File size: 4,976 Bytes
18573e4 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 | package bg.bas.dcl.LLMs.IfGPTDataset;
import java.io.File;
import java.io.FileOutputStream;
import java.io.OutputStreamWriter;
import java.io.Writer;
import java.util.Scanner;
import org.json.simple.JSONArray;
import org.json.simple.JSONObject;
import bg.bas.dcl.general.FileHandler;
/**
* Processes the MARCELL Bulgarian annotated corpus.
*
* Licence: CC0-1.0 (fixed for all MARCELL documents).
* Domain: "Държавно управление" (State governance).
* Style: "Административен".
*/
public class MarcellProcessor extends BaseSourceProcessor {
private static final String LICENCE = "CC0-1.0";
private static final String LICENCE_LINK =
"https://elrc-share.eu/static/metashare/licences/CC0-1.0.pdf";
private static final String DOMAIN = "Държавно управление";
private static final String STYLE = "Административен";
private static final String PREFIX = "bg_MARCELL_";
private static final String EXT = ".conllup";
@Override
public void process(String indir, String outdir) {
try {
FileHandler fh = new FileHandler();
JSONObject json = new JSONObject();
JSONArray descrArray = new JSONArray();
for (File f : fh.getFileListing(new File(indir))) {
if (!f.isFile()) continue;
System.out.println("Processing: " + f.getAbsolutePath());
String tfname = PREFIX + f.getName().replace(EXT, "");
JSONObject fdescr = newBaseDescriptor(tfname);
fdescr.put("Licence", LICENCE);
fdescr.put("LicenceLink", LICENCE_LINK);
fdescr.put("Domain", DOMAIN);
fdescr.put("Style", STYLE);
Writer out = new OutputStreamWriter(
new FileOutputStream(outdir + tfname + ".txt"), "UTF-8");
Scanner s = new Scanner(f, "UTF-8");
int nw = 0, ns = 0, np = 0, nt = 0;
while (s.hasNextLine()) {
String line = s.nextLine();
// --- Metadata extraction ---
if (line.startsWith("# date =")) {
fdescr.put("PublicationDate", line.replace("# date =", "").trim());
} else if (line.startsWith("# title =")) {
fdescr.put("DocumentTitle", line.replace("# title =", "").trim());
} else if (line.startsWith("# issuer =")) {
fdescr.put("Author", line.replace("# issuer =", "").trim());
} else if (line.startsWith("# type =")) {
fdescr.put("Type", line.replace("# type =", "").trim());
} else if (line.startsWith("# url =")) {
fdescr.put("Url", line.replace("# url =", "").trim());
}
// --- Structure counting ---
else if (line.startsWith("# sent_id =")) {
ns++;
} else if (line.startsWith("# newpar id =")) {
np++;
out.write("\n");
}
// --- Text output ---
else if (line.startsWith("# text =")) {
out.write(line.replace("# text =", "").trim() + "\n");
out.flush();
} else {
// CoNLL-UP token line: count words and tokens
String[] cols = line.split("\t");
if (cols.length > 5) {
nt++;
if (!cols[3].equals("PUNCT")) nw++;
}
}
}
s.close();
out.flush();
out.close();
fdescr.put("NumberWords", nw);
fdescr.put("NumberSentences", ns);
fdescr.put("NumberParagraphs", np);
fdescr.put("NumberTokens", nt);
descrArray.add(fdescr);
}
json.put("metadata", descrArray);
writeMetadata(json, outdir, "metadata.json");
} catch (Exception e) {
e.printStackTrace();
}
}
// -----------------------------------------------------------------------
@SuppressWarnings("unchecked")
private void writeMetadata(JSONObject json, String outdir, String filename)
throws Exception {
String outMetaPath = outdir + filename;
Writer outMeta = new OutputStreamWriter(
new FileOutputStream(outMetaPath), "UTF-8");
json.writeJSONString(outMeta);
outMeta.flush();
outMeta.close();
convertJsonToCSV(json, outMetaPath + "_CSV.csv");
System.out.println("Metadata written to: " + outMetaPath);
}
}
|