Making dtSearch highlight one hit per phrase, rather than one hit per word-in-a-phrase

Posted by Chris on Stack Overflow See other posts from Stack Overflow or by Chris
Published on 2010-04-26T20:01:34Z Indexed on 2010/04/26 20:03 UTC
Read the original article Hit count: 426

Filed under:

dtsearch

|

hit-highlighting

I'm using dtSearch to highlight text search matches within a document. The code to do this, minus some details and cleanup, is roughly along these lines:

SearchJob sj = new SearchJob();
sj.Request = "\"audit trail\""; // the user query
sj.FoldersToSearch.Add(path_to_src_document);
sj.Execute();
FileConverter fileConverter = new FileConverter();
fileConverter.SetInputItem(sj.Results, 0);
fileConvert.BeforeHit = "<a name=\"HH_%%ThisHit%%\"/><b>";
fileConverter.AfterHit = "</b>";
fileConverter.Execute();
string myHighlightedDoc = fileConverter.OutputString;

If I give dtSearch a quoted phrase query like

"audit trail"

then dtSearch will do hit highlighting like this:

An <a name="HH_0"/>audit <a name="HH_1"/>trail is a fun thing to have an <a name="HH_2"/>audit <a name="HH_last"/>trail about!

Note that each word of the phrase is highlighted separately. Instead, I would like phrases to get highlighted as whole units, like this:

An <a name="HH_0"/>audit trail is a fun thing to have an <a name="HH_last"/>audit trail about!

This would A) make highlighting look better, B) improve behavior of my javascript that helps users navigate from hit to hit, and C) give more accurate counts of total # hits.

Is there good ways to make dtSearch highlight phrases this way?

Developer IT