Hi Dara,
I hereby copy the developer mailing list to broaden the discussion. Below more responses.
On 29 May 2007, at 11:03 AM, Vertex Dara.Dowd wrote:
Hi Jeroen,
Thanks for your mail. I was away on holidays so only catching up now.
Jeroen Ticheler wrote:
- Yes, a language check seems kind of logical as a solution. Is the 80Mb space requirements something that is part of the installer!? I assume not, but just to be sure
Otherwise, if that's for the indexing on the system it doesn't seem a big drawback to me. I'm sure there are issues that need to be resolved in this respect also for the other languages; for instance ensure that the Lucene indexes take care of the language the metadata is in while indexing. As far as I know it doesn't do that now.
The 80Mb isn't part of the installer itself. It's just the size of the dictionary after it has been built.
I see no problem there.
- Would you like to take this to the developer mailing list and possibly to a test environment on a branch in the Subversion?
Sounsd good. How do I go about getting access to the SVN repository?
I created a branch on SVN (). Make sure you keep that branch in synch with what's new in the trunk regularly so you keep yours closely aligned! This will make merging much easier when ready.
Also, don't commit to the trunk at this stage! Make sure you discuss anything you want to commit to the trunk before doing so until we feel comfortable with the proposed changes
- Also, you may want to create a WIKI page on GeoNetwork opensource Developer website under R&D to describe what you want to do?
Will do. Should I just go ahead and create an OSGEO login?
Yes
I will need both your OSGEO and SourceForge.net usernames (you can send them directly to me if you wish)
Ciao,
Jeroen
Cheers,
Dara
Ciao,
JeroenOn 21 May 2007, at 5:50 AM, donal_regan10@...266... wrote:
Hi Jeroen,
Thank you for reply and apologies for the delay in getting back to you.
I too think it would be better to have a package that allows searching in any language. The Japanese Analyzer allows english to be searched without any problems, however other languages are problematic. For example, Français and Español get indexed as "Fran" "ç" "ais" and "Espa" "ñ" "ol" respectively.
The simplest solution would probably be to allow the user to specify which language (via a select box maybe) the search string is in and to then use the appropriate Analyzer.
The main change I made to the source was the following in SearchManager.java
IndexWriter writer = new IndexWriter(_luceneDir, new StandardAnalyzer(new String {}), false);
becomes
IndexWriter writer = new IndexWriter(_luceneDir, new JapaneseAnalyzer(), false);
It would be easy to select between the two depending on search language.
One possible issue though is that the Japanese Analyzer requires a dictionary to be built which takes up over 80 MB of space.
Best regards,
Dara
Jeroen Ticheler-3 wrote:
Hi Dara,
That's an interesting proposal. The big issue to be discussed now is
how to (1) integrate this in a way both searching in Japanese and in
other languages work, or (2) add this as a separate package that can
be used to build a Japanese only version.Do you have any ideas? My preference would be the first option, but
it may be impossible and may not even needed. Could it be done in a
way that allows a user to select the language of the metadata he
wants to search?Once a good approach is found, we could start with adding your code
to a branch. Using an Ant build script it should be possible to build
one of the two solutions quickly.Ciao,
JeroenOn 11 May 2007, at 9:41 AM, Dara Dowd wrote:
Hello,
Our team have developed a Japanese version of Geonetwork and based
on the
GPL would like to contribute it to the community.We noticed that the metadata search functionality didn't work for
Japanese
data. The reason for this being that the Lucene Analyzer shipped with
Geonetwork doesn't support Japanese. To rectify this we decided to
add a
'sen' based Japanese Analyzer to geonetwork. (sen:
https://sen.dev.java.net/)One thing is that this localized version is currently only suitable
for
Japanese searching.What would the best way be to send our code?
Best regards,
Dara Dowd
(sent on behalf of Akira Sugawara)
--View this message in context: http://www.nabble.com/Japanese-
version-tf3725954s18419.html#a10427333
Sent from the geonetwork-devel mailing list archive at Nabble.com.----------------------------------------------------------------------
---
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
GeoNetwork-devel mailing list
GeoNetwork-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-devel
GeoNetwork OpenSource is maintained at http://sourceforge.net/
projects/geonetwork-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
GeoNetwork-devel mailing list
GeoNetwork-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-devel
GeoNetwork OpenSource is maintained at
http://sourceforge.net/projects/geonetworkQuoted from:
http://www.nabble.com/Japanese-version-tf3725954s18419.html#a10427997