But its wipes out the existing index and creates a new empty one - the
pages for Ferret are currently unavailable - just wondering anybody else
has come across this? Is the 'create=>false' just not working or do I
have it wrong here ? ( I have also tried : create=>'false' - still the
same thing).
Cheers
John
How do I get the version of Ferret I'm using by the way - I would have
posted that here if I knew how !
-- Search.rb --
require 'rubygems'
require 'ferret'
index=Ferret::I.new(:path => 'index', :create=>false);
index.search_each('thane') do |doc,score|
puts doc, score
end
index.close
-- end
···
--
Posted via http://www.ruby-forum.com/.
As far as i know, Ferret-built indexes are not binary-compatible with Lucene anymore.
To find out the version, open a console and type:
gem list ferret
the highest version behind the ferret gem is the one in use.
Regards,
Florian Gilcher
···
On Nov 2, 2008, at 11:03 PM, John Pritchard-williams wrote:
But its wipes out the existing index and creates a new empty one - the
pages for Ferret are currently unavailable - just wondering anybody else
has come across this? Is the 'create=>false' just not working or do I
have it wrong here ? ( I have also tried : create=>'false' - still the
same thing).
Cheers
John
How do I get the version of Ferret I'm using by the way - I would have
posted that here if I knew how !
-- Search.rb --
require 'rubygems'
require 'ferret'
index=Ferret::I.new(:path => 'index', :create=>false);
index.search_each('thane') do |doc,score|
puts doc, score
end
index.close
-- end
--
Posted via http://www.ruby-forum.com/\.
Thanks Florian, that would be shame if we've lost that binary
compatibility ah well...
The version is:
*** LOCAL GEMS ***
ferret (0.11.6, 0.11.4)
And ignoring the Lucene/Ferret thing for the moment - the 'search.rb'
program above seems to write over the index directory in any case (if I
run it twice , I get new mod times on the 'segments' files at least).
Cheers
John
···
--
Posted via http://www.ruby-forum.com/.
I don't remember the reasoning behind it, but as far as i know, this happend quite a while ago.
If you search http://www.ruby-forum.com/forum/5 for it, you might get more.
I think the ability to use Lucene directly from JRuby already makes the ability to manipulate it
with ferret a cornercase most people will ignore.
Regards,
Florian Gilcher
···
On Nov 2, 2008, at 11:45 PM, John Pritchard-williams wrote:
Thanks Florian, that would be shame if we've lost that binary
compatibility ah well...
The version is:
*** LOCAL GEMS ***
ferret (0.11.6, 0.11.4)
And ignoring the Lucene/Ferret thing for the moment - the 'search.rb'
program above seems to write over the index directory in any case (if I
run it twice , I get new mod times on the 'segments' files at least).
Cheers
John
--
Posted via http://www.ruby-forum.com/\.
I have not explicitly checked for this behaviour with the ff.rb
I modified (without touching that bit), but certainly the index is not
rebuilt, that would take ages.
http://www.eng.cse.dmu.ac.uk/~hgs/ruby/ff.rb
http://www.eng.cse.dmu.ac.uk/~hgs/ruby/ferret_helper.rb
See
http://www.eng.cse.dmu.ac.uk/~hgs/ruby/#ff.rb
for a bit of blurb about what I did.
You might like to experiment with that to see if its behaviour is useful.
[For those reading the archive: that machine is doomed to die, and the
URL will change. I'd like to keep this stuff on the web somewhere, but
I'm not sure how that will work out]
Hugh
···
On Mon, 3 Nov 2008, John Pritchard-williams wrote:
And ignoring the Lucene/Ferret thing for the moment - the 'search.rb'
program above seems to write over the index directory in any case (if I
run it twice , I get new mod times on the 'segments' files at least).