Grammalecte  Check-in [1184e8ba6d]

    s = re.sub(r"isRealStart *\(\)", 'before("^ *$")', s)
    s = re.sub(r"isStart0 *\(\)", 'before0("^ *$|, *$")', s)
    s = re.sub(r"isRealStart0 *\(\)", 'before0("^ *$")', s)
    s = re.sub(r"isEnd *\(\)", 'after("^ *$|^,")', s)
    s = re.sub(r"isRealEnd *\(\)", 'after("^ *$")', s)
    s = re.sub(r"isEnd0 *\(\)", 'after0("^ *$|^,")', s)
    s = re.sub(r"isRealEnd0 *\(\)", 'after0("^ *$")', s)
    s = re.sub(r"(select|exclude)[(][\\](\d+)", '\\1(dTokenPos, m.start(\\2), m.group(\\2)', s)
    s = re.sub(r"define[(][\\](\d+)", 'define(dTokenPos, m.start(\\1)', s)
    s = re.sub(r"(morph|morphex|displayInfo)[(][\\](\d+)", '\\1((m.start(\\2), m.group(\\2))', s)
    s = re.sub(r"(morph|morphex|displayInfo)[(]", '\\1(dTokenPos, ', s)
    s = re.sub(r"(sugg\w+|switch\w+)\(@", '\\1(m.group(i[4])', s)
    s = re.sub(r"word\(\s*1\b", 'nextword1(s, m.end()', s)                                  # word(1)
    s = re.sub(r"word\(\s*-1\b", 'prevword1(s, m.start()', s)                               # word(-1)
    s = re.sub(r"word\(\s*(\d)", 'nextword(s, m.end(), \\1', s)                             # word(n)
    s = re.sub(r"word\(\s*-(\d)", 'prevword(s, m.start(), \\1', s)                          # word(-n)
    s = re.sub(r"before\(\s*", 'look(s[:m.start()], ', s)                                   # before(s)
    s = re.sub(r"after\(\s*", 'look(s[m.end():], ', s)                                      # after(s)
    s = re.sub(r"textarea\(\s*", 'look(s, ', s)                                             # textarea(s)
    s = re.sub(r"before_chk1\(\s*", 'look_chk1(dTokenPos, s[:m.start()], 0, ', s)           # before_chk1(s)
    s = re.sub(r"after_chk1\(\s*", 'look_chk1(dTokenPos, s[m.end():], m.end(), ', s)        # after_chk1(s)
    s = re.sub(r"textarea_chk1\(\s*", 'look_chk1(dTokenPos, s, 0, ', s)                     # textarea_chk1(s)
    s = re.sub(r"/0", 'sx[m.start():m.end()]', s)                                           # /0
    s = re.sub(r"before0\(\s*", 'look(sx[:m.start()], ', s)                                 # before0(s)
    s = re.sub(r"after0\(\s*", 'look(sx[m.end():], ', s)                                    # after0(s)
    s = re.sub(r"textarea0\(\s*", 'look(sx, ', s)                                           # textarea0(s)
    s = re.sub(r"before0_chk1\(\s*", 'look_chk1(dTokenPos, sx[:m.start()], 0, ', s)         # before0_chk1(s)
    s = re.sub(r"after0_chk1\(\s*", 'look_chk1(dTokenPos, sx[m.end():], m.end(), ', s)      # after0_chk1(s)
    s = re.sub(r"textarea0_chk1\(\s*", 'look_chk1(dTokenPos, sx, 0, ', s)                   # textarea0_chk1(s)
    s = re.sub(r"isEndOfNG\(\s*\)", 'isEndOfNG(dTokenPos, s[m.end():], m.end())', s)        # isEndOfNG(s)
    s = re.sub(r"isNextNotCOD\(\s*\)", 'isNextNotCOD(dTokenPos, s[m.end():], m.end())', s)  # isNextNotCOD(s)
    s = re.sub(r"isNextVerb\(\s*\)", 'isNextVerb(dTokenPos, s[m.end():], m.end())', s)      # isNextVerb(s)
    s = re.sub(r"\bspell *[(]", '_oSpellChecker.isValid(', s)
    s = re.sub(r"[\\](\d+)", 'm.group(\\1)', s)
    return s


def uppercase (s, sLang):
    "(flag i is not enough): converts regex to uppercase regex: 'foo' becomes '[Ff][Oo][Oo]', but 'Bar' becomes 'B[Aa][Rr]'."

    # creating file with all functions callable by rules
    print("  creating callables...")
    sPyCallables = "# generated code, do not edit\n"
    sJSCallables = "// generated code, do not edit\nconst oEvalFunc = {\n"
    for sFuncName, sReturn in lFUNCTIONS:
        cType = sFuncName[0:1]
        if cType == "c": # condition
            sParams = "s, sx, m, dTokenPos, sCountry, bCondMemo"
        elif cType == "m": # message
            sParams = "s, m"
        elif cType == "s": # suggestion
            sParams = "s, m"
        elif cType == "p": # preprocessor
            sParams = "s, m"
        elif cType == "d": # disambiguator
            sParams = "s, m, dTokenPos"
        else:
            print("# Unknown function type in [" + sFuncName + "]")
            continue
        sPyCallables += "def {} ({}):\n".format(sFuncName, sParams)
        sPyCallables += "    return " + sReturn + "\n"
        sJSCallables += "    {}: function ({})".format(sFuncName, sParams) + " {\n"
        sJSCallables += "        return " + jsconv.py2js(sReturn) + ";\n"

#### 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}

# URL
__<i>(p_URL)__
    https?://[\w./?&!%=+*"'@$#-]+ <<- ~>> *
__<i](p_URL2)__
    ((?:{w_1}[.])*)({w_2})([.](?:com|net|org|info|fr|ca|be|ch|i[ot]|co[.]uk|tk|es|jp|zh|ru|us|nl|xyz)) @@0,**,$
    <<- ~1>> *
    <<- ~2>> =\2.capitalize()

    <<- ~3>> *

# Numéro de chapitre
__<i>(p_chapitre)__
    ^\d+[.][\d.-]* <<- ~>> *

# Numéro suivi de plusieurs espaces, considéré comme une numérotation de chapitre

TEST: dès qu’il le {{voie}}
TEST: donnant à entendre qu’il avait l’intention de violer Laura dès qu’il en aurait l’occasion


# verbe que + subjonctif
__vmode_qqch_que_subjonctif1__
    [>afin|>avant|>pour|>quoi|>permettre|>falloir|>vouloir|>ordonner|>exiger|>désirer|>préférer|>suffire]  [que|qu’|qu]  @:(?:Os|M)  @:I¬:[GYS]
        <<- /vmode/ -4>> =suggVerbMode(\4, ":S", \3)                                                # Après « \1 que », ce verbe devrait être au subjonctif.

    >douter  [que|qu’|qu]  @:(?:Os|M)  @:I¬:(?:[GYSK]|If)
        <<- /vmode/ morph(\1, ":V", ":N") -4>> =suggVerbMode(\4, ":S", \3)                          # Après « \1 que », ce verbe devrait être au subjonctif.

TEST: Il suffit qu’il {{court}} plus
TEST: Je veux qu’il {{finit}} son repas.
TEST: quoi qu’il en {{conclut}}
TEST: Je ne veux pas que tu {{es}} des ennuis
TEST: Avant que tu {{pars}}, je voudrais qu’on discute.
TEST: Nul doute qu’elle nourrira à brève échéance la haine de demain à notre égard.