Help regarding BM25Similarity

2018-01-04 Thread Parit Bansal
Hi, I am trying to tweak BM25Similarity for my use case wherein, I want to avoid the effects of field-length normalization for certain fields (return a constant value irrespective of how long was the document). Currently, both computeWeight and computeNorm methods are defined final in BM25Sim

Re: Help regarding BM25Similarity

2018-01-04 Thread Uwe Schindler
How about just indexing the field without norms? Uwe Am January 4, 2018 3:58:27 PM UTC schrieb Parit Bansal : >Hi, > >I am trying to tweak BM25Similarity for my use case wherein, I want to >avoid the effects of field-length normalization for certain fields >(return a constant value irrespective

Re: Help regarding BM25Similarity

2018-01-04 Thread Robert Muir
You don't need to do any subclassing for this: just pass parameter b=0 to the constructor. On Thu, Jan 4, 2018 at 10:58 AM, Parit Bansal wrote: > Hi, > > I am trying to tweak BM25Similarity for my use case wherein, I want to avoid > the effects of field-length normalization for certain fields (re

Re: Help regarding BM25Similarity

2018-01-05 Thread Parit Bansal
Hi Uwe, You are right. Thankx! :) - Best Parit Bansal On 01/04/2018 05:02 PM, Uwe Schindler wrote: How about just indexing the field without norms? Uwe Am January 4, 2018 3:58:27 PM UTC schrieb Parit Bansal : Hi, I am trying to tweak BM25Similarity for my use case wherein, I want to avoid

Re: Help regarding BM25Similarity

2018-01-05 Thread Parit Bansal
Hi Robert, passing b = 0 will influence the similarity across all the fields (no?) . I wanted it to be field specific. I think Uwe's suggestion of not indexing norms for specific fields should work better. Thankx again. - Best Parit Bansal On 01/04/2018 08:34 PM, Robert Muir wrote: You do

Re: Help regarding BM25Similarity

2018-01-05 Thread Parit Bansal
Hi Robert, passing b = 0 will influence the similarity across all the fields (no?) . I wanted it to be field specific. I think Uwe's suggestion of not indexing norms for specific fields should work better. - Best Parit Bansal On 01/04/2018 08:34 PM, Robert Muir wrote: You don't need to do

Re: Help regarding BM25Similarity

2018-01-05 Thread Adrien Grand
You can use PerFieldSimilarityWrapper to have different BM25 settings per field. Le ven. 5 janv. 2018 à 10:37, Parit Bansal a écrit : > Hi Robert, > > passing b = 0 will influence the similarity across all the fields (no?) > . I wanted it to be field specific. I think Uwe's suggestion of not > i

Re: Help regarding BM25Similarity

2018-01-05 Thread Parit Bansal
Thankx Adrien. I'll try this approach too. - Best Parit Bansal On 01/05/2018 10:43 AM, Adrien Grand wrote: You can use PerFieldSimilarityWrapper to have different BM25 settings per field. Le ven. 5 janv. 2018 à 10:37, Parit Bansal a écrit : Hi Robert, passing b = 0 will influence the simil