[Student-projects] Blog update

[Sujith V]

Hello Aashik,

The reason why I moved to sqlite is that while I was working on
transliteration module I observed it to was faster to store CMU dictionary
in an sqlite database than to parse data (approx 3.7 MB) into map which was
getting slower on adding roughly  1.3 lac elements on each object creation.
So when I tried this approach and it seemed to run much faster. After that
I was started experimenting on other modules too. For other modules like
Payyans, Stemmer I have retained text file rules as there were no
comparable difference as files were very small. But I have changed for
fortune module.
Jishnu Mohan had earlier suggested a method to convert files into binary
data files for faster processing however I am still left to experiment that
approach.
Please let me know what you feel about this approach.

Regards


On Wed, Jul 30, 2014 at 4:11 PM, Aashik Shahmiya <aashiks at gmail.com> wrote:

>  Hello Sujith,
>
> May I ask for your reasons for pushing rules into SQLite database (rather
> than text files)?
>
> Regards
> Ashik Salahudeen
>
>
> On 07/30/2014 12:40 AM, Sujith V wrote:
>
>
> Hello,
>
>  Here is
> Link to blog : http://hashtagcoding.wordpress.com/
>  Link to project : https://github.com/SujithVadakkepat/silpa-sdk
>
>  Thank you
> --
> Sujith V
> PES Institute of Technology, Bangalore
>
>
>


-- 
Sujith V
PES Institute of Technology, Bangalore
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.smc.org.in/pipermail/student-projects-smc.org.in/attachments/20140730/7dc95052/attachment.html>