Laboratoire ICAR » Plateforme TXM

// Copyright © 2010-2013 ENS de Lyon.

// Copyright © 2007-2010 ENS de Lyon, CNRS, INRP, University of

// Lyon 2, University of Franche-Comté, University of Nice

// Sophia Antipolis, University of Paris 3.

//

// The TXM platform is free software: you can redistribute it

// and/or modify it under the terms of the GNU General Public

// License as published by the Free Software Foundation,

// either version 2 of the License, or (at your option) any

// later version.

//

// The TXM platform is distributed in the hope that it will be

// useful, but WITHOUT ANY WARRANTY; without even the implied

// warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR

// PURPOSE. See the GNU General Public License for more

// details.

//

// You should have received a copy of the GNU General

// Public License along with the TXM platform. If not, see

// http://www.gnu.org/licenses.

//

// This file is part of the TXM platform.

//

// The TXM platform is free software: you can redistribute it and/or modif y

// it under the terms of the GNU General Public License as published by

// the Free Software Foundation, either version 3 of the License, or

// (at your option) any later version.

//

// The TXM platform is distributed in the hope that it will be useful,

// but WITHOUT ANY WARRANTY; without even the implied warranty of

// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the

// GNU General Public License for more details.

//

// You should have received a copy of the GNU General Public License

// along with the TXM platform.  If not, see <http://www.gnu.org/licenses/>.

//

//

//

// $LastChangedDate: 2017-05-02 11:55:17 +0200 (mar. 02 mai 2017) $

// $LastChangedRevision: 3436 $

// $LastChangedBy: mdecorde $

//

package org.txm.scripts.importer.bfm;

import org.txm.scripts.importer.*;

import org.xml.sax.Attributes;

import org.txm.importer.scripts.filters.*;

import java.io.File;

import java.io.IOException;

import java.util.ArrayList;

import javax.xml.parsers.SAXParserFactory;

import javax.xml.parsers.ParserConfigurationException;

import javax.xml.parsers.SAXParser;

import javax.xml.stream.*;

import java.net.URL;

import org.xml.sax.InputSource;

import org.xml.sax.helpers.DefaultHandler;

// TODO: Auto-generated Javadoc

/**

 * Build BFM texts HTML edition from the TEI-TXM files. <br/>

 * The result is similar to the one of the XSLT script made by Alexis Lavrentiev

 * (BFM & ICAR3). <br/>

 *

 * @author mdecorde

 */

class pager {

        private List<String> NoSpaceBefore;

        /** The No space after. */

        private List<String> NoSpaceAfter;

        /** The wordcount. */

        private int wordcount = 0;

        /** The pagecount. */

        private int pagecount = 0;

        /** The wordmax. */

        private int wordmax = 0;

        /** The wordid. */

        private String wordid;

        /** The first word. */

        private boolean firstWord = true;

        /** The wordvalue. */

        private String wordvalue;

        /** The interpvalue. */

        private String interpvalue;

        /** The lastword. */

        private String lastword = " ";

        /** The wordtype. */

        private String wordtype;

        /** The flagform. */

        private boolean flagform = false;

        /** The flaginterp. */

        private boolean flaginterp = false;

        /** The url. */

        private URL url;

        /** The input data. */

        private def inputData;

        /** The factory. */

        private def factory;

        /** The parser. */

        private XMLStreamReader parser;

        /** The writer. */

        private OutputStreamWriter writer;

        /** The multiwriter. */

        private OutputStreamWriter multiwriter = null;

        StaxStackWriter pagedWriter = null;

        /** The infile. */

        private File infile;

        /** The outfile. */

        private File outfile;

        private File outDir;

        /** The pages. */

        ArrayList<File> pages = new ArrayList<File>();

        /** The idxstart. */

        ArrayList<String> idxstart = new ArrayList<String>();

        /** The titre id. */

        private String titreId;

        private String basename;

        private String textname;

        private def xpathProperties;

        /**

         * Instantiates a new pager.

         *

         * @param infile the file to convert

         * @param outfile the result file

         * @param NoSpaceBefore the punctuation marks that don't have a space before

         * @param NoSpaceAfter the punctuation marks that don't have a space after

         * @param max the max number of word per page

         */

        pager(File infile, File outDir, String textname, List<String> NoSpaceBefore,

        List<String> NoSpaceAfter, int max, String basename, Properties xpathprops) {

                this.wordmax = max;

                this.NoSpaceBefore = NoSpaceBefore;

                this.NoSpaceAfter = NoSpaceAfter;

                this.basename = basename;

                this.textname = textname;

                this.outDir = outDir;

                this.xpathProperties = xpathprops;

                this.url = infile.toURI().toURL();

                this.infile = infile;

                inputData = url.openStream();

                factory = XMLInputFactory.newInstance();

                parser = factory.createXMLStreamReader(inputData);

                process();

        }

        private def closeMultiWriter()

        {

                if (pagedWriter != null) {

                        def tags = pagedWriter.getTagStack().clone();

                        if (firstWord) { // there was no words

                                this.idxstart.add("w_0")

                                pagedWriter.write("<span id=\"w_0\"/>");

                        }

                        pagedWriter.writeEndElements();

                        pagedWriter.close();

                        return tags;

                } else {

                        return [];

                }

        }

        /**

         * Creates the next output.

         *

         * @return true, if successful

         */

        private boolean createNextOutput()

        {

                wordcount = 0;

                try {

                        def tags = closeMultiWriter();

                        for (int i = 0 ; i < tags.size() ; i++) {

                                String tag = tags[i]

                                if ("body" != tag) {

                                        tags.remove(i--)

                                } else {

                                        tags.remove(i--) // remove "body"

                                        break; // remove elements until "body tag

                                }

                        }

                        File outfile = new File(outDir, textname+"_"+(++pagecount)+".html")

                        pages.add(outfile);

                        firstWord = true; // waiting for next word

                        pagedWriter = new StaxStackWriter(outfile, "UTF-8");

                        pagedWriter.writeStartDocument("UTF-8", "1.0")

                        pagedWriter.writeStartElement("html")

                        pagedWriter.writeEmptyElement("meta", ["http-equiv":"Content-Type", "content":"text/html","charset":"UTF-8"]);

                        pagedWriter.writeStartElement("head")

                        pagedWriter.writeStartElement("title")

                        pagedWriter.writeCharacters(basename.toUpperCase()+" - "+textname+" - Edition - Page "+pagecount)

                        pagedWriter.writeEndElement() // </title>

                        pagedWriter.writeEndElement() // </head>

                        pagedWriter.writeStartElement("body") //<body>

                        pagedWriter.writeStartElements(tags);

                        return true;

                } catch (Exception e) {

                        System.out.println(e.getLocalizedMessage());

                        return false;

                }

        }

        /**

         * Creates the output.

         *

         * @param outfile the outfile

         * @return true, if successful

         */

        private boolean createOutput(File outfile) {

                try {

                        return createNextOutput();

                } catch (Exception e) {

                        System.out.println(e.getLocalizedMessage());

                        return false;

                }

        }

        /**

         * Gets the page files.

         *

         * @return the pages create during the processing

         */

        public ArrayList<File> getPageFiles() {

                return pages;

        }

        /**

         * Gets the idx.

         *

         * @return the word's indexes

         */

        public ArrayList<String> getIdx() {

                return idxstart;

        }

        /**

         * process the infile.

         */

        void process() {

                String localname = "";

                boolean closeTitre = false;

                boolean closeAmen = false;

                boolean flagchoice = false;

                boolean flagcorr = false;

                boolean flagsic = false;

                boolean flagreg = false;

                boolean flagexpan = false;

                boolean flagorig = false;

                boolean flagabbr = false;

                boolean flagforeign = false;

                int levelSupplied = 0;

                boolean flagSurplus = false;

                boolean flagDel = false;

                // option for metadata from BFM TEI header

                def mValues = [:];

                def xpathprocessor = new XPathResult(infile)

                for (String name : xpathProperties.keySet()) {

                        String value = xpathprocessor.getXpathResponse(xpathProperties.get(name), "N/A");

                        mValues.put(name, value.replace("\n", " "));

                }

                if (mValues.containsKey("forme") && !mValues.get("forme").equals("N/A")) {

                        if (mValues.get("forme").startsWith("#forme_")) {

                                mValues.put("forme", mValues.get("forme").substring(7));

                        } else if (mValues.get("forme").startsWith("#")) {

                                mValues.put("forme", mValues.get("forme").substring(1));

                        } else {

                                mValues.put("forme", mValues.get("forme"));

                        }

                }

                createNextOutput();

                for (int event = parser.next(); event != XMLStreamConstants.END_DOCUMENT; event = parser.next())

                {

                        switch (event) {

                                case XMLStreamConstants.START_ELEMENT:

                                        localname = parser.getLocalName();

                                        switch (localname) {

                                                case "text":

                                                        pagedWriter.writeStartElement("h1",

                                                        ["align":"center", "style":"color:darkviolet;font-family:Times;font-weight:bold;font-style:italic;font-size=200%", "class":"text"])

                                                        pagedWriter.writeCharacters(mValues.get("titre"))

                                                        pagedWriter.writeEndElement(); // h1

                                                        pagedWriter.writeStartElement("p")

                                                        pagedWriter.writeEmptyElement("hr")

                                                        pagedWriter.writeCharacters("auteur : "+mValues.get("auteur"))

                                                        pagedWriter.writeEmptyElement("br");

                                                        pagedWriter.writeCharacters("date de composition : "+mValues.get("datecompolibre"));

                                                        pagedWriter.writeEmptyElement("br");

                                                        pagedWriter.writeCharacters("domaine : "+mValues.get("domaine"))

                                                        pagedWriter.writeEmptyElement("br");

                                                        pagedWriter.writeCharacters("genre : "+mValues.get("genre"))

                                                        pagedWriter.writeEmptyElement("br");

                                                        pagedWriter.writeCharacters("forme : "+mValues.get("forme"))

                                                        pagedWriter.writeEmptyElement("br");

                                                        pagedWriter.writeCharacters("dialecte : "+mValues.get("dialecte"))

                                                        pagedWriter.writeEmptyElement("br");

                                                        pagedWriter.writeEmptyElement("hr")

                                                        pagedWriter.writeEndElement(); // p

                                                        break;

                                                case "head":

                                                        pagedWriter.writeStartElement("h2")

                                                        break;

                                                case "lg":

                                                case "p":

                                                        pagedWriter.writeStartElement("p")

                                                        break;

                                                case "pb":

                                                        createNextOutput();

                                                        if (parser.getAttributeValue(null,"n") != null) {

                                                                pagedWriter.writeStartElement("p", ["style":"color:red", "align":"center"]);

                                                                pagedWriter.writeCharacters("- "+parser.getAttributeValue(null,"n")+" -")

                                                                pagedWriter.writeEndElement() // p

                                                        }

                                                        break;

                                                case "ab":

                                                        if (parser.getAttributeValue(null,"n") != null) {

                                                                pagedWriter.writeStartElement("p", ["align":"center"]);

                                                                pagedWriter.writeCharacters(parser.getAttributeValue(null,"n"))

                                                                pagedWriter.writeEndElement() // p

                                                        }

                                                        break;

                                                case "l":

                                                        pagedWriter.writeEmptyElement("br");

                                                        break;

                                                case "div":

                                                        if (parser.getAttributeValue(null,"type") == "titre") {

                                                                pagedWriter.writeStartElement("h3")

                                                                closeTitre = true;

                                                        } else if (parser.getAttributeValue(null,"type") == "amen") {

                                                                pagedWriter.writeStartElement("h3")

                                                                closeAmen = true;

                                                        }

                                                        break;

                                                case "w":

                                                        interpvalue = "";

                                                        for (int i = 0 ; i < parser.getAttributeCount(); i++) {

                                                                if (parser.getAttributeLocalName(i).equals("type")) {

                                                                        interpvalue = " pos: "+parser.getAttributeValue(i);

                                                                }

                                                        }

                                                        wordid = parser.getAttributeValue(null, "id");

                                                        wordcount++;

                                                        if (wordcount >= wordmax) {

                                                                createNextOutput();

                                                        }

                                                        if (firstWord) {

                                                                firstWord = false;

                                                                this.idxstart.add(wordid);

                                                        }

                                                        break;

                                                case "choice":

                                                        flagchoice = true;

                                                        break;

                                                case "corr":

                                                        flagcorr = true;

                                                        break;

                                                case "sic":

                                                        flagsic = true;

                                                        break;

                                                case "reg":

                                                        flagreg = true;

                                                        break;

                                                case "orig":

                                                        flagorig = true;

                                                        break;

                                                case "foreign":

                                                        flagforeign = true;

                                                        break;

                                                case "supplied":

                                                        levelSupplied = levelSupplied +1;

                                                        if(flagform)

                                                                wordvalue = wordvalue+"[";

                                                        break;

                                                case "surplus":

                                                        flagSurplus = true;

                                                        pagedWriter.writeStartElement("span", ["class": "surplus", "style":"color:red;"])

                                                        break;

                                                case "del":

                                                        flagDel = true;

                                                        pagedWriter.writeStartElement("span", ["class": "del", "style":"color:red;text-decoration:line-through;"])

                                                        break;

                                                case "ana":

                                                        flaginterp=true;

                                                        interpvalue+=" "+parser.getAttributeValue(null,"type").substring(1)+":"

                                                        break;

                                                case "form":

                                                        wordvalue="";

                                                        flagform=true;

                                        }

                                        break;

                                case XMLStreamConstants.END_ELEMENT:

                                        localname = parser.getLocalName();

                                        switch (localname) {

                                                case "head":

                                                        pagedWriter.writeEndElement() // h2

                                                        break;

                                                case "div":

                                                        if (closeTitre) {

                                                                pagedWriter.writeEndElement() // h3

                                                                closeTitre = false;

                                                        } else if(closeAmen) {

                                                                pagedWriter.writeEndElement() // h3

                                                                closeTitre = false;

                                                        }

                                                        break;

                                                case "lb":

                                                        pagedWriter.writeEmptyElement("br")

                                                        break;

                                                case "lg":

                                                case "p":

                                                        pagedWriter.writeEndElement() // p

                                                        break;

                                                case "choice":

                                                        flagchoice = false;

                                                        break;

                                                case "corr":

                                                        flagcorr = false;

                                                        break;

                                                case "sic":

                                                        flagsic = false;

                                                        break;

                                                case "reg":

                                                        flagreg = false;

                                                        break;

                                                case "orig":

                                                        flagorig = false;

                                                        break;

                                                case "foreign":

                                                        flagforeign = false;

                                                        break;

                                                case "supplied":

                                                        levelSupplied = levelSupplied -1;

                                                        if (flagform)

                                                                wordvalue = wordvalue+"]";

                                                        break;

                                                case "surplus":

                                                        flagSurplus = false;

                                                        pagedWriter.writeCharacters(")")

                                                        pagedWriter.writeEndElement() // span

                                                        break;

                                                case "del":

                                                        flagDel = false;

                                                        pagedWriter.writeEndElement() // span

                                                        break;

                                                case "form":

                                                        flagform = false

                                                        break;

                                                case "ana":

                                                        flaginterp = false

                                                        break;

                                                case "w":

                                                        int l = lastword.length();

                                                        String color = "";

                                                        if (flagcorr)

                                                                color = "color: green;";

                                                        else if (flagreg)

                                                                color = "color: darkgreen;";

                                                        else if (flagforeign)

                                                                color = "color: darkred;";

                                                        else if (levelSupplied == 1)

                                                                color = "color: blue;";

                                                        else if (levelSupplied == 2)

                                                                color = "color: darkblue;";

                                                        if (!flagchoice || flagcorr || flagreg) {

                                                                String endOfLastWord = "";

                                                                if (l > 0)

                                                                        endOfLastWord = lastword.subSequence(l-1, l);

                                                                if (NoSpaceBefore.contains(wordvalue) ||

                                                                NoSpaceAfter.contains(lastword) ||

                                                                wordvalue.startsWith("-") ||

                                                                NoSpaceAfter.contains(endOfLastWord))

                                                                {

                                                                        //multiwriter.write("<span title=\""+interpvalue+"\" id=\""+wordid+"\" "+color+">");

                                                                } else {

                                                                        //multiwriter.write(" <span title=\""+interpvalue+"\" id=\""+wordid+"\" "+color+">");

                                                                        pagedWriter.writeCharacters(" ")

                                                                }

                                                                pagedWriter.writeStartElement("span", ["title":interpvalue, "id":wordid, "style":color]) // span

                                                                pagedWriter.writeCharacters(wordvalue)

                                                                pagedWriter.writeEndElement() // span

                                                                lastword=wordvalue;

                                                        }

                                                        break;

                                        }

                                        break;

                                case XMLStreamConstants.CHARACTERS:

                                        if (flagform && parser.getText().length() > 0)

                                                        wordvalue += (parser.getText());

                                        if (flaginterp && parser.getText().length() > 0)

                                                        interpvalue += (parser.getText());

                                        break;

                        }

                }

                if (parser != null) parser.close();

                if (inputData != null) inputData.close();

                closeMultiWriter();

        }

}