/**
* convert a list of APF files to XML files containing in-line markup for sentences, timex,
* mentions, entity types, and names. Takes the following arguments:
*
* <ul>
* <li>year: the year (2002, 2003, 2004, 2005) the APF file was created, which determines its
* format
* <li>inDir: the directory containing the text and APF files
* <li>outDir: the directory containing the NE files
* <li>filelist: a file containing a list of document names
* <li>apfExtension: the suffix added to the doc name to create the APF file name
* <li>outExtension: the suffix added to the doc name to create the output file name
* <li>gazetteer: (for 2004 files only) a Jet gazetteer, used to resolve the name/noun
* distinction for GPE words categorized as PRE
* <li>PREdict: (for 2004 files only) a name vs. noun list, used to resolve the name/noun
* distinction for non-GPE words categorized as PRE
* <li>flag ... : one or more of 'sentences', 'timex', 'mentions', 'extents', 'types' or
* 'names', indicating the information to be included in the output file
* </ul>
*/
public static void main(String[] args) throws IOException {
init(args);
processFileList(fileList);
if (year.equals("2004")) {
System.out.println("\nUnclassified items: " + unknownPre.size());
for (String word : unknownPre) System.out.println(word);
}
}
/** write 'fileText' out as file 'XMLfileName' with ENAMEX tags for the names in the document */
static void addENAMEXtags(Document doc, AceDocument aceDoc) {
ArrayList<AceEntity> entities = aceDoc.entities;
for (int i = 0; i < entities.size(); i++) {
AceEntity entity = entities.get(i);
ArrayList<AceEntityName> names = entity.names;
for (int j = 0; j < names.size(); j++) {
AceEntityName name = names.get(j);
Span aceSpan = name.extent;
Span jetSpan = new Span(aceSpan.start(), aceSpan.end() + 1);
doc.annotate("ENAMEX", jetSpan, new FeatureSet("TYPE", entity.type));
}
// for 2004 we have to examine PRE mentions and decide which are names
if (year.equals("2004")) {
ArrayList<AceEntityMention> mentions = entity.mentions;
for (int j = 0; j < mentions.size(); j++) {
AceEntityMention mention = mentions.get(j);
String htext = Resolve.normalizeName(mention.headText);
String[] mentionName = Gazetteer.splitAtWS(htext);
String preClass = preDict.get(htext.toLowerCase());
if (mention.type.equals("PRE")) {
if (gazetteer.isNationality(mentionName)
|| gazetteer.isLocation(mentionName)
|| "N".equals(preClass)) {
Span aceSpan = mention.head;
Span jetSpan = new Span(aceSpan.start(), aceSpan.end() + 1);
doc.annotate("ENAMEX", jetSpan, new FeatureSet("TYPE", entity.type));
} else if (preClass != null) {
// do nothing
} else {
System.out.println(
"Unclassified PRE: " + mention.text + " {" + mention.headText + ")");
unknownPre.add(htext.toLowerCase());
}
}
}
}
}
}
