public static IList <CoreLabel> ToCoreLabelListWithCharacterOffsets(string[] words, string[] tags)
{
System.Diagnostics.Debug.Assert(tags.Length == words.Length);
IList <CoreLabel> tokens = new List <CoreLabel>(words.Length);
int offset = 0;
for (int i = 0; i < sz; i++)
{
CoreLabel cl = new CoreLabel();
cl.SetWord(words[i]);
cl.SetTag(tags[i]);
cl.Set(typeof(CoreAnnotations.CharacterOffsetBeginAnnotation), offset);
offset += words[i].Length;
cl.Set(typeof(CoreAnnotations.CharacterOffsetEndAnnotation), offset);
offset++;
// assume one space between words :-)
tokens.Add(cl);
}
return(tokens);
}
public virtual ILabel NewLabel(ILabel oldLabel)
{
if (oldLabel is CoreLabel)
{
return(new CoreLabel((CoreLabel)oldLabel));
}
else
{
//Map the old interfaces to the correct key/value pairs
//Don't need to worry about HasIndex, which doesn't appear in any legacy code
CoreLabel label = new CoreLabel();
if (oldLabel is IHasWord)
{
label.SetWord(((IHasWord)oldLabel).Word());
}
if (oldLabel is IHasTag)
{
label.SetTag(((IHasTag)oldLabel).Tag());
}
if (oldLabel is IHasOffset)
{
label.SetBeginPosition(((IHasOffset)oldLabel).BeginPosition());
label.SetEndPosition(((IHasOffset)oldLabel).EndPosition());
}
if (oldLabel is IHasCategory)
{
label.SetCategory(((IHasCategory)oldLabel).Category());
}
if (oldLabel is IHasIndex)
{
label.SetIndex(((IHasIndex)oldLabel).Index());
}
label.SetValue(oldLabel.Value());
return(label);
}
}
/// <summary>
/// Default constructor; uses
/// <see cref="CoreLabel"/>
/// default constructor
/// </summary>
public IndexedWord()
{
// = 0;
label = new CoreLabel();
}
/// <summary>Construct an IndexedWord from a CoreLabel just as for a CoreMap.</summary>
/// <remarks>
/// Construct an IndexedWord from a CoreLabel just as for a CoreMap.
/// <i>Implementation note:</i> this is a the same as the constructor
/// that takes a CoreMap, but is needed to ensure unique most specific
/// type inference for selecting a constructor at compile-time.
/// </remarks>
/// <param name="w">A Label to initialize this IndexedWord from</param>
public IndexedWord(CoreLabel w)
{
label = w;
}
public static string WordToString <T>(T o, bool justValue, string separator)
{
if (justValue && o is ILabel)
{
if (o is CoreLabel)
{
CoreLabel l = (CoreLabel)o;
string w = l.Value();
if (w == null)
{
w = l.Word();
}
return(w);
}
else
{
return(((ILabel)o).Value());
}
}
else
{
if (o is CoreLabel)
{
CoreLabel l = ((CoreLabel)o);
string w = l.Value();
if (w == null)
{
w = l.Word();
}
if (l.Tag() != null)
{
if (separator == null)
{
return(w + CoreLabel.TagSeparator + l.Tag());
}
else
{
return(w + separator + l.Tag());
}
}
return(w);
}
else
{
// an interface that covered these next four cases would be
// nice, but we're moving away from these data types anyway
if (separator != null && o is TaggedWord)
{
return(((TaggedWord)o).ToString(separator));
}
else
{
if (separator != null && o is LabeledWord)
{
return(((LabeledWord)o).ToString(separator));
}
else
{
if (separator != null && o is WordLemmaTag)
{
return(((WordLemmaTag)o).ToString(separator));
}
else
{
if (separator != null && o is WordTag)
{
return(((WordTag)o).ToString(separator));
}
else
{
return(o.ToString());
}
}
}
}
}
}
}
/// <summary><inheritDoc/></summary>
public virtual ILabelFactory LabelFactory()
{
return(CoreLabel.Factory());
}