Google like search


(alheim) #1

Hi,

I'm trying to implement the same kind of "natural" search query than google.

I'm using query strings on fields analyzed with a stemmer filter.

What is a good configuration of the query string to get the same results
than google would give ?

  • default operator
  • phrase slop
  • auto generate phrase

Thanks


(alheim) #2

no answer..I understand that this could look like a newbie question, but
it's really important to undersand how to offer a natural search.

On Monday, March 19, 2012 12:38:31 PM UTC+1, Alexandre Heimburger wrote:

Hi,

I'm trying to implement the same kind of "natural" search query than
google.

I'm using query strings on fields analyzed with a stemmer filter.

What is a good configuration of the query string to get the same results
than google would give ?

  • default operator
  • phrase slop
  • auto generate phrase

Thanks


(David Pilato) #3

Hi Alexandre,

At the present time, I use a "google" like search on my project.
It's not really natural language but helps a lot to deal with french
queries.

I created a GIST here to show a test case for that.
I'm not sure that it will fit to your own needs, but I hope it could help.

https://gist.github.com/2146038 https://gist.github.com/2146038

Take care
David.

Le 28 mars 2012 à 12:49, Alexandre Heimburger ahb@bluekiwi-software.com a
écrit :

could look like a newbie question, but it's really important to undersand
how to offer a natural search.

On Monday, March 19, 2012 12:38:31 PM UTC+1, Alexandre Heimburger wrote:

I'm trying to implement the same kind of "natural" search query than
google.

I'm using query strings on fields analyzed with a stemmer filter.

What is a good configuration of the query string to get the same results
than google would give ?

  • default operator
  • phrase slop
  • auto generate phrase

Thanks

--
David Pilato
http://dev.david.pilato.fr/
Twitter : @dadoonet


(alheim) #4

Thanks david. I've got the same configuration. I suppose it is a good one :slight_smile:

Btw, I know that you have a talk slot during Devoxx event. Maybe we could
meet.

Cheers

On Wed, Mar 28, 2012 at 2:56 PM, david@pilato.fr david@pilato.fr wrote:

**

Hi Alexandre,

At the present time, I use a "google" like search on my project.

It's not really natural language but helps a lot to deal with french
queries.

I created a GIST here to show a test case for that.

I'm not sure that it will fit to your own needs, but I hope it could help.

https://gist.github.com/2146038

Take care

David.

Le 28 mars 2012 à 12:49, Alexandre Heimburger ahb@bluekiwi-software.com
a écrit :

no answer..I understand that this could look like a newbie question, but
it's really important to undersand how to offer a natural search.

On Monday, March 19, 2012 12:38:31 PM UTC+1, Alexandre Heimburger wrote:

Hi,

I'm trying to implement the same kind of "natural" search query than
google.

I'm using query strings on fields analyzed with a stemmer filter.

What is a good configuration of the query string to get the same results
than google would give ?

  • default operator
  • phrase slop
  • auto generate phrase

Thanks

--
David Pilato
http://dev.david.pilato.fr/
Twitter : @dadoonet

--
Alexandre Heimburger
VP Engineering
blueKiwi Software
tel : +33687880997
email : ahb@bluekiwi-software.com
adress : 93 rue Vieille du Temple, 75003 Paris

blueKiwi is the innovation leader in Enterprise Social Software. Our
solutions enable enterprises worldwide to engage and interact with their
internal and external social networks in multiple business domains.
blueKiwi has been consistently recognized by Gartner Inc. as a visionary
provider since 2007.


(David Pilato) #5

Sure ! Glad to meet you there !

BTW, there's something missing in elision filter (the d letter used in
"police d'assurance" for example). I opened a bug here
https://issues.apache.org/jira/browse/LUCENE-3931
https://issues.apache.org/jira/browse/LUCENE-3931

But, you can fix it on ES side with defining the elision filter :

"elision" : {
"type" : "elision",
"articles" : ["l", "m", "t", "qu", "n", "s", "j", "d"]
}

I updated the gist.

Take care
David.

Le 28 mars 2012 à 15:15, Alexandre Heimburger ahb@bluekiwi-software.com a
écrit :

configuration. I suppose it is a good one :slight_smile:

Btw, I know that you have a talk slot during Devoxx event. Maybe we could
meet.

Cheers

On Wed, Mar 28, 2012 at 2:56 PM, david@pilato.fr mailto:david@pilato.fr
< david@pilato.fr mailto:david@pilato.fr > wrote:

Hi Alexandre,

At the present time, I use a "google" like search on my project.
It's not really natural language but helps a lot to deal with french
queries.

I created a GIST here to show a test case for that.
I'm not sure that it will fit to your own needs, but I hope it could help.

https://gist.github.com/2146038 https://gist.github.com/2146038

Take care
David.

Le 28 mars 2012 à 12:49, Alexandre Heimburger < ahb@bluekiwi-software.com
mailto:ahb@bluekiwi-software.com > a écrit :

could look like a newbie question, but it's really important to undersand
how to offer a natural search.

On Monday, March 19, 2012 12:38:31 PM UTC+1, Alexandre Heimburger wrote:

I'm trying to implement the same kind of "natural" search query than
google.

I'm using query strings on fields analyzed with a stemmer filter.

What is a good configuration of the query string to get the same results
than google would give ?

  • default operator
  • phrase slop
  • auto generate phrase

Thanks

--
David Pilato
http://dev.david.pilato.fr/ http://dev.david.pilato.fr/
Twitter : @dadoonet

--
Alexandre Heimburger
VP Engineering
blueKiwi Software
tel : +33687880997
email : ahb@bluekiwi-software.com mailto:ahb@bluekiwi-software.com
adress : 93 rue Vieille du Temple, 75003 Paris

blueKiwi is the innovation leader in Enterprise Social Software. Our
solutions enable enterprises worldwide to engage and interact with their
internal and external social networks in multiple business domains.
blueKiwi has been consistently recognized by Gartner Inc. as a visionary
provider since 2007.

--
David Pilato
http://dev.david.pilato.fr/
Twitter : @dadoonet


(system) #6