Am 16.03.2013 15:18, schrieb Ulrich Pegelow: > Am 16.03.2013 15:14, schrieb Sebastian Kraft: >> >> One way to improve the matching with the fuzzy search would be to >> evaluate the score returned with the search result. If the score is high >> you know the search result matches quite good. Have you tried this? >> >> For me this gives a very good indication about the search result >> quality. At least I could not find an example where the fuzzy search >> returned wrong entries with a high score. >> > > Is the score value documented anywhere? What would be a good threshold > to accept or refuse a match? >
Unfortunately not :( I just found out that this exists a few weeks ago. You can compile the tfun.cpp example code and try a search with several lens names supplied by commandline arguments. A perfect match is around 80 points. The Tamron 17-50 lens with the "(model 16)" part yields 67 points. So maybe a threshold at 50 points should work? For better details we would need to investigate the search algorithm more deeply... Sebastian ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_mar _______________________________________________ darktable-devel mailing list darktable-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/darktable-devel
