public static ProcessedEntity Classify(this StanfordCoreNLP nlp, string source)
{
CoreDocument document = new CoreDocument(source);
nlp.annotate(document);
return(document.sentences()
.toArray()
.OfType <CoreSentence>()
.Select(s => new ParsedSentence(s))
.Aggregate(new ProcessedEntity(), (r, s) => ProcessedEntity.Union(r, s.ToProcessedEntity())));
}
public static ProcessedEntity Union(ProcessedEntity a, ProcessedEntity b)
{
return(new ProcessedEntity
{
Text = a.Text.Union(b.Text).ToArray(),
Reference = a.Reference.Union(b.Reference).ToArray(),
ParentReference = a.ParentReference.Union(b.ParentReference).ToArray(),
Entities = a.Entities.Union(b.Entities).ToArray(),
Triples = a.Triples.Union(b.Triples).ToArray(),
Children = a.Children.Union(b.Children).ToList(),
Dates = a.Dates.Union(b.Dates).ToArray()
});
}
private static string SaveToJsonFile(DirectoryInfo output, Series series, ProcessedEntity entity)
{
var path = Path.GetFullPath(Path.Combine(output.FullName, Path.ChangeExtension(series.File, "json")));
using (var fs = File.Open(path, FileMode.Create, FileAccess.Write, FileShare.None))
using (var sw = new StreamWriter(fs))
using (var jw = new JsonTextWriter(sw))
{
Serializer.Serialize(jw, entity);
}
return(path);
}
