Hi Sam, You are right that if TurboLucne just returned the hits collection then your could just use the data from the index. When I originally wrote that bit of code it was with the idea that I wanted to get back the actual model objects that match, not just the data. It is an assumption, that could be removed to make things more flexible.
I suggest 2 things: 1) That we move this conversation over the the TurboLucene mailing list as it is not really a TurboGears issue, I have CCed the TL list to start the thread over there. 2) That you open a ticket for this on the TurboLucene site and I will include the change in the next release. In the mean time you can just edit/patch the turbolucene code and adjust your results_formatter accordingly. Hope this helps. It's nice to see people using TurboLucene in ways I did not think of. It's certainly making me think! :) Krys On Saturday 2 June 2007 17:12, SamDonaldson wrote: > Ok, I will. I have a lot of questions on this and I really need to > push forward with a more direct API. I'll look into your code. > > One question: Is it absolutely necessary to return results back as a > query to get back the models? I see this as a waste. Why can't I > return the data that I indexed. Why am I have to hit the db in the IN > list query to get all the models? I'm assuming I could just return > the results whichever way I want in results_formatter? > > Thanks, > Sam > > On May 29, 4:45 am, Krys <[EMAIL PROTECTED]> wrote: > > Hi Sam, > > > > I am the author ofTurboLucene. :) All the source is in the > > __init__.py. > > > > TL is still pretty young and evolving (albeit more slowly than I would > > like.) I do plan on breaking the out of __init__.py into separate > > modules. I also plan on providing some kind of full PyLucene access > > at some point. (I just personally do not need it.) > > > > The way I dealt with the threading issues is to have the main Python > > loop create an indexer thread and a searcher factory thread (as > > PythonThreads). Then the API just sends messages through some > > Queues. It's convoluted, but I think it is better than trying to > > monkey patch CherryPy. > > > > Anyway, hope this helps. Why not join theTurboLucenemailing list? > > That's probably the best place for TL-related questions, and I'd > > certainly love to hear any thoughts/suggestions you have about TL use > > and PyLucene use in general. > > > > Thanks, > > Krys > > > > On May 28, 11:54 pm, SamDonaldson <[EMAIL PROTECTED]> wrote: > > > Guys, > > > > > > I downloaded theturbolucenesource code, but I just don't see any of > > > the files that have to do with the source? How does Turblucene > > > connect to pylucene? Where's the code that takes care of the > > > threadying issue in CherryPy? I would love to be able to connect to > > > pylucene directly without using another layer of API's provided by > > >Turbolucene. Can somebody help here? > > > > > > Thanks. > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "TurboGears" group. To post to this group, send email to turbogears@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/turbogears?hl=en -~----------~----~----~----~------~----~------~--~---