public ChunkedText Process(String sentence)
{
ChunkedText result = null;
if (tokenizer == null)
{
InitializeTokenizer();
}
if (tagger == null)
{
InitializePOSTagger();
}
if (chunker == null)
{
InitializeChunker();
}
if (tokenizer != null && tagger != null & chunker != null)
{
result = new ChunkedText();
result.Text = sentence; //.Replace("'s", ""); //To prevent the chunker categorizing 's as a term
result.TokenArray = tokenizer.tokenize(result.Text);
result.POSTagArray = tagger.tag(result.TokenArray);
result.ChunkTagArray = chunker.chunk(result.TokenArray, result.POSTagArray);
}
return(result);
}
public virtual string[] GetPOSTags(string[] words)
{
lock (this)
{
return(tagger.tag(words));
}
}
public virtual string[] GetPOSTags(string[] words)
{
UninterruptableMonitor.Enter(this);
try
{
return(tagger.tag(words));
}
finally
{
UninterruptableMonitor.Exit(this);
}
}
public override RawCorefSample read()
{
RawCorefSample sample = samples.read();
if (sample != null)
{
IList <Parse> enhancedParses = new List <Parse>();
IList <string[]> sentences = sample.Texts;
foreach (var sentence in sentences)
{
Parse p = FullParseCorefEnhancerStream.createIncompleteParse(sentence);
p.Type = AbstractBottomUpParser.TOP_NODE;
Parse[] parseTokens = p.Children;
// construct incomplete parse here ..
string[] tags = posTagger.tag(sentence);
for (int i = 0; i < parseTokens.Length; i++)
{
p.insert(new Parse(p.Text, parseTokens[i].Span, tags[i], 1d, parseTokens[i].HeadIndex));
}
// insert tags into incomplete parse
Span[] chunks = chunker.chunkAsSpans(sentence, tags);
foreach (Span chunk in chunks)
{
if ("NP".Equals(chunk.Type))
{
p.insert(new Parse(p.Text, new Span(0, 0), chunk.Type, 1d, p.HeadIndex));
}
}
enhancedParses.Add(p);
}
sample.Parses = enhancedParses;
return(sample);
}
else
{
return(null);
}
}