/// <summary>
/// Annotates the given sequence of <see cref="Document"/> objects by adding a <b>_highlight</b> field;
/// the <b>_highlight</b> field will contain the best matching text fragment from the <see cref="Document"/>
/// object's full-text field.
/// </summary>
/// <param name="hits">The sequence of <see cref="Document"/> objects.</param>
/// <param name="criteria">The search criteria that produced the hits.</param>
/// <returns>
/// The original sequence of Document objects, with a <b>_highlight</b> field added to each Document.
/// </returns>
public static IEnumerable<Document> GenerateHighlights(this IEnumerable<Document> hits, SearchCriteria criteria)
{
if (hits == null)
throw new ArgumentNullException(nameof(hits));
if (criteria == null)
throw new ArgumentNullException(nameof(criteria));
if (String.IsNullOrWhiteSpace(criteria.Query))
throw new ArgumentException("SearchCriteria.Query cannot be empty");
var documents = hits.ToList();
try
{
var indexDirectory = new RAMDirectory();
var analyzer = new FullTextAnalyzer();
var config = new IndexWriterConfig(analyzer).SetRAMBufferSizeMB(_ramBufferSizeMB);
var writer = new IndexWriter(indexDirectory, config);
BuidIndex(documents, writer);
GenerateHighlights(documents, writer, criteria);
writer.DeleteAll();
writer.Commit();
writer.Close();
indexDirectory.Close();
}
catch (Exception ex)
{
_log.Error(ex);
}
return documents;
}
/// <summary>
/// Breaks up the input text into individual tokens.
/// </summary>
/// <param name="text">The input text.</param>
/// <param name="enableStemming">if set to <c>true</c>, the FullTextIndex will stem
/// the tokens that make up the texts, using the Porter stemming algorithm.</param>
/// <param name="ignoreCase">if set to <c>true</c>, character casing is ignored.</param>
/// <param name="separatorChars">A string whose component characters will be used to split the texts into tokens.</param>
/// <returns></returns>
public static IEnumerable<string> Tokenize(string text, bool enableStemming = true, bool ignoreCase = true, string separatorChars = DEFAULT_SEPARATOR_CHARS)
{
if (String.IsNullOrWhiteSpace(text))
throw new ArgumentException("text cannot be null or blank");
if (String.IsNullOrWhiteSpace(separatorChars))
separatorChars = DEFAULT_SEPARATOR_CHARS;
using (var analyzer = new FullTextAnalyzer(enableStemming, ignoreCase, separatorChars))
{
using (var stream = analyzer.TokenStream("text", text))
{
var attrib = stream.AddAttribute(typeof(CharTermAttribute)) as CharTermAttribute;
stream.Reset();
while (stream.IncrementToken())
{
yield return attrib.ToString();
}
stream.End();
}
}
}
