Laboratoire ICAR » Plateforme TXM

// Copyright © 2018 ENS de Lyon, CNRS, University of Franche-Comté

// Licensed under the terms of the GNU General Public License (http://www.gnu.org/licenses)

// @author sheiden

/*

This macro calls NumberElement successively

*/

package org.txm.macro.xml

res = gse.run(NumberElementMacro, [

	"args":[

		"inputDirectory":new File("/home/sheiden/Documents/projet-susana-mars-reers/corpus/tei"),

		"outputDirectory":new File("/home/sheiden/Documents/projet-susana-mars-reers/corpus/tei/out"),

		"elementName":"pb",

		"elementAttribute":"n",

		"countStart":94,

		"valuePrefix":"",

		"valueSuffix":"",

		"debug":"true"

	],

				"selection":selection,

				"selections":selections,

				"corpusViewSelection":corpusViewSelection,

				"corpusViewSelections":corpusViewSelections,

				"monitor":monitor])

if (!res) println "** problem calling NumberElementMacro."

res = gse.run(NumberElementMacro, [

	"args":["inputDirectory":new File("/home/sheiden/Documents/projet-susana-mars-reers/corpus/tei/out"),

	"outputDirectory":new File("/home/sheiden/Documents/projet-susana-mars-reers/corpus/tei/out/out"),

	"elementName":"pb",

	"elementAttribute":"facs",

	"countStart":102,

	"valuePrefix":"https://gallica.bnf.fr/iiif/ark:/12148/bpt6k35936/f",

	"valueSuffix":"/full/full/0/native.jpg",

	"debug":"true"

	],

				"selection":selection,

				"selections":selections,

				"corpusViewSelection":corpusViewSelection,

				"corpusViewSelections":corpusViewSelections,

				"monitor":monitor])

if (!res) println "** problem calling NumberElementMacro."

package org.txm.macro.xml

// STANDARD DECLARATIONS

import org.kohsuke.args4j.*

import groovy.transform.Field

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

import org.txm.importer.ApplyXsl2

// BEGINNING OF PARAMETERS

@Field @Option(name="inputDirectory", usage="an example folder", widget="Folder", required=true, def="in")

def inputDirectory = new File(System.getProperty("user.home"),"/tmp/xsltest")

@Field @Option(name="outputDirectory", usage="an example folder", widget="Folder", required=true, def="out")

def outputDirectory = new File("/tmp/xsltest/out")

@Field @Option(name="omitXmlDeclaration", usage="omit xml declaration", widget="Boolean", required=false, def="false")

def omitXmlDeclaration

@Field @Option(name="suppressIndentation", usage="space delimited list of elements not to indent", widget="Text", required=false, def="")

def suppressIndentation

@Field @Option(name="doubleSpace", usage="space delimited list of elements with forced line break before", widget="Text", required=false, def="")

def doubleSpace

@Field @Option(name="encoding", usage="output character encoding", widget="String", required=false, def="utf-8")

def encoding

@Field @Option(name="namespacesDeclaration", usage="space delimited list of namespace declarations (eg \"xmlns:dc='http://purl.org/dc/elements/1.1/'\")", widget="Text", required=false, def="")

def namespacesDeclaration

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

def debug

if (!ParametersDialog.open(this)) return

// END OF PARAMETERS

doubleSpace = doubleSpace.trim()

if (doubleSpace.length() > 0) doubleSpace = " saxon:double-space=\"$doubleSpace\""

suppressIndentation = suppressIndentation.trim()

if (suppressIndentation.length() > 0) suppressIndentation = " saxon:suppress-indentation=\"$suppressIndentation\""

encoding = encoding.trim()

if (encoding.length() > 0) encoding = " encoding=\"$encoding\""

if (omitXmlDeclaration) omitXmlDeclarationP = " omit-xml-declaration=\"yes\"" else omitXmlDeclarationP = " omit-xml-declaration=\"no\""

namespacesDeclaration = namespacesDeclaration.trim()

if (namespacesDeclaration.length() > 0) namespacesDeclaration = " $namespacesDeclaration"

def xslParameters = ""

IndentStyleSheet = """

<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform" xmlns:saxon="http://saxon.sf.net/"${namespacesDeclaration}>

  <xsl:output method="xml" indent="yes"${omitXmlDeclarationP}${encoding}${doubleSpace}${suppressIndentation}/>

	<!-- saxon:indent-spaces="6" for Saxon-PE only -->

  <xsl:template match="node()|@*">

    <xsl:copy>

      <!-- Including any attributes it has and any child nodes -->

      <xsl:apply-templates select="@*|node()"/>

    </xsl:copy>

  </xsl:template>

</xsl:stylesheet>"""

def XSLFile = File.createTempFile("IndentStyleSheet",".xsl")

XSLFile.deleteOnExit()

XSLFile << IndentStyleSheet

outputDirectory.mkdir()

xslParameters = xslParameters.trim()

if (xslParameters.length() == 0) {

	xslParameters = null

} else {

	def split = xslParameters.split("\n")

	xslParameters = []

	for (def str : split) {

		def split2 = str.split("=", 2)

		if (split2.size() == 2) {

			xslParameters << split2

		}

	}

}

println "Processed directory: $inputDirectory."

def files = [] 

ApplyXsl2 a = new ApplyXsl2(XSLFile.getAbsolutePath())

inputDirectory.eachFileMatch(~/.+\.(xml|XML)/) { XMLFile ->

	String name = XMLFile.getName()

	try {

		def outFile = new File(outputDirectory, name)

		if (xslParameters != null) {

			for(def param : xslParameters) {

				a.setParam(param[0], param[1])

			}

		}

		a.process(XMLFile, outFile)

		a.resetParams()

		files << XMLFile

	} catch (Exception e) {

			msgParse = (e.getMessageAndLocation() =~ /^(.*) lineNumber: ([0-9]+); columnNumber: ([0-9]+); (.*).$/)

			if (msgParse.size() == 1) {

				def lineNumber = msgParse[0][2]

				def colNumber = msgParse[0][3]

				def errMsg = msgParse[0][4]

				println "** $name processing aborted, [line $lineNumber, character $colNumber]: $errMsg."

			} else {

				println e.getMessageAndLocation()

			}

		XSLFile.delete()

		return 0

	}

}

XSLFile.delete()

return 1

// Copyright © 2018 ENS de Lyon, CNRS, University of Franche-Comté

// Licensed under the terms of the GNU General Public License (http://www.gnu.org/licenses)

// @author sheiden

package org.txm.macro.xml

// STANDARD DECLARATIONS

import java.nio.file.Files

import org.kohsuke.args4j.*

import groovy.transform.Field

import org.txm.rcpapplication.commands.*

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

import org.txm.importer.ApplyXsl2

import net.sf.saxon.TransformerFactoryImpl

import org.txm.utils.saxon.SaxonNodeSet

import javax.xml.transform.stream.StreamSource

import javax.xml.transform.stream.StreamResult

import org.eclipse.core.filesystem.EFS

import org.eclipse.ui.PlatformUI

import org.eclipse.ui.ide.IDE

import org.eclipse.jface.text.ITextOperationTarget

// BEGINNING OF PARAMETERS

@Field @Option(name="inputDirectory", usage="an example folder", widget="Folder", required=true, def="in")

def inputDirectory = new File(System.getProperty("user.home"),"/tmp/xsltest")

@Field @Option(name="edit", usage="edit XML files with errors", widget="Boolean", required=true, def="true")

def edit

@Field @Option(name="maxEditors", usage="maximum number of files opened", widget="Integer", required=true, def="4")

def maxEditors

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

def debug

if (!ParametersDialog.open(this)) return

// END OF PARAMETERS

def xslParameters = ""

def nEditors = 0

IdentityStyleSheet = """

<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform" xmlns:saxon="http://saxon.sf.net/">

  <xsl:output method="xml"/>

  <xsl:template match="node()|@*">

    <xsl:copy>

      <xsl:apply-templates select="@*|node()"/>

    </xsl:copy>

  </xsl:template>

</xsl:stylesheet>"""

def XSLFile = File.createTempFile("IdentityStyleSheet",".xsl")

XSLFile.deleteOnExit()

addShutdownHook {

    XSLFile.delete()

    println XSLFile+" deleted."

}

XSLFile << IdentityStyleSheet

def outputDirectory = Files.createTempDirectory("Validate")

outputDirectory = outputDirectory.toFile()

outputDirectory.deleteOnExit()

xslParameters = xslParameters.trim()

if (xslParameters.length() == 0) {

	xslParameters = null

} else {

	def split = xslParameters.split("\n")

	xslParameters = []

	for (def str : split) {

		def split2 = str.split("=", 2)

		if (split2.size() == 2) {

			xslParameters << split2

		}

	}

}

println "Processed directory: $inputDirectory."

// let's build a Saxon XSLT processor

System.setProperty("javax.xml.transform.TransformerFactory", "net.sf.saxon.TransformerFactoryImpl")

tFactory = new net.sf.saxon.TransformerFactoryImpl()

tConfiguration = tFactory.getConfiguration()

tConfiguration.registerExtensionFunction(new SaxonNodeSet()) // ?

// the following features don't seem to work, yet

tConfiguration.setConfigurationProperty(net.sf.saxon.lib.FeatureKeys.LINE_NUMBERING, true)

tConfiguration.setConfigurationProperty(net.sf.saxon.lib.FeatureKeys.RECOVERY_POLICY, net.sf.saxon.Configuration.RECOVER_WITH_WARNINGS)

// === TODO: other configuration tuning options ===

// Basically:

// - schema validation

// - dtd validation

// - etc.

//

// DTD_VALIDATION

// DTD_VALIDATION_RECOVERABLE

// LINE_NUMBERING

// RECOVERY_POLICY: Configuration.RECOVER_SILENTLY, Configuration.RECOVER_WITH_WARNINGS, or Configuration.DO_NOT_RECOVER.

// RECOVERY_POLICY_NAME (get?)

// SCHEMA_VALIDATION

// SCHEMA_VALIDATION_MODE: None, Strict, Lax, Preserve

// STANDARD_ERROR_OUTPUT_FILE

// STRIP_WHITESPACE

// SUPPRESS_XSLT_NAMESPACE_CHECK

// TRACE_LISTENER

// VALIDATION_COMMENTS

// VALIDATION_WARNINGS

// XSD_VERSION

// XSLT_SCHEMA_AWARE

def transformer = tFactory.newTransformer(new StreamSource(XSLFile))

inputDirectory.eachFileMatch(~/.+\.(xml|XML)/) { xmlInputFile ->

	name = xmlInputFile.getName()

	def editor = ""

	try {

		def xmlOutputFile = new File(outputDirectory, name)

		xmlOutputFile.deleteOnExit()

		try {

		// redirect XSLT processor error output

		tmperr = new ByteArrayOutputStream()

		tFactory.getErrorListener().setErrorOutput(new PrintStream(tmperr))

                strWriter = new StringWriter()

                out = new StreamResult(strWriter)

		transformer.transform(new StreamSource(xmlInputFile), out)

		if (debug) println "stderr = "+tmperr.toString(java.nio.charset.StandardCharsets.UTF_8)

		//System.setErr(stderr)

		w = new BufferedWriter(new FileWriter(xmlOutputFile))

		w.write(strWriter.toString())

                w.close()

		// reset transformer to free memory

		transformer = tFactory.newTransformer(new StreamSource(XSLFile))

		} catch (Exception e) {

			// org.xml.sax.SAXParseException; systemId: file:/home/sheiden/Documents/txm/xml/sample.xml; lineNumber: 3; columnNumber: 1; Le type d'élément "a" doit être suivi des spécifications d'attribut, ">" ou "/>".

			msgParse = (e.getMessageAndLocation() =~ /^(.*) lineNumber: ([0-9]+); columnNumber: ([0-9]+); (.*).$/)

			if (msgParse.size() == 1) {

				def lineNumber = msgParse[0][2]

				def colNumber = msgParse[0][3]

				def errMsg = msgParse[0][4]

				println "** $name [line $lineNumber, character $colNumber]: $errMsg."

				if (edit && (nEditors < maxEditors++)) {

					def fileStore = EFS.getLocalFileSystem().getStore(xmlInputFile.toURI())

					monitor.syncExec(new Runnable() {

						public void run() {

    							def page = PlatformUI.getWorkbench().getActiveWorkbenchWindow().getActivePage()

							editor = IDE.openEditorOnFileStore(page, fileStore)

							if (editor.getClass() == org.eclipse.wst.xml.ui.internal.tabletree.XMLMultiPageEditorPart) {

								editor = editor.fTextEditor

							}

							def dprov = editor.getDocumentProvider()

							def doc = dprov.getDocument(editor.getEditorInput())

							def nlines = doc.getNumberOfLines()

							def line = Integer.parseInt(lineNumber)-1

							if (nlines < line) line = nlines-1

							def ncols = doc.getLineLength(line)

							def col = Integer.parseInt(colNumber)-1

							if (ncols <= col) col = ncols-1

							def offset = doc.getLineOffset(line)+col

							def textv = editor.getAdapter(ITextOperationTarget.class)

							textv.getTextWidget().setSelection(offset)

							textv.revealRange(offset, 10)

						} // run

					    } // Runnable

					) // syncExec

				}

			} else {

				println e.getMessageAndLocation()

			}

		}

	} catch (Exception ee) {

		//println "Warning: XSL transformation of '$name' failed with error=$e"

		//if (debug) " with: "+e.printStackTrace()

		XSLFile.delete()

		outputDirectory.deleteDir()

		return 0

	}

}

XSLFile.delete()

outputDirectory.deleteDir()

return 1