Laboratoire ICAR » Plateforme TXM

// Copyright © 2019 ENS de Lyon, CNRS, University of Franche-Comté

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

// @author mdecorde

// STANDARD DECLARATIONS

package org.txm.macro.tiger.exploit

import groovy.transform.Field

import org.txm.searchengine.core.SearchEnginesManager

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

import org.txm.searchengine.ts.TIGERSearchEngine

import org.txm.utils.TableReader

def scriptName = this.class.getSimpleName()

def selection = []

for (def s : corpusViewSelections) {

	if (s instanceof CQPCorpus) selection << s

	else if (s instanceof Partition) selection.addAll(s.getParts())

}

if (selection.size() == 0) {

	println "** $scriptName: please select a Corpus or a Partition in the Corpus view: "+corpusViewSelections

	return false

} else {

	for (def c : selection) c.compute(false)

}

@Field @Option(name="query_table_file", usage="A Full TIGERSearch query", widget="FileOpen", required=true, def="queries.ods")

		def query_table_file

@Field @Option(name="count_subgraph", usage="A Full TIGERSearch query", widget="Boolean", required=true, def="true")

		def count_subgraph

@Field @Option(name="sujet_value", usage="", widget="String", required=true, def=".*subj.*")

		def sujet_value

@Field @Option(name="object_value", usage="", widget="String", required=true, def=".*obj.*")

		def object_value

@Field @Option(name="proposition_type", usage="", widget="String", required=true, def=".*")

		def proposition_type

@Field @Option(name="proposition_subordonee", usage="", widget="StringArray", metaVar="Sans	Seulement	Avec", required=true, def="Sans")

		def proposition_subordonee

@Field @Option(name="coordinations", usage="", widget="StringArray", metaVar="Sans	Seulement	Avec", required=true, def="Sans")

		def coordinations

@Field @Option(name="debug", usage="Show internal variable content", widget="StringArray", metaVar="OFF	ON	ALL	REALLY ALL", required=true, def="OFF")

		debug

if (!ParametersDialog.open(this)) return

	if (debug == "OFF") debug = 0; else if (debug == "ON") debug = 1; else if (debug == "ALL") debug = 2 else if (debug == "REALLY ALL") debug = 3

TIGERSearchEngine tse = SearchEnginesManager.getTIGERSearchEngine()

String based_query = """#pivot:[pos="VERB"] & #clause:[cat="root" & type="VFin"] & #clause >L #pivot & #clause >D #obj:[cat=("obj"|"ccomp"|"obj\\:advneg"|"obj\\:advmod")]  & #clause >D #suj:[cat=("nsubj"|"csubj")]  & #obj    >L #objhead:[]  & #suj    >L #sujhead:[]"""

println "based_query=$based_query"

def SVO_queries = new LinkedHashMap<String, String>()

SVO_queries["SVO"] = "& #sujhead .* #pivot   & #pivot   .* #objhead"

SVO_queries["SOV"] = "& #sujhead .* #objhead & #objhead .* #pivot"

SVO_queries["OSV"] = "& #objhead .* #sujhead & #sujhead .* #pivot"

SVO_queries["OVS"] = "& #objhead .* #pivot   & #pivot   .* #sujhead"

SVO_queries["VSO"] = "& #pivot   .* #sujhead & #sujhead .* #objhead"

SVO_queries["VOS"] = "& #pivot   .* #objhead & #objhead .* #sujhead"

def test_queries = []

TableReader reader = new TableReader(query_table_file)

reader.readHeaders()

def headers = reader.getHeaders()

if (!(headers.contains("mesure") && headers.contains("valeur") && headers.contains("requête"))) { // mesure, valeur, requête

	println "Error: 'mesure', 'valeur', 'requête' columns not found"

	return

}

while (reader.readRecord()) {

	test_queries << [reader.get("mesure"), reader.get("valeur"), reader.get("requête")]

}

println "mesure	valeur	SVO	SOV	OSV	OVS	VSO	VOS"

def results = new LinkedHashMap()

for (def corpus : selection) {

	def root = corpus.getRootCorpusBuild();

	File buildDirectory = new File(root.getProjectDirectory(), "tiger");

	if (!tse.hasIndexes(corpus)) {

		println "Warning: skipping $corpus: no TIGERSearch indexes found."

		continue;

	}

	def tcorpus = tse.getTSCorpus(corpus);

	def based_sentences_min_max = tse.getSentMinMax(corpus);

	for (def test : test_queries) {

		def test_query = test[2]

		print test[0]+"\t"+test[1]

		for (def k : SVO_queries.keySet()) {

			def svo_query = based_query+"\n"+SVO_queries[k]+"\n"+test_query

			def test_result = tcorpus.manager.processQuery(svo_query, based_sentences_min_max[0], based_sentences_min_max[1], 9999999);

			if (count_subgraph) {

				print "	"+test_result.submatchSize()

			} else {

				print "	"+test_result.size()

			}

		}

		println ""

		//			for (def svo_match : svo_result.getMatches()) {

		//

		//			}

	}

}

println ""

println "Done."

return results

		module.orderedFiles = new ArrayList<String>();

		for (def f : xmlFiles) {

			String id = f.getName()

			id = id.substring(0, id.indexOf(".xml"))

			module.orderedFiles.add(id)

		}

	ArrayList<String> orderedFiles = null;

		List<org.txm.searchengine.cqp.corpus.query.Match> matches = cqpCorpus.getMatches();

		if (matches.size() == 0) {

			return new int[]{0,0};

		}

		return getSentMinMax(cqpCorpus, matches.get(0).getStart(), matches.get(matches.size()-1).getEnd());

	}

	/**

	 * 

	 * @param cqpCorpus the targeted CQPCorpus 

	 * @return the first sentence and last sentence id (from 0 to N, N the number of sentences). WARNING: this is not the list of sentences in the targeted CQPCorpus. unless the CQPcorpus is contigues

	 * 

	 * @throws UnexpectedAnswerException

	 * @throws IOException

	 * @throws CqiServerError

	 * @throws CqiClientException

	 */

	public static int[] getSentMinMax(CQPCorpus cqpCorpus, int start, int end) throws UnexpectedAnswerException, IOException, CqiServerError, CqiClientException {

		int[] cpos = new int[] {start, end};