• datamapper.org

• github.com/datamapper/dm-is-searchable

A DataMapper search plugin api to search for resources from one repository and load from another.

Typically a full text search adapter that can only produce partial resources is searched and the resulting resource collection is then loaded from your default repository.

require "rubygems"
require "dm-core"
require "dm-is-searchable"

DataMapper.setup(:default, :adapter => 'in_memory')

# Your search adapter configuration. See your search adapters setup options.
DataMapper.setup(:search, :adapter => 'your_adapter')

class Cow
  include DataMapper::Resource
  property :name, String, :key => true
  property :likes, String
  property :cowpats, Integer

  is :searchabe # This defaults to repository(:search), you could also do
  # is :searchable, :repository => :some_searchable_repository
end

class Chicken
  include DataMapper::Resource
  property :name,  String, :key => true
  property :likes, String
  property :eggs, Integer

  is :searchable
  repository(:search) do
    # We only want to be able to search by name.
    properties(:search).clear
    property :name, String
  end
end

Mixing in is-searchable by default defines a single class method in your resource with the signature of:

# ==== Parameters
# search_options<Hash>::
#   DM::Query conditions to pass to the searchable repository. Unless you
#   explicitly defined a searchable repository this is repository(:search).
#
# options<Hash>::
#   Optionsal DM::Query conditions to pass to the repository holding the full
#   resource. Without a scoped search this is repository(:default).
#
# ==== Returns
# DM::Collection:: Zero or more DM::Resource objects.
MyModel#search(search_options = {}, options = {})

A basic full text search for cows called ‘Pete’, ‘Peter’, ‘Pedro’ (dependant on the search adapter) would look like:

puts Cow.search(:name => 'pete').inspect
#=> [<Cow name="peter" cowpats="12" ...>, <Cow name='pete' cowpats="1024" ...>, ...]

Adding extra conditions to apply to the default repository allows you to do interesting things to the search adapter results. For example conditions on properties not available in your search adapter or unsupported operators:

# Unsupported #gt operator in search adapter.
puts Cow.search({:name => 'pete'}, {:cowpats.gt => 1000}).inspect
#=> [<Cow name="pete" cowpats="1024" ...>]

# Unknown property in search adapter.
puts Chicken.search({:name => 'steve'}, {:eggs => (100..200)}).inspect
#=> [<Chicken name="steve" eggs="120" ...>]

Like all DM adapters a custom search adapter implements the DM::AbstractAdapter interface.

The key differences from a typical adapter are:

An Array of Hashes in the form of [{:id => 12}, {:id => 53}] will work just fine. In fact so long as the following snippet would run on the returned value you are free to return whatever you like.

ids = read_many_result.collect{|doc| doc[:id]} #=> Array of ID's.

No need to DM::Model#load here. Just return your partial resource as a Hash in the form of {:id => 12}. Like #read_many the returned value really only needs to respond to #[:id].

read_one_result[:id] #=> The resource ID to load.

Gem

dm-ferret-adapter

Git

git://github.com/datamapper/dm-ferret-adapter.git

(not currently longer maintained)

Gem

dm-sphinx-adapter

Git

git://github.com/shanna/dm-sphinx-adapter.git