You are viewing a plain text version of this content. The canonical link for it is here.
Posted to java-user@lucene.apache.org by saikrishna venkata pendyala <pv...@gmail.com> on 2007/10/08 15:36:47 UTC

Re : porting lucene on to mobile phone

Hai,

      Has any one tried to port lucene on to a mobile phone. I need a
text-based search engine for mobile phone as part of my project.

      Or is there any other J2ME based API which can be used to search the
text in mobile phones.

thanks,
Saikrishna.

Re: Re : porting lucene on to mobile phone

Posted by Chris Hostetter <ho...@fucit.org>.
: > You want to create 1,7 million documents containing a short and
: > indexed field: title, and a stored field: file name?

: Yes this is what I need.

if i may make a modest proposal: i would suggest that before you dig ito 
the details of what it takes to get LUcene working on a mobile platform, 
you spend 30 minutes writing a simple command line (JRE) app that uses 
lucene to index your source data the way you expect ot index it (ie: index 
the title, store the filename) and see how big that index is.  depending 
on how long these titles/filenames are and how many unique terms are in 
the titles, you may find that your index won't even fit in the mount of 
storage available on your target device -- so the issues of getting Lucene 
running on that device may be moot.

if thta is the case; perhaps a webservice (or SMS) based approach would 
work better .. Solr might be able to simplify the "server' side, all you 
need to do is write the client and feed Solr a CSV files of your data.




-Hoss


---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org


Re: Re : porting lucene on to mobile phone

Posted by Karl Wettin <ka...@gmail.com>.
8 okt 2007 kl. 19.17 skrev saikrishna venkata pendyala:

>> You only require boolean term queries? And no scoring, sorting or
>> anything like that?
>>
>> In that case, it should not be that hard to rewrite Lucene for your
>> needs.
>
>
> Can I rewrite lucene to match the given query with some title ?

Sorry, I still don't understand.

If you only want matching and no scoring (i.e. you will not get to  
know what document was the better or worse match), then it is no that  
hard to get rid of all floating points.

See LUCENE-584 for seperation of Matcher and Scorer.


-- 
karl

---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org


Re: Re : porting lucene on to mobile phone

Posted by saikrishna venkata pendyala <pv...@gmail.com>.
Sorry I couldn't frame my question correctly due to typos.

this is what you said earlier

"""
You only require boolean term queries? And no scoring, sorting or
anything like that?

In that case, it should not be that hard to rewrite Lucene for your
needs.

"""

Can I rewrite lucene to match the given query with some title ?


On 10/8/07, Karl Wettin <ka...@gmail.com> wrote:
>
>
> 8 okt 2007 kl. 18.52 skrev saikrishna venkata pendyala:
>
> > Can I use ucene with some changs, for the purpose of matching and
> > with OUT
> > SCORING the query given.
>
> Sorry, but I don't understand your question?
>
> In order to run Lucene on J2ME you need CLDC1.1 or later,
> alternatively modify Lucene to use some sort of floating point library.
>
>
> --
> karl
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-user-help@lucene.apache.org
>
>

Re: Re : porting lucene on to mobile phone

Posted by Karl Wettin <ka...@gmail.com>.
8 okt 2007 kl. 18.52 skrev saikrishna venkata pendyala:

> Can I use ucene with some changs, for the purpose of matching and  
> with OUT
> SCORING the query given.

Sorry, but I don't understand your question?

In order to run Lucene on J2ME you need CLDC1.1 or later,  
alternatively modify Lucene to use some sort of floating point library.


-- 
karl

---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org


Re: Re : porting lucene on to mobile phone

Posted by saikrishna venkata pendyala <pv...@gmail.com>.
On 10/8/07, Karl Wettin <ka...@gmail.com> wrote:
>
>
> 8 okt 2007 kl. 18.13 skrev saikrishna venkata pendyala:
>
> > I need the name of the file in which the best matched title exists.
> > Don't I
> > need scoring here.
>
> I don't know how you would decide what the best matched document is
> without scoring the results?



Yes we need scoring to decide what is best matched.

Can I use ucene with some changs, for the purpose of matching and with OUT
SCORING the query given.

thnks,
Saikrishna.

--
> karl
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-user-help@lucene.apache.org
>
>

Re: Re : porting lucene on to mobile phone

Posted by Karl Wettin <ka...@gmail.com>.
8 okt 2007 kl. 18.13 skrev saikrishna venkata pendyala:

> I need the name of the file in which the best matched title exists.  
> Don't I
> need scoring here.

I don't know how you would decide what the best matched document is  
without scoring the results?


-- 
karl

---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org


Re: Re : porting lucene on to mobile phone

Posted by saikrishna venkata pendyala <pv...@gmail.com>.
On 10/8/07, Karl Wettin <ka...@gmail.com> wrote:
>
>
> 8 okt 2007 kl. 17.40 skrev saikrishna venkata pendyala:
>
> >  I have abt 17,00,000 documents in 1000 files. I need to index the
> > titles of
> > those documents with out storing the titles and storing the file name.
>
> Not sure I understand what you say here.
>
> You want to create 1,7 million documents containing a short and
> indexed field: title, and a stored field: file name?


Yes this is what I need.


> When ever I issue a query I need to get the name of the file in
> > which that
> > document exists.
>
> You only require boolean term queries? And no scoring, sorting or
> anything like that?



I need the name of the file in which the best matched title exists. Don't I
need scoring here.

Suppose my query is "q1 q2 q3" and their are titles like "q1
q2","q3","q1",etc.  Among this the best mateched title is "q1 q2", and I
need the name of the file in which the document with title "q1 q2" occur.



In that case, it should not be that hard to rewrite Lucene for your
> needs.
>
> See LUCENE-584 for seperation of Matcher and Scorer.
>
>
> --
> karl
>
>
> >
> > thanks,
> > Saikrishna.
> >
> > On 10/8/07, Karl Wettin <ka...@gmail.com> wrote:
> >>
> >>
> >> 8 okt 2007 kl. 15.36 skrev saikrishna venkata pendyala:
> >>
> >>>       Has any one tried to port lucene on to a mobile phone. I
> >>> need a
> >>> text-based search engine for mobile phone as part of my project.
> >>
> >>
> >> There was some discussion about this here on the forums some year(s?)
> >> ago. The major problem back then was the lack of floaing points.
> >> Also, as Lucene moves towards Java 1.5 the possible compabillity with
> >> J2ME moves further away.
> >>
> >>>       Or is there any other J2ME based API which can be used to
> >>> search the
> >>> text in mobile phones.
> >>
> >> Before asking too much you should tell us some about your project.
> >> How large is the corpus, what features do you require, et c? I.e. do
> >> you really need Lucene, is it perhaps sufficient with something
> >> simple as a trie-pattern?
> >>
> >> --
> >> karl
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
> >> For additional commands, e-mail: java-user-help@lucene.apache.org
> >>
> >>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-user-help@lucene.apache.org
>
>

Re: Re : porting lucene on to mobile phone

Posted by Karl Wettin <ka...@gmail.com>.
8 okt 2007 kl. 17.40 skrev saikrishna venkata pendyala:

>  I have abt 17,00,000 documents in 1000 files. I need to index the  
> titles of
> those documents with out storing the titles and storing the file name.

Not sure I understand what you say here.

You want to create 1,7 million documents containing a short and  
indexed field: title, and a stored field: file name?

> When ever I issue a query I need to get the name of the file in  
> which that
> document exists.

You only require boolean term queries? And no scoring, sorting or  
anything like that?

In that case, it should not be that hard to rewrite Lucene for your  
needs.

See LUCENE-584 for seperation of Matcher and Scorer.


-- 
karl


>
> thanks,
> Saikrishna.
>
> On 10/8/07, Karl Wettin <ka...@gmail.com> wrote:
>>
>>
>> 8 okt 2007 kl. 15.36 skrev saikrishna venkata pendyala:
>>
>>>       Has any one tried to port lucene on to a mobile phone. I  
>>> need a
>>> text-based search engine for mobile phone as part of my project.
>>
>>
>> There was some discussion about this here on the forums some year(s?)
>> ago. The major problem back then was the lack of floaing points.
>> Also, as Lucene moves towards Java 1.5 the possible compabillity with
>> J2ME moves further away.
>>
>>>       Or is there any other J2ME based API which can be used to
>>> search the
>>> text in mobile phones.
>>
>> Before asking too much you should tell us some about your project.
>> How large is the corpus, what features do you require, et c? I.e. do
>> you really need Lucene, is it perhaps sufficient with something
>> simple as a trie-pattern?
>>
>> --
>> karl
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
>> For additional commands, e-mail: java-user-help@lucene.apache.org
>>
>>


---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org


Re: Re : porting lucene on to mobile phone

Posted by saikrishna venkata pendyala <pv...@gmail.com>.
Hai,

     Thank you for responding.

 I have abt 17,00,000 documents in 1000 files. I need to index the titles of
those documents with out storing the titles and storing the file name.

When ever I issue a query I need to get the name of the file in which that
document exists.

thanks,
Saikrishna.

On 10/8/07, Karl Wettin <ka...@gmail.com> wrote:
>
>
> 8 okt 2007 kl. 15.36 skrev saikrishna venkata pendyala:
>
> >       Has any one tried to port lucene on to a mobile phone. I need a
> > text-based search engine for mobile phone as part of my project.
>
>
> There was some discussion about this here on the forums some year(s?)
> ago. The major problem back then was the lack of floaing points.
> Also, as Lucene moves towards Java 1.5 the possible compabillity with
> J2ME moves further away.
>
> >       Or is there any other J2ME based API which can be used to
> > search the
> > text in mobile phones.
>
> Before asking too much you should tell us some about your project.
> How large is the corpus, what features do you require, et c? I.e. do
> you really need Lucene, is it perhaps sufficient with something
> simple as a trie-pattern?
>
> --
> karl
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-user-help@lucene.apache.org
>
>

Re: Re : porting lucene on to mobile phone

Posted by Karl Wettin <ka...@gmail.com>.
8 okt 2007 kl. 15.36 skrev saikrishna venkata pendyala:

>       Has any one tried to port lucene on to a mobile phone. I need a
> text-based search engine for mobile phone as part of my project.


There was some discussion about this here on the forums some year(s?)  
ago. The major problem back then was the lack of floaing points.  
Also, as Lucene moves towards Java 1.5 the possible compabillity with  
J2ME moves further away.

>       Or is there any other J2ME based API which can be used to  
> search the
> text in mobile phones.

Before asking too much you should tell us some about your project.  
How large is the corpus, what features do you require, et c? I.e. do  
you really need Lucene, is it perhaps sufficient with something  
simple as a trie-pattern?

-- 
karl

---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org