Laboratoire ICAR » Plateforme TXM

package org.txm.macro.frolex

import java.io.File

import java.io.IOException

import java.io.PrintWriter

import java.nio.charset.Charset

import java.util.ArrayList

import java.util.HashMap

import javax.xml.stream.XMLStreamException

import org.txm.utils.CsvReader

import org.txm.utils.io.IOUtils

import org.kohsuke.args4j.*

import groovy.transform.Field

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

// PARAMETERS

@Field @Option(name="workingDirectory", usage="workingDirectory containing all files needed", widget="Folder", required=true, def="dir")

File workingDirectory = null

if (!ParametersDialog.open(this)) {

        System.out.println("Aborting CLBFMSITELEX creation.")

        return null

}

File dmfsiteTSVFile = new File(workingDirectory, "dmf_site.tsv")

if (!dmfsiteTSVFile.exists()) {

        println "Aborting: file is missing: $dmfsiteTSVFile"

        return

}

File tsvFile = new File(workingDirectory, "cldmf-site.tsv")

//START

HashMap<String, Entry2> lemmas = new HashMap<String, Entry2>()

CsvReader reader = new CsvReader(dmfsiteTSVFile.getAbsolutePath(), "\t".charAt(0), Charset.forName("UTF-8"))

reader.readHeaders()

while (reader.readRecord()) {

        String lemma = reader.get("dmf_lemma")

        String category = reader.get("category")

        String corresp = convertCorresp(reader.get("source"))

        String corresp_lemma = reader.get("source_lemma")

        String key = lemma+"_"+category

        if (!lemmas.containsKey(key)) {

                lemmas.put(key, new Entry2(lemma, category))

                lemmas.get(key).lemma_source = "DMF"

        }

        Entry2 current = lemmas.get(key)

        current.lemma_source = "DMF"

        if (current.corresps.keySet().contains(corresp)) {

                current.corresps.put(corresp, corresp_lemma)

        } else {

                System.out.println("Error unknown lemma="+lemma+" with corresp="+corresp+" and corresp_lemma="+corresp_lemma)

        }

}

reader.close()

PrintWriter writer = IOUtils.getWriter(tsvFile)

writer.print("lemma\tcategory")

Entry2 tmp = new Entry2()

for (String corresp : tmp.corresps.keySet()) {

        writer.print("\t"+corresp)

}

writer.print("\tlemma_source")

writer.println("")

ArrayList<String> keys = new ArrayList<String>(lemmas.keySet())

keys.sort()

for (String key : keys) {

        writer.print(lemmas.get(key))

}

writer.close()

return true

// END

String convertCorresp(String corresp) {

        if (corresp.equals("AND")) {

                return "AND"

        } else if (corresp.matches("\\*?DEAF.*")) {

                return "DEAF"

        } else if (corresp.equals("DÉCT")) {

                return "DECT"

        } else if (corresp.startsWith("FEW")) {

                return "FEW"

        } else if (corresp.equals("GD")) {

                return "GDF"

        } else if (corresp.startsWith("GDC")) {

                return "GDC"

        } else if (corresp.equals("gs")) {

                return "TLF"

        } else if (corresp.equals("HUG")) {

                return "HUG"

        } else if (corresp.startsWith("T-L")) {

                return "TL"

        } else if (corresp.startsWith("TLF")) {

                return "TLF"

        } else {

                return "UNDEF"

        }

}