Laboratoire ICAR » Plateforme TXM

// Copyright © 2021 ENS Lyon

// Licensed under the terms of the GNU General Public License version 3 (http://www.gnu.org/licenses/gpl-3.0.html)

// @author mdecorde

// @author sheiden

package org.txm.macro.annotation

// STANDARD DECLARATIONS

import org.txm.rcpapplication.swt.widget.parameters.*

import org.txm.searchengine.cqp.corpus.*

import org.txm.xml.DOMIdentityHook

import org.txm.xml.LocalNameHookActivator

import org.txm.xml.XMLProcessor

import org.txm.objects.*

import org.txm.rcp.utils.JobHandler

import org.txm.rcp.views.corpora.CorporaView

import org.eclipse.core.runtime.Status

import org.txm.annotation.kr.core.KRAnnotationEngine

import org.txm.annotation.kr.core.repository.*

import org.txm.annotation.kr.rcp.commands.SaveAnnotations

import org.txm.annotation.kr.rcp.concordance.WordAnnotationToolbar

import org.txm.importer.StaxIdentityParser

def scriptName = this.class.getSimpleName()

if (!(corpusViewSelection instanceof MainCorpus)) {

	println "** $scriptName: please select a Corpus in the Corpus view."

	return 0

}

MainCorpus mcorpus = corpusViewSelection

Project project = mcorpus.getProject()

File txmDir = new File(project.getProjectDirectory(), "txm/"+mcorpus.getName())

@Field @Option(name="nWordsPerText", usage="the number of words per text", widget="Integer", required=true, def="")

File nWordsPerText

@Field @Option(name="debug", usage="Afficher les messages de debug", widget="Boolean", required=true, def="false")

def debug

// Open the parameters input dialog box

if (!ParametersDialog.open(this)) return

if (!outputDirectory.exists()) {

	println "Error: output directory not found: "+outputDirectory

	return false

}

values_to_ignore_regex = /$values_to_ignore_regex/

project.compute();

for (Text text : project.getChildren(Text.class)) {

	println "Text: "+text

	File orig = text.getXMLTXMFile()

	if ((orig == null) || !orig.exists()) {

		println "** Error: no XML-TXM file found for "+text

		return // next !

	}

	File result = new File(outputDirectory, orig.getName())

	if (false && result.lastModified() > orig.lastModified()) {

		println "Skipping $text: result file is more recent: $result"

		continue // next !

	}

	XMLProcessor xp = new XMLProcessor(orig);

	LocalNameHookActivator activator = new LocalNameHookActivator("w")

	nWords = 0

	new DOMIdentityHook("word", activator, xp) {

		@Override

		public void processDom() {

			//println dom

			use(groovy.xml.dom.DOMCategory) {

				nWords++

				if (nWords >  nWordsPerText) {

					dom = null

				}

			}

		}

	}

	xp.process(result)

	if (nWords == 0) {

		println "Removing empty <text>: $text: $result"

		result.delete()

	}

}

println "Done: update the corpus using the 'Corpus > Update' main menu to see the modifications in commnds."