Laboratoire ICAR » Plateforme TXM

package org.txm.scripts.export.conll2009

import java.io.File;

import java.io.FileOutputStream;

import java.io.OutputStreamWriter;

import java.io.Writer;

import java.util.Arrays;

import java.util.List;

import org.txm.searchengine.cqp.CqpDataProxy;

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

import org.txm.searchengine.cqp.corpus.query.CQLQuery

class ToCoNLL2009 {

        boolean debug = true

        /**

         * To CoNLL2009.

         *

         * @param outfile the outfile, where is result is saved

         * @param corpus the corpus or sub-corpus to export

         * @param sentenceUnit the StructuralUnitProperty that identify a sentence

         * @param word the word property

         * @param lemme the lemme property

         * @param pos the pos property

         * @param encoding the encoding of the outfile

         * @return true, if successful

         */

        public boolean process(File outfile, CQPCorpus corpus, StructuralUnitProperty sentenceUnit, Property word, Property lemme, Property pos, String encoding)

        {

                def matches = corpus.query(new CQLQuery("<s> []+ </s>"), "TMPEXPORTCONNL", false).getMatches()

                if (debug) println matches.size()

                int npositions = 0;

                for (def match : matches) npositions += match.size() +1

                if (debug) println "npositions= $npositions"

                int[] positions = new int[npositions+1]

                int i = 0;

                for (def match : matches) {

                        for (int p : match.getRange())

                                positions[i++] = p

                }

                List<String> words = CorpusManager.getCorpusManager().getCqiClient().cpos2Str(word.getQualifiedName(), positions);

                List<String> lemmes = null;

                if (lemme != null) {

                        lemmes = CorpusManager.getCorpusManager().getCqiClient().cpos2Str(lemme.getQualifiedName(), positions);

                }

                List<String> poss = null;

                if (pos != null) {

                        poss =  CorpusManager.getCorpusManager().getCqiClient().cpos2Str(pos.getQualifiedName(), positions);

                }

                Writer writer = new OutputStreamWriter(new FileOutputStream(outfile), "UTF-8"); //$NON-NLS-1$;

                String lemmestr;

                String posstr;

                int id = 0;

                int iposition = 0;

                for (def match : matches) {

                        id = 0 // first position in sentence

                        for (i = match.start ; i < match.end ; i++) { // loop n times, n = mathc length

                                if (lemme != null)

                                        lemmestr = lemmes.get(iposition);

                                else

                                        lemmestr = "_";

                                if (pos != null)

                                        posstr = poss.get(iposition);

                                else

                                        posstr = "_";

                                //print(""+id+"\t"+words.get(iposition)+"\t"+lemmestr+"\t_\t"+posstr+"\t_\t_\t_\t_\t_\t_\t_\t_\t_\t_\n");

                                writer.write(""+id+"\t"+words.get(iposition)+"\t"+lemmestr+"\t_\t"+posstr+"\t_\t_\t_\t_\t_\t_\t_\t_\t_\t_\n");

                                id++; // next word

                                iposition++ // nextposition

                        }

                }

                writer.flush();

                writer.close();

                return true;

        }

        public static void main(String[] args) {

                CQPCorpus corpus = null

                File outfile = new File("/home/mdecorde/Bureau/conn_export.tsv")

                String encoding = "UTF-8"

                String sentenceProperty = "s_id"

                String posProperty = "frpos"

                String lemmaProperty = "frlemma"

                def split = sentenceProperty.split("_", 2)

                StructuralUnitProperty s = null;//corpus.getStructuralUnit(split[0]).getProperty(split[1])

                Property word = null;//corpus.getProperty("word")

                Property lemma = null;//corpus.getProperty(lemmaProperty)

                Property pos = null;//corpus.getProperty(posProperty)

                //                if (s == null) { println "Error sentence property: $sentenceProperty"; return}

                //                if (word == null) { println "Error no word property"; return}

                //                if (pos == null) { println "Error pos property: $posProperty"; return}

                //                if (lemma == null) { println "Error lemma property: $lemmaProperty"; return}

                ToCoNLL2009 processor = new ToCoNLL2009()

                processor.process(outfile, corpus, s, word, lemma, pos, encoding)

        }

}