                        self.dErrorPriority[nErrorStart] = nPriority

                                        if bDebug:
                                            echo("    NEW_ERROR: {}".format(self.dError[nErrorStart]))
                            elif cActionType == "~":
                                # text processor
                                nTokenStart = nTokenOffset + eAct[0]  if eAct[0] > 0  else nLastToken + eAct[0]
                                nTokenEnd = nTokenOffset + eAct[1]  if eAct[1] > 0  else nLastToken + eAct[1]
                                self._tagAndPrepareTokenForRewriting(sWhat, nTokenStart, nTokenEnd, nTokenOffset, nLastToken, eAct[2], bDebug)