Grammalecte  Diff

        // parse paragraph
        try {
            this.parseText(this.sText, this.sText0, true, 0, sCountry, dOpt, bShowRuleId, bDebug, bContext);
        }
        catch (e) {
            console.error(e);
        }
        this.lTokens = null;
        this.lTokens0 = null;
        let lParagraphErrors = null;
        if (bFullInfo) {
            lParagraphErrors = Array.from(this.dError.values());
            this.dSentenceError.clear();
        }
        // parse sentence
        let sText = this._getCleanText();
        let lSentences = [];
        let oSentence = null;
        for (let [iStart, iEnd] of text.getSentenceBoundaries(sText)) {
            try {
                this.sSentence = sText.slice(iStart, iEnd);
                this.sSentence0 = this.sText0.slice(iStart, iEnd);
                this.nOffsetWithinParagraph = iStart;
                this.lTokens = Array.from(gc_engine.oTokenizer.genTokens(this.sSentence, true));
                this.dTokenPos.clear();
                for (let dToken of this.lTokens) {
                    if (dToken["sType"] != "INFO") {
                        this.dTokenPos.set(dToken["nStart"], dToken);
                    }
                }
                if (bFullInfo) {
                    this.lTokens0 = Array.from(this.lTokens);





                    // the list of tokens is duplicated, to keep all tokens from being deleted when analysis
                }
                this.parseText(this.sSentence, this.sSentence0, false, iStart, sCountry, dOpt, bShowRuleId, bDebug, bContext);
                if (bFullInfo) {
                    for (let oToken of this.lTokens0) {
                        if (oToken["sType"] == "WORD") {
                            oToken["bValidToken"] = gc_engine.oSpellChecker.isValidToken(oToken["sValue"]);
                        }
                        if (!oToken.hasOwnProperty("lMorph")) {
                            oToken["lMorph"] = gc_engine.oSpellChecker.getMorph(oToken["sValue"]);
                        }
                        gc_engine.oSpellChecker.setLabelsOnToken(oToken);
                    }
                    lSentences.push({
                        "nStart": iStart,
                        "nEnd": iEnd,
                        "sSentence": this.sSentence0,
                        "lTokens": this.lTokens0,
                        "lGrammarErrors": Array.from(this.dSentenceError.values())
                    });
                    this.dSentenceError.clear();
                }
            }
            catch (e) {
                console.error(e);
            }
        }

            if (this.dTokenPos.gl_get(oToken["nStart"], {}).hasOwnProperty("lMorph")) {
                oToken["lMorph"] = this.dTokenPos.get(oToken["nStart"])["lMorph"];
            }
            if (this.dTokenPos.gl_get(oToken["nStart"], {}).hasOwnProperty("aTags")) {
                oToken["aTags"] = this.dTokenPos.get(oToken["nStart"])["aTags"];
            }
        }
        this.lTokens = lNewToken;
        this.dTokenPos.clear();
        for (let oToken of this.lTokens) {
            if (oToken["sType"] != "INFO") {
                this.dTokenPos.set(oToken["nStart"], oToken);
            }
        }
        if (bDebug) {
            console.log("UPDATE:");
            console.log(this.asString());

    }

    parseGraph (oGraph, sCountry="${country_default}", dOptions=null, bShowRuleId=false, bDebug=false, bContext=false) {
        // parse graph with tokens from the text and execute actions encountered
        let lPointer = [];
        let bTagAndRewrite = false;
        try {
            for (let [iToken, oToken] of this.lTokens.entries()) {
                if (bDebug) {
                    console.log("TOKEN: " + oToken["sValue"]);
                }
                // check arcs for each existing pointer
                let lNextPointer = [];
                for (let oPointer of lPointer) {
                    lNextPointer.push(...this._getNextPointers(oToken, oGraph, oPointer, bDebug));

                    // Suggestion    [ option, condition, "-", replacement/suggestion/action, iTokenStart, iTokenEnd, cStartLimit, cEndLimit, bCaseSvty, nPriority, sMessage, iURL ]
                    // TextProcessor [ option, condition, "~", replacement/suggestion/action, iTokenStart, iTokenEnd, bCaseSvty ]
                    // Disambiguator [ option, condition, "=", replacement/suggestion/action ]
                    // Tag           [ option, condition, "/", replacement/suggestion/action, iTokenStart, iTokenEnd ]
                    // Immunity      [ option, condition, "!", "",                            iTokenStart, iTokenEnd ]
                    // Test          [ option, condition, ">", "" ]
                    if (!sOption || dOptions.gl_get(sOption, false)) {
                        bCondMemo = !sFuncCond || gc_functions[sFuncCond](this.lTokens, nTokenOffset, nLastToken, sCountry, bCondMemo, this.dTags, this.sSentence, this.sSentence0);
                        //bCondMemo = !sFuncCond || oEvalFunc[sFuncCond](this.lTokens, nTokenOffset, nLastToken, sCountry, bCondMemo, this.dTags, this.sSentence, this.sSentence0);
                        if (bCondMemo) {
                            if (cActionType == "-") {
                                // grammar error
                                let [iTokenStart, iTokenEnd, cStartLimit, cEndLimit, bCaseSvty, nPriority, sMessage, iURL] = eAct;
                                let nTokenErrorStart = (iTokenStart > 0) ? nTokenOffset + iTokenStart : nLastToken + iTokenStart;
                                if (!this.lTokens[nTokenErrorStart].hasOwnProperty("bImmune")) {
                                    let nTokenErrorEnd = (iTokenEnd > 0) ? nTokenOffset + iTokenEnd : nLastToken + iTokenEnd;
                                    let nErrorStart = this.nOffsetWithinParagraph + ((cStartLimit == "<") ? this.lTokens[nTokenErrorStart]["nStart"] : this.lTokens[nTokenErrorStart]["nEnd"]);
                                    let nErrorEnd = this.nOffsetWithinParagraph + ((cEndLimit == ">") ? this.lTokens[nTokenErrorEnd]["nEnd"] : this.lTokens[nTokenErrorEnd]["nStart"]);
                                    if (!this.dError.has(nErrorStart) || nPriority > this.dErrorPriority.gl_get(nErrorStart, -1)) {
                                        this.dError.set(nErrorStart, this._createErrorFromTokens(sWhat, nTokenOffset, nLastToken, nTokenErrorStart, nErrorStart, nErrorEnd, sLineId, sRuleId, bCaseSvty,
                                                                                                 sMessage, gc_rules_graph.dURL[iURL], bShowRuleId, sOption, bContext));
                                        this.dErrorPriority.set(nErrorStart, nPriority);
                                        this.dSentenceError.set(nErrorStart, this.dError.get(nErrorStart));
                                        if (bDebug) {
                                            console.log("    NEW_ERROR: ",  this.dError.get(nErrorStart));
                                        }
                                    }
                                }
                            }
                            else if (cActionType == "~") {
                                // text processor
                                let nTokenStart = (eAct[0] > 0) ? nTokenOffset + eAct[0] : nLastToken + eAct[0];
                                let nTokenEnd = (eAct[1] > 0) ? nTokenOffset + eAct[1] : nLastToken + eAct[1];
                                this._tagAndPrepareTokenForRewriting(sWhat, nTokenStart, nTokenEnd, nTokenOffset, nLastToken, eAct[2], bDebug);
                                bChange = true;
                                if (bDebug) {
                                    console.log(`    TEXT_PROCESSOR: [${this.lTokens[nTokenStart]["sValue"]}:${this.lTokens[nTokenEnd]["sValue"]}]  > ${sWhat}`);
                                }
                            }
                            else if (cActionType == "=") {
                                // disambiguation
                                gc_functions[sWhat](this.lTokens, nTokenOffset, nLastToken);
                                //oEvalFunc[sWhat](this.lTokens, nTokenOffset, nLastToken);
                                if (bDebug) {
                                    console.log(`    DISAMBIGUATOR: (${sWhat})  [${this.lTokens[nTokenOffset+1]["sValue"]}:${this.lTokens[nLastToken]["sValue"]}]`);
                                }
                            }
                            else if (cActionType == ">") {
                                // we do nothing, this test is just a condition to apply all following actions
                                if (bDebug) {
                                    console.log("    COND_OK");
                                }
                            }
                            else if (cActionType == "/") {
                                // Tag
                                let nTokenStart = (eAct[0] > 0) ? nTokenOffset + eAct[0] : nLastToken + eAct[0];
                                let nTokenEnd = (eAct[1] > 0) ? nTokenOffset + eAct[1] : nLastToken + eAct[1];
                                for (let i = nTokenStart; i <= nTokenEnd; i++) {
                                    if (this.lTokens[i].hasOwnProperty("aTags")) {
                                        this.lTokens[i]["aTags"].add(...sWhat.split("|"))
                                    } else {
                                        this.lTokens[i]["aTags"] = new Set(sWhat.split("|"));
                                    }
                                }
                                if (bDebug) {
                                    console.log(`    TAG:  ${sWhat} > [${this.lTokens[nTokenStart]["sValue"]}:${this.lTokens[nTokenEnd]["sValue"]}]`);
                                }
                                for (let sTag of sWhat.split("|")) {
                                    if (!this.dTags.has(sTag)) {
                                        this.dTags.set(sTag, [nTokenStart, nTokenEnd]);
                                    } else {
                                        this.dTags.set(sTag, [Math.min(nTokenStart, this.dTags.get(sTag)[0]), Math.max(nTokenEnd, this.dTags.get(sTag)[1])]);
                                    }
                                }
                            }
                            else if (cActionType == "!") {
                                // immunity
                                if (bDebug) {
                                    console.log("    IMMUNITY: " + sLineId + " / " + sRuleId);
                                }
                                let nTokenStart = (eAct[0] > 0) ? nTokenOffset + eAct[0] : nLastToken + eAct[0];
                                let nTokenEnd = (eAct[1] > 0) ? nTokenOffset + eAct[1] : nLastToken + eAct[1];
                                if (nTokenEnd - nTokenStart == 0) {
                                    this.lTokens[nTokenStart]["bImmune"] = true;
                                    let nErrorStart = this.nOffsetWithinParagraph + this.lTokens[nTokenStart]["nStart"];
                                    if (this.dError.has(nErrorStart)) {
                                        this.dError.delete(nErrorStart);
                                    }
                                } else {
                                    for (let i = nTokenStart;  i <= nTokenEnd;  i++) {
                                        this.lTokens[i]["bImmune"] = true;
                                        let nErrorStart = this.nOffsetWithinParagraph + this.lTokens[i]["nStart"];
                                        if (this.dError.has(nErrorStart)) {
                                            this.dError.delete(nErrorStart);
                                        }
                                    }
                                }
                            } else {
                                console.log("# error: unknown action at " + sLineId);

        return this._createError(nStart, nEnd, sLineId, sRuleId, sOption, sMessage, lSugg, sURL, bContext);
    }

    _createErrorFromTokens (sSugg, nTokenOffset, nLastToken, iFirstToken, nStart, nEnd, sLineId, sRuleId, bCaseSvty, sMsg, sURL, bShowRuleId, sOption, bContext) {
        // suggestions
        let lSugg = [];
        if (sSugg.startsWith("=")) {
            sSugg = gc_functions[sSugg.slice(1)](this.lTokens, nTokenOffset, nLastToken);
            //sSugg = oEvalFunc[sSugg.slice(1)](this.lTokens, nTokenOffset, nLastToken);
            lSugg = (sSugg) ? sSugg.split("|") : [];
        } else if (sSugg == "_") {
            lSugg = [];
        } else {
            lSugg = this._expand(sSugg, nTokenOffset, nLastToken).split("|");
        }
        if (bCaseSvty && lSugg.length > 0 && this.lTokens[iFirstToken]["sValue"].slice(0,1).gl_isUpperCase()) {
            lSugg = (this.sSentence.slice(nStart, nEnd).gl_isUpperCase()) ? lSugg.map((s) => s.toUpperCase()) : capitalizeArray(lSugg);
        }
        // Message
        let sMessage = (sMsg.startsWith("=")) ? gc_functions[sMsg.slice(1)](this.lTokens, nTokenOffset, nLastToken) : this._expand(sMsg, nTokenOffset, nLastToken);
        //let sMessage = (sMsg.startsWith("=")) ? oEvalFunc[sMsg.slice(1)](this.lTokens, nTokenOffset, nLastToken) : this._expand(sMsg, nTokenOffset, nLastToken);
        if (bShowRuleId) {
            sMessage += "  #" + sLineId + " / " + sRuleId;
        }
        //
        return this._createError(nStart, nEnd, sLineId, sRuleId, sOption, sMessage, lSugg, sURL, bContext);
    }

        return oErr;
    }

    _expand (sText, nTokenOffset, nLastToken) {
        let m;
        while ((m = /\\(-?[0-9]+)/.exec(sText)) !== null) {
            if (m[1].slice(0,1) == "-") {
                sText = sText.replace(m[0], this.lTokens[nLastToken+parseInt(m[1],10)+1]["sValue"]);
            } else {
                sText = sText.replace(m[0], this.lTokens[nTokenOffset+parseInt(m[1],10)]["sValue"]);
            }
        }
        return sText;
    }

    rewriteText (sText, sRepl, iGroup, m, bUppercase) {
        // text processor: write sRepl in sText at iGroup position"

    }

    _tagAndPrepareTokenForRewriting (sWhat, nTokenRewriteStart, nTokenRewriteEnd, nTokenOffset, nLastToken, bCaseSvty, bDebug) {
        // text processor: rewrite tokens between <nTokenRewriteStart> and <nTokenRewriteEnd> position
        if (sWhat === "*") {
            // purge text
            if (nTokenRewriteEnd - nTokenRewriteStart == 0) {
                this.lTokens[nTokenRewriteStart]["bToRemove"] = true;
            } else {
                for (let i = nTokenRewriteStart;  i <= nTokenRewriteEnd;  i++) {
                    this.lTokens[i]["bToRemove"] = true;
                }
            }
        }
        else if (sWhat === "␣") {
            // merge tokens
            this.lTokens[nTokenRewriteStart]["nMergeUntil"] = nTokenRewriteEnd;
        }
        else if (sWhat === "_") {
            // neutralized token
            if (nTokenRewriteEnd - nTokenRewriteStart == 0) {
                this.lTokens[nTokenRewriteStart]["sNewValue"] = "_";
            } else {
                for (let i = nTokenRewriteStart;  i <= nTokenRewriteEnd;  i++) {
                    this.lTokens[i]["sNewValue"] = "_";
                }
            }
        }
        else {
            if (sWhat.startsWith("=")) {
                sWhat = gc_functions[sWhat.slice(1)](this.lTokens, nTokenOffset, nLastToken);
                //sWhat = oEvalFunc[sWhat.slice(1)](this.lTokens, nTokenOffset, nLastToken);
            } else {
                sWhat = this._expand(sWhat, nTokenOffset, nLastToken);
            }
            let bUppercase = bCaseSvty && this.lTokens[nTokenRewriteStart]["sValue"].slice(0,1).gl_isUpperCase();
            if (nTokenRewriteEnd - nTokenRewriteStart == 0) {
                // one token
                if (bUppercase) {
                    sWhat = sWhat.gl_toCapitalize();
                }
                this.lTokens[nTokenRewriteStart]["sNewValue"] = sWhat;
            }
            else {
                // several tokens
                let lTokenValue = sWhat.split("|");
                if (lTokenValue.length != (nTokenRewriteEnd - nTokenRewriteStart + 1)) {
                    if (bDebug) {
                        console.log("Error. Text processor: number of replacements != number of tokens.");
                    }
                    return;
                }
                let j = 0;
                for (let i = nTokenRewriteStart;  i <= nTokenRewriteEnd;  i++) {
                    let sValue = lTokenValue[j];
                    if (!sValue || sValue === "*") {
                        this.lTokens[i]["bToRemove"] = true;
                    } else {
                        if (bUppercase) {
                            sValue = sValue.gl_toCapitalize();
                        }
                        this.lTokens[i]["sNewValue"] = sValue;
                    }
                    j++;
                }
            }
        }
    }

    rewriteFromTags (bDebug=false) {
        // rewrite the sentence, modify tokens, purge the token list
        if (bDebug) {
            console.log("REWRITE");
        }
        let lNewToken = [];
        let lNewTokens0 = [];
        let nMergeUntil = 0;
        let oMergingToken = null;
        for (let [iToken, oToken] of this.lTokens.entries()) {
            let bKeepToken = true;
            let bKeepToken0 = true;
            if (oToken["sType"] != "INFO") {
                if (nMergeUntil && iToken <= nMergeUntil) {
                    oMergingToken["sValue"] += " ".repeat(oToken["nStart"] - oMergingToken["nEnd"]) + oToken["sValue"];
                    oMergingToken["nEnd"] = oToken["nEnd"];
                    if (bDebug) {
                        console.log("  MERGED TOKEN: " + oMergingToken["sValue"]);
                    }
                    bKeepToken = false;
                    bKeepToken0 = false;
                }
                if (oToken.hasOwnProperty("nMergeUntil")) {
                    if (iToken > nMergeUntil) { // this token is not already merged with a previous token
                        oMergingToken = oToken;
                    }
                    if (oToken["nMergeUntil"] > nMergeUntil) {
                        nMergeUntil = oToken["nMergeUntil"];

                    this.dTokenPos.delete(oToken["nStart"]);
                }
                catch (e) {
                    console.log(this.asString());
                    console.log(oToken);
                }
            }
            if (this.lTokens0 !== null && bKeepToken0) {
                lNewTokens0.push(oToken);
            }
        }
        if (bDebug) {
            console.log("  TEXT REWRITED: " + this.sSentence);
        }
        this.lTokens.length = 0;
        this.lTokens = lNewToken;
        if (this.lTokens0 !== null) {
            this.lTokens0.length = 0;
            this.lTokens0 = lNewTokens0;
        }
    }
};


if (typeof(exports) !== 'undefined') {
    exports.lang = gc_engine.lang;
    exports.locales = gc_engine.locales;