Ferret talk - Oct 2006 - hello, acts_as_ferret questions, any help greatly appreciate

hi, ive been reading up on ferret, acts_as_ferret, and other search 
plugins for rails.

after reading about ferret, i found out about the acts_as_ferrt plugin.

my first question about acts_as_ferret:

1. from reading about ferret, do i still need to manually save the IDX 
and add a IDX column field to my model table for acts_as_ferret to work? 
they say that acts_as_ferret handles everything, but i wasnt sure what 
exactly does it handle when compared to the ferret tutorials that i 
read.

2. is there a complete tutorial online for acts_as_ferret plugin? so far 
all the blogs that i came across pretty much say the same thing.

3. does acts_as_ferret do this? if i specify acts_as_ferret to index 
just title column in my ''recipe'' table, and if someone types
''bbq plate''
in the search box, will these recipes with the following title match?

"island bbq style plate"
"bbq ribs"
"potato plate salad"
"my bbq plate"

if so, will the one with the most matching text like "my bbq plate" be
listed as #1?


Thanks for the help!

-- 
Posted via http://www.ruby-forum.com/.

On 10/9/06, bbqTree <none at none.com> wrote:
> hi, ive been reading up on ferret, acts_as_ferret, and other search
> plugins for rails.
>
> after reading about ferret, i found out about the acts_as_ferrt plugin.
>
> my first question about acts_as_ferret:
>
> 1. from reading about ferret, do i still need to manually save the IDX
> and add a IDX column field to my model table for acts_as_ferret to work?
> they say that acts_as_ferret handles everything, but i wasnt sure what
> exactly does it handle when compared to the ferret tutorials that i
> read.
Personally I don''t use acts_as_ferret but I''ll try and answer
these
questions as best I can. The following will automatically index the
fields of the Foo model:

 class Foo < ActiveRecord::Base
   acts_as_ferret
 end

See http://projects.jkraemer.net/acts_as_ferret/wiki/AdvancedUsage for
more advanced usage.
> 2. is there a complete tutorial online for acts_as_ferret plugin? so far
> all the blogs that i came across pretty much say the same thing.
That above link is the most complete tutorial I know of. Maybe someone
else will know of other resources. If you need to do anything more
advanced then you should probably learn more about Ferret itself.
> 3. does acts_as_ferret do this? if i specify acts_as_ferret to index
> just title column in my ''recipe'' table, and if someone
types ''bbq plate''
> in the search box, will these recipes with the following title match?
>
> "island bbq style plate"
> "bbq ribs"
> "potato plate salad"
> "my bbq plate"
>
> if so, will the one with the most matching text like "my bbq
plate" be
> listed as #1?
Ferret is pretty easy to experiment with. Here is an example:

    require ''rubygems''
    require ''ferret''

    index = Ferret::I.new

    [
      "island bbq style plate",
      "bbq ribs",
      "potato plate salad",
      "my bbq plate"
    ].each {|text| index << text}

    puts index.search(''bbq plate'')
    puts index.search(''"bbq plate"'')

And the output:

    TopDocs: total_hits = 4, max_score = 0.883883 [
            3 "my bbq plate": 0.883883
            0 "island bbq style plate": 0.707107
            1 "bbq ribs": 0.220971
            2 "potato plate salad": 0.176777
    ]
    TopDocs: total_hits = 1, max_score = 1.250000 [
            3 "my bbq plate": 1.250000
    ]

So to answer your question, yes, the results are ordered by relevency.

> Thanks for the help!
no problem.

Dave

On Mon, Oct 09, 2006 at 03:27:34PM +0200, bbqTree wrote:
> hi, ive been reading up on ferret, acts_as_ferret, and other search 
> plugins for rails.
> 
> after reading about ferret, i found out about the acts_as_ferrt plugin.
> 
> my first question about acts_as_ferret:
> 
> 1. from reading about ferret, do i still need to manually save the IDX 
> and add a IDX column field to my model table for acts_as_ferret to work? 
> they say that acts_as_ferret handles everything, but i wasnt sure what 
> exactly does it handle when compared to the ferret tutorials that i 
> read.
I don''t know what that IDX column should be good for, where did you
read
about that ?

basically acts_as_ferret can handle everything for you. just call
acts_as_ferret in your model class definition, and you''ll get (besides
other goodies and methods for special use cases):
- automatic index creation for existing records
- new records are automatically added to the index
- automatic index updates, when you update your model data
- automatic removal of documents from the index when the corresponding
  model object is destroyed.
- Model.find_by_contents(''querystring'') to retrieve model
instances
  matching the querystring.

the acts_as_ferret method offers several options to customize the way
aaf works, i.e. which fields of your model should get indexed, and with
which indexing options (refer to the api docs at
http://projects.jkraemer.net/acts_as_ferret/rdoc)
> 2. is there a complete tutorial online for acts_as_ferret plugin? so far 
> all the blogs that i came across pretty much say the same thing.
the API docs for the acts_as_ferret and find_by_content methods should
get you started. You might also want to check out the demo project from
the acts_as_ferret svn (see the wiki at
http://projects.jkraemer.net/acts_as_ferret/ for the repository url)
> 3. does acts_as_ferret do this? if i specify acts_as_ferret to index 
> just title column in my ''recipe'' table, and if someone
types ''bbq plate''
> in the search box, will these recipes with the following title match?
> 
> "island bbq style plate"
> "bbq ribs"
> "potato plate salad"
> "my bbq plate"
since acts_as_ferret defaults to ANDed queries, only records 1 and 4
would match. you can however make aaf use ORed queries by specifying 
:or_default => false in your call to acts_as_ferret.
> if so, will the one with the most matching text like "my bbq
plate" be
> listed as #1?
I suppose it would be #1, not because the exact phrase matches, but
because it''s the most relevant hit (the overall size of the text is
smaller than that of the other document matching the query, and a match
in a short text generally scores higher than one in a large text).
Anybody please correct me if I''m wrong here.


cheers,
Jens

-- 
webit! Gesellschaft f?r neue Medien mbH          www.webit.de
Dipl.-Wirtschaftsingenieur Jens Kr?mer       kraemer at webit.de
Schnorrstra?e 76                         Tel +49 351 46766  0
D-01069 Dresden                          Fax +49 351 46766 66