Laboratoire ICAR » Plateforme TXM

import org.txm.utils.FileUtils

	def primarySpeakerIdRegex

	String primarySpeakerId

	Boolean debug

	String otherNonPrimaryLocutor = "other"

	public FixTranscription(File xmlfile, String primarySpeakerIdPrefix, String otherNonPrimaryLocutor, Boolean debug) {

		this.debug = debug

		this.otherNonPrimaryLocutor = otherNonPrimaryLocutor

		if (primarySpeakerIdPrefix != null && primarySpeakerIdPrefix.length() > 0) {

			String id = FileUtils.stripExtension(xmlfile)

			this.primarySpeakerIdRegex = /$primarySpeakerIdPrefix.*/

			def rez = (id =~ primarySpeakerIdRegex).findAll()

			def rez2 = (id =~ /$primarySpeakerIdPrefix/).findAll()

			if (rez2.size() != 1) {

				if (debug) println "WARNING: found the ${rez2.size()} matches of primary speaker prefix in the '$id' file name"

				this.primarySpeakerIdRegex = null

			} else {

				primarySpeakerId = rez[0]

				//if (debug) println "Detected primary speaker: $primarySpeakerId"

			}

		}

								if (primarySpeakerIdRegex == null || turnInfos["speaker"] ==~ primarySpeakerIdRegex) {

									tmpInfos["speaker"] = "other"

								} else {

									tmpInfos["speaker"] = primarySpeakerId

								}

		File trsFile

		File trsDirectory

		File resultDirectory

@Field @Option(name="primarySpeakerIdPrefix", usage="speaker ID of the primary speaker", widget="String", required=false, def="")

		String primarySpeakerIdPrefix

		@Field @Option(name="otherNonPrimaryLocutor", usage="other non primary id of the other turns", widget="String", required=false, def="")

		String otherNonPrimaryLocutor

@Field @Option(name="debug", usage="speaker ID of the primary speaker", widget="Boolean", required=false, def="false")

		Boolean debug

	FixTranscription fixer = new FixTranscription(file, primarySpeakerIdPrefix, otherNonPrimaryLocutor, debug)

package org.txm.macro.transcription

import java.nio.charset.Charset

import java.time.LocalTime

import java.time.format.DateTimeFormatter

import org.txm.utils.*

import org.txm.utils.logger.*

@Field @Option(name="trsDirectory", usage="Dossier qui contient les fichiers TRS", widget="Folder", required=true, def="")

File trsDirectory;

@Field @Option(name="outputTrsDirectory", usage="Dossier résultat qui contient les fichiers TRS", widget="Folder", required=true, def="")

File outputTrsDirectory;

@Field @Option(name="idRegex", usage="Colonne de jointure de transcription", widget="String", required=true, def="")

def idRegex

@Field @Option(name="nameRegex", usage="Colonne de jointure de transcription", widget="String", required=true, def="")

def nameRegex

@Field @Option(name="newID", usage="Colonne de jointure de transcription", widget="String", required=true, def="")

def newID

@Field @Option(name="newName", usage="Colonne de jointure de transcription", widget="String", required=true, def="")

def newName

def files = trsDirectory.listFiles()

if (files == null) {

	println "Error: no files in $trsDirectory"

	return false

}

outputTrsDirectory.mkdirs()

if (!outputTrsDirectory.exists()) {

	println "Error: can't create $outputTrsDirectory"

	return false;

}

boolean ok = true

for (File trsFile : files) {

	if (!trsFile.getName().toLowerCase().endsWith(".trs")) continue;

	File outputTrsFile = new File(outputTrsDirectory, trsFile.getName());

	ok = ok && gse.runMacro(RenameSpeakerMacro, ["trsFile":trsFile, "outputTrsFile":outputTrsFile, "idRegex":idRegex, "nameRegex":nameRegex, "newID":newID, "newName":newName])

}

return ok

package org.txm.macro.transcription

import org.txm.scripts.importer.*;

import org.txm.utils.*;

import org.txm.metadatas.*;

import java.io.File;

import org.w3c.dom.Document;

import org.w3c.dom.Element;

import javax.xml.parsers.*;

import javax.xml.xpath.*;

import javax.xml.transform.*;

import javax.xml.transform.dom.DOMSource;

import javax.xml.transform.stream.StreamResult;

/** 

 * Renames speakers given an id or name regex

 **/

public class ChangeSpeaker {

	File outfile, transcriptionfile

	/** The doc. */

	Document doc

	def idRegex, nameRegex

	String newId, newName

	/**

	 * Instantiates a new change speaker.

	 *

	 * @param transcriptionfile the transcriptionfile

	 * @param outfile the outfile

	 * @param id the id

	 * @param newid the newid

	 */

	public ChangeSpeaker(File transcriptionfile, File outfile, String idRegexString, String nameRegexString, String newId, String newName) {

		this.transcriptionfile = transcriptionfile

		this.outfile = outfile;

		if (idRegexString != null && idRegexString.length() > 0) {

			this.idRegex = /$idRegexString/

		}

		if (nameRegexString != null && nameRegexString.length() > 0) {

			this.nameRegex = /$nameRegexString/

		}

		this.newId = newId

		this.newName = newName

	}

	/**

	 * Save.

	 *

	 * @return true, if successful

	 */

	public boolean process() {

		try {

			DocumentBuilderFactory domFactory = DocumentBuilderFactory.newInstance();

			domFactory.setNamespaceAware(true); // never forget this!

			domFactory.setXIncludeAware(true);

			DocumentBuilder builder = domFactory.newDocumentBuilder();

			doc = builder.parse(transcriptionfile);

			def expr2 = XPathFactory.newInstance().newXPath().compile("//Speaker");

			def nodes2 = expr2.evaluate(doc, XPathConstants.NODESET);

			// fix speaker declarations

			def replacedIds = [] // list of IDs replaced, the list is used later to update the Turn locutors ids

			for (def node : nodes2) {

				if (node == null) continue;

				Element elem = (Element)node;

				String id = elem.getAttribute("id")

				String name = elem.getAttribute("name")

				if (idRegex != null && id =~ idRegex) { // patch Speaker@id

					if (newId != null) {

						elem.setAttribute("id", newId);

						replacedIds << id

					}

					if (newName != null) {

						elem.setAttribute("name", newName);

					}

				}

				if (nameRegex != null && id =~ nameRegex) { // patch Speaker@name

					if (newName != null) {

						elem.setAttribute("name", newName);

					}

				}

			}

			def expr = XPathFactory.newInstance().newXPath().compile("//Turn");

			def nodes = expr.evaluate(doc, XPathConstants.NODESET);

			// fix speaker turns

			for (def node : nodes) {

				if (node == null) continue;

				Element elem = (Element)node;

				String id = elem.getAttribute("speaker");

				if (replacedIds.contains(id)) {

					elem.setAttribute("speaker", newId);

				}

			}

			// Création de la source DOM

			Source source = new DOMSource(doc);

			// Création du fichier de sortie

			Writer writer = new BufferedWriter(new OutputStreamWriter(new FileOutputStream(outfile), "UTF-8"));

			Result resultat = new StreamResult(writer);

			// Configuration du transformer

			TransformerFactory fabrique = new net.sf.saxon.TransformerFactoryImpl();

			Transformer transformer = fabrique.newTransformer();

			transformer.setOutputProperty(OutputKeys.METHOD, "xml");

			transformer.setOutputProperty(OutputKeys.INDENT, "yes");

			transformer.setOutputProperty(OutputKeys.ENCODING, "UTF-8");

			// Transformation

			transformer.transform(source, resultat);

			writer.close();

			return true;

		} catch (Exception e) {

			org.txm.utils.logger.Log.printStackTrace(e);

			return false;

		}

	}

	/**

	 * The main method.

	 *

	 * @param args the arguments

	 */

	public static void main(String[] args) {

		String homedir = System.getProperty("user.home")

		File trs1 = new File(homedir, "xml/concattrs/int18_1.trs")

		File trs2 = new File(homedir, "xml/concattrs/int18_1-renamed.trs")

		new RenameSpeaker(trs1, trs2, "spk2", null, "spk4", null).process();

	}

}

package org.txm.macro.transcription

import java.nio.charset.Charset

import java.time.LocalTime

import java.time.format.DateTimeFormatter

import org.txm.utils.*

import org.txm.utils.logger.*

@Field @Option(name="trsFile", usage="Dossier qui contient les fichiers TRS", widget="FileOpen", required=true, def="")

File trsFile;

@Field @Option(name="outputTrsFile", usage="Dossier qui contient les fichiers TRS", widget="FileSave", required=true, def="")

File outputTrsFile;

@Field @Option(name="idRegex", usage="Colonne de jointure de transcription", widget="String", required=true, def="")

def idRegex

@Field @Option(name="nameRegex", usage="Colonne de jointure de transcription", widget="String", required=true, def="")

def nameRegex

@Field @Option(name="idRegex", usage="Colonne de jointure de transcription", widget="String", required=true, def="")

def newID

@Field @Option(name="nameRegex", usage="Colonne de jointure de transcription", widget="String", required=true, def="")

def newName

def cs = new RenameSpeaker(trsFile, outputTrsFile, idRegex, nameRegex, newID, newName);

return cs.process();