dezi search speed

[Dimitar Kenanov]

Hi Peter,

firstly thank you for the fine job with Dezi! It works very well.

I have just recently found about Dezi. I have been using Lucy before.. So i played with Dezi and managed to get it to index and search my docs. So i created the same index with Dezi and Lucy. Both huge indices. So i tested both with same query and saw a pretty large difference in speed.

lucy  - 0.7s

dezi - 6s

That is with query which returns around 13K results.

So i search in the group conversations about speed and found only 2 options you mentioned: custom aggregator and switch off auto_commit. I tried  Using the --no-ac option when starting dezi server, it did not improve the speed.

I checked the aggregator code you have provided but it seems to be just about how to index in memory, instead of parsing xml files.

As i understand the speed is such because the HTTP overhead.

So my question is. Is there a way to use the ready index in different way? I mean without using Dezi::Client and avoid the HTTP overhead.

Thank you for your time and any help or hint you can provide!

Best regards
Dimitar

[Peter Karman]

Thanks for your questions.

I am traveling and can’t reply at length but check out 

https://metacpan.org/dist/Dezi-App/view/bin/deziapp which might help.

pek

--
Dezi search platform . http://dezi.org/
---
You received this message because you are subscribed to the Google Groups "dezi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dezi-search...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dezi-search/1f6dce79-5317-43fe-905b-32b93ffa154an%40googlegroups.com.

--

Peter Karman .  he/him/his  .  https://karpet.github.io/ . 785.337.0405