Grammalecte  Diff

#### Parsing

def parse (sText, sCountry="${country_default}", bDebug=False, dOptions=None, bContext=False):
    "analyses the paragraph sText and returns list of errors"
    #sText = unicodedata.normalize("NFC", sText)
    aErrors = None
    sRealText = sText

    dPriority = {}  # Key = position; value = priority
    dOpt = _dOptions  if not dOptions  else dOptions
    bShowRuleId = option('idrule')

    # parse paragraph
    try:
        sNew, aErrors = _proofread(None, sText, sRealText, 0, True, dPriority, sCountry, dOpt, bShowRuleId, bDebug, bContext)
        if sNew:
            sText = sNew
    except:
        raise

    # cleanup
    if " " in sText:
        sText = sText.replace(" ", ' ') # nbsp
    if " " in sText:
        sText = sText.replace(" ", ' ') # nnbsp
    if "'" in sText:
        sText = sText.replace("'", "’")
    if "‑" in sText:
        sText = sText.replace("‑", "-") # nobreakdash

    # parse sentences
    for iStart, iEnd in _getSentenceBoundaries(sText):
        if 4 < (iEnd - iStart) < 2000:

            try:
                oSentence = TokenSentence(sText[iStart:iEnd], sRealText[iStart:iEnd], iStart)
                _, errs = _proofread(oSentence, sText[iStart:iEnd], sRealText[iStart:iEnd], iStart, False, dPriority, sCountry, dOpt, bShowRuleId, bDebug, bContext)
                aErrors.update(errs)
            except:
                raise
    return aErrors.values() # this is a view (iterable)


_zEndOfSentence = re.compile(r'([.?!:;…][ .?!… »”")]*|.$)')
_zBeginOfParagraph = re.compile(r"^\W*")
_zEndOfParagraph = re.compile(r"\W*$")

def _getSentenceBoundaries (sText):
    iStart = _zBeginOfParagraph.match(sText).end()
    for m in _zEndOfSentence.finditer(sText):
        yield (iStart, m.end())
        iStart = m.end()


def _proofread (oSentence, s, sx, nOffset, bParagraph, dPriority, sCountry, dOptions, bShowRuleId, bDebug, bContext):
    dErrs = {}
    bParagraphChange = False
    bSentenceChange = False
    dTokenPos = oSentence.dTokenPos if oSentence else {}
    for sOption, lRuleGroup in _getRules(bParagraph):
        if sOption == "@@@@":
            # graph rules
            if not bParagraph and bSentenceChange:
                oSentence.update(s)
                bSentenceChange = False
            for sGraphName, sLineId in lRuleGroup:

            for zRegex, bUppercase, sLineId, sRuleId, nPriority, lActions in lRuleGroup:
                if sRuleId not in _aIgnoredRules:
                    for m in zRegex.finditer(s):
                        bCondMemo = None
                        for sFuncCond, cActionType, sWhat, *eAct in lActions:
                            # action in lActions: [ condition, action type, replacement/suggestion/action[, iGroup[, message, URL]] ]
                            try:
                                bCondMemo = not sFuncCond or globals()[sFuncCond](s, sx, m, dTokenPos, sCountry, bCondMemo)
                                if bCondMemo:
                                    if cActionType == "-":
                                        # grammar error
                                        nErrorStart = nOffset + m.start(eAct[0])
                                        if nErrorStart not in dErrs or nPriority > dPriority[nErrorStart]:
                                            dErrs[nErrorStart] = _createRegexError(s, sx, sWhat, nOffset, m, eAct[0], sLineId, sRuleId, bUppercase, eAct[1], eAct[2], bShowRuleId, sOption, bContext)
                                            dPriority[nErrorStart] = nPriority
                                    elif cActionType == "~":
                                        # text processor
                                        s = _rewrite(s, sWhat, eAct[0], m, bUppercase)
                                        bParagraphChange = True
                                        bSentenceChange = True
                                        if bDebug:
                                            echo("~ " + s + "  -- " + m.group(eAct[0]) + "  # " + sLineId)
                                    elif cActionType == "=":
                                        # disambiguation
                                        if not bParagraph:
                                            globals()[sWhat](s, m, dTokenPos)
                                            if bDebug:
                                                echo("= " + m.group(0) + "  # " + sLineId)
                                    elif cActionType == ">":
                                        # we do nothing, this test is just a condition to apply all following actions
                                        pass
                                    else:
                                        echo("# error: unknown action at " + sLineId)
                                elif cActionType == ">":
                                    break

#### common functions

def option (sOpt):
    "return True if option sOpt is active"
    return _dOptions.get(sOpt, False)


def displayInfo (dTokenPos, tWord):
    "for debugging: retrieve info of word"
    if not tWord:
        echo("> nothing to find")
        return True
    lMorph = _oSpellChecker.getMorph(tWord[1])
    if not lMorph:
        echo("> not in dictionary")
        return True
    if tWord[0] in dTokenPos and "lMorph" in dTokenPos[tWord[0]]:
        echo("DA: " + str(dTokenPos[tWord[0]]["lMorph"]))
    echo("FSA: " + str(lMorph))
    return True


def morph (dTokenPos, tWord, sPattern, bStrict=True, bNoWord=False):
    "analyse a tuple (position, word), return True if sPattern in morphologies (disambiguation on)"
    if not tWord:
        return bNoWord
    lMorph = dTokenPos[tWord[0]]["lMorph"]  if tWord[0] in dTokenPos and "lMorph" in dTokenPos[tWord[0]]  else _oSpellChecker.getMorph(tWord[1])
    if not lMorph:
        return False
    p = re.compile(sPattern)
    if bStrict:
        return all(p.search(s)  for s in lMorph)
    return any(p.search(s)  for s in lMorph)


def morphex (dTokenPos, tWord, sPattern, sNegPattern, bNoWord=False):
    "analyse a tuple (position, word), returns True if not sNegPattern in word morphologies and sPattern in word morphologies (disambiguation on)"
    if not tWord:
        return bNoWord
    lMorph = dTokenPos[tWord[0]]["lMorph"]  if tWord[0] in dTokenPos and "lMorph" in dTokenPos[tWord[0]]  else _oSpellChecker.getMorph(tWord[1])
    if not lMorph:
        return False
    # check negative condition
    np = re.compile(sNegPattern)
    if any(np.search(s)  for s in lMorph):
        return False
    # search sPattern

    if sNegPattern and re.search(sNegPattern, s):
        return False
    if re.search(sPattern, s):
        return True
    return False


def look_chk1 (dTokenPos, s, nOffset, sPattern, sPatternGroup1, sNegPatternGroup1=None):
    "returns True if s has pattern sPattern and m.group(1) has pattern sPatternGroup1"
    m = re.search(sPattern, s)
    if not m:
        return False
    try:
        sWord = m.group(1)
        nPos = m.start(1) + nOffset
    except:
        return False
    if sNegPatternGroup1:
        return morphex(dTokenPos, (nPos, sWord), sPatternGroup1, sNegPatternGroup1)
    return morph(dTokenPos, (nPos, sWord), sPatternGroup1, False)


#### Disambiguator

def select (dTokenPos, nPos, sWord, sPattern, lDefault=None):
    if not sWord:
        return True
    if nPos not in dTokenPos:
        print("Error. There should be a token at this position: ", nPos)
        return True
    lMorph = _oSpellChecker.getMorph(sWord)
    if not lMorph or len(lMorph) == 1:
        return True
    lSelect = [ sMorph  for sMorph in lMorph  if re.search(sPattern, sMorph) ]
    if lSelect:
        if len(lSelect) != len(lMorph):
            dTokenPos[nPos]["lMorph"] = lSelect
    elif lDefault:
        dTokenPos[nPos]["lMorph"] = lDefault
    return True


def exclude (dTokenPos, nPos, sWord, sPattern, lDefault=None):
    if not sWord:
        return True
    if nPos not in dTokenPos:
        print("Error. There should be a token at this position: ", nPos)
        return True
    lMorph = _oSpellChecker.getMorph(sWord)
    if not lMorph or len(lMorph) == 1:
        return True
    lSelect = [ sMorph  for sMorph in lMorph  if not re.search(sPattern, sMorph) ]
    if lSelect:
        if len(lSelect) != len(lMorph):
            dTokenPos[nPos]["lMorph"] = lSelect
    elif lDefault:
        dTokenPos[nPos]["lMorph"] = lDefault
    return True


def define (dTokenPos, nPos, lMorph):
    if nPos not in dTokenPos:
        print("Error. There should be a token at this position: ", nPos)
        return True
    dTokenPos[nPos]["lMorph"] = lMorph
    return True


#### GRAMMAR CHECKER PLUGINS

${plugins}