stretcher 1.3.2 → 1.3.3

data/lib/stretcher.rb

@@ -5,6 +5,8 @@ require 'faraday'
 require 'faraday_middleware'
 require "stretcher/version"
 require 'stretcher/request_error'
+require 'stretcher/search_results'
+require 'stretcher/es_component'
 require 'stretcher/server'
 require 'stretcher/index'
 require 'stretcher/index_type'

data/lib/stretcher/es_component.rb

@@ -0,0 +1,24 @@
+module Stretcher
+  # Elasticsearch has symmetry across API endpoints for Server, Index, and Type, lets try and provide some common ground
+  class EsComponent
+
+    private
+    
+    def do_search(generic_opts={}, explicit_body=nil)
+      uri_str = '/_search'
+      body = nil
+      if explicit_body
+        uri_str << '?' + Util.querify(generic_opts)
+        body = explicit_body
+      else
+        body = generic_opts
+      end
+
+      logger.info { "Stretcher Search: curl -XGET '#{uri_str}' -d '#{body.to_json}'" }
+      response = @server.request(:get, path_uri(uri_str)) do |req|
+        req.body = body
+      end
+      SearchResults.new(:raw => response)      
+    end
+  end
+end

data/lib/stretcher/index.rb

@@ -2,9 +2,9 @@ require 'stretcher/search_results'
 module Stretcher
   # Represents an Index context in elastic search.
   # Generally should be instantiated via Server#index(name).
-  class Index
+  class Index < EsComponent
     attr_reader :server, :name, :logger
-    
+
     def initialize(server, name, options={})
       @server = server
       @name = name
@@ -14,14 +14,14 @@ module Stretcher
     # Returns a Stretcher::IndexType object for the type +name+.
     # Optionally takes a block, which will be passed a single arg with the Index obj
     # The block syntax returns the evaluated value of the block
-    # 
+    #
     #    my_index.index(:foo) # => #<Stretcher::Index ...>
     #    my_index.index(:foo) {|idx| 1+1} # => 2
     def type(name, &block)
       t = IndexType.new(self, name)
       block ? block.call(t) : t
     end
-    
+
     # Given a hash of documents, will bulk index
     #
     #    docs = [{"_type" => "tweet", "_id" => 91011, "text" => "Bulked"}]
@@ -42,7 +42,7 @@ module Stretcher
         req.body = options
       end
     end
-    
+
     # Deletes the index
     def delete
       @server.request :delete, path_uri
@@ -52,27 +52,27 @@ module Stretcher
     def stats
       @server.request :get, path_uri("/_stats")
     end
-    
+
     # Retrieves status for this index
     def status
-      @server.request :get, path_uri("/_status")      
+      @server.request :get, path_uri("/_status")
     end
-    
+
     # Retrieve the mapping for this index
     def get_mapping
       @server.request :get, path_uri("/_mapping")
     end
-    
+
     # Retrieve settings for this index
     def get_settings
       @server.request :get, path_uri("/_settings")
     end
-    
+
     # Check if the index has been created on the remote server
     def exists?
       @server.http.head(path_uri).status != 404
     end
-    
+
     # Issues a search with the given query opts and body, both should be hashes
     #
     #    res = server.index('foo').search(size: 12, {query: {match_all: {}}})
@@ -82,22 +82,10 @@ module Stretcher
     #    res.results # => [#<Hashie::Mash _id="123" text="Hello">]
     #    res.raw     # => #<Hashie::Mash ...> Raw JSON from the search
     def search(generic_opts={}, explicit_body=nil)
-      uri_str = '/_search'
-      body = nil
-      if explicit_body
-        uri_str << '?' + Util.querify(generic_opts)
-        body = explicit_body
-      else
-        body = generic_opts
-      end
-      
-      logger.info { "Stretcher Search: curl -XGET '#{uri_str}' -d '#{body.to_json}'" }
-      response = @server.request(:get, path_uri(uri_str)) do |req|
-        req.body = body
-      end
-      SearchResults.new(raw: response)
+      # Written this way to be more RDoc friendly
+      do_search(generic_opts, explicit_body)
     end
-    
+
     # Searches a list of queries against only this index
     # This deviates slightly from the official API in that *ONLY*
     # queries are requried, the empty {} preceding them are not
@@ -108,7 +96,7 @@ module Stretcher
     def msearch(queries=[])
       raise ArgumentError, "msearch takes an array!" unless queries.is_a?(Array)
       req_body = queries.reduce([]) {|acc,q|
-        acc << {index: name}
+        acc << {:index => name}
         acc << q
         acc
       }
@@ -116,7 +104,7 @@ module Stretcher
     end
 
     # Implements the Analyze API
-    # EX: 
+    # EX:
     #    index.analyze("Candles", analyzer: :snowball)
     #    # => #<Hashie::Mash tokens=[#<Hashie::Mash end_offset=7 position=1 start_offset=0 token="candl" type="<ALPHANUM>">]>
     def analyze(text, analysis_params)

data/lib/stretcher/index_type.rb

@@ -1,9 +1,9 @@
 module Stretcher
   # Represents an index  scoped to a specific type.
   # Generally should be instantiated via Index#type(name).
-  class IndexType
+  class IndexType < EsComponent
     attr_reader :server, :index, :name, :logger
-    
+
     def initialize(index, name, options={})
       @index = index
       @server = index.server
@@ -17,12 +17,12 @@ module Stretcher
       res = server.request(:get, path_uri("/#{id}"))
       raw ? res : res["_source"]
     end
-    
+
     # Index an item with a specific ID
     def put(id, source)
       server.request(:put, path_uri("/#{id}"), source)
     end
-    
+
     # Uses the update api to modify a document with a script
     # To update a doc with ID 987 for example:
     # type.update(987, script: "ctx._source.message = 'Updated!'")
@@ -30,11 +30,11 @@ module Stretcher
     def update(id, body)
       server.request(:post, path_uri("/#{id}/_update"), body)
     end
-    
+
     # Deletes the document with the given ID
     def delete(id)
       res = server.http.delete path_uri("/#{id}")
-      
+
       # Since 404s are just not a problem here, let's simply return false
       if res.status == 404
         false
@@ -49,7 +49,7 @@ module Stretcher
     def get_mapping
       @server.request :get, path_uri("/_mapping")
     end
-    
+
     # Delete the mapping for this type. Note this will delete
     # All documents of this type as well
     # http://www.elasticsearch.org/guide/reference/api/admin-indices-delete-mapping.html
@@ -63,7 +63,7 @@ module Stretcher
         req.body = body
       }
     end
-    
+
     # Check if this index type is defined, if passed an id
     # this will check if the given document exists
     def exists?(id=nil)
@@ -72,8 +72,9 @@ module Stretcher
 
     # Issues an Index#search scoped to this type
     # See Index#search for more details
-    def search(generic_opts={}, body=nil)
-      @index.search(generic_opts.merge(type: name), body)
+    def search(generic_opts={}, explicit_body=nil)
+      # Written this way to be more RDoc friendly
+      do_search(generic_opts, explicit_body)
     end
 
     # Full path to this index type

data/lib/stretcher/search_results.rb

@@ -1,25 +1,25 @@
 require 'hashie/dash'
 module Stretcher
   # Conveniently represents elastic search results in a more compact fashion
-  # 
+  #
   # Available properties:
-  # 
+  #
   # * raw : The raw response from elastic search
   # * total : The total number of matched docs
   # * facets : the facets hash
   # * results : The hit results with _id merged in to _source
   class SearchResults < Hashie::Dash
-    property :raw, required: true
+    property :raw, :required => true
     property :total
     property :facets
     property :results
-    
+
     def initialize(*args)
       super
       self.total = raw.hits.total
       self.facets = raw.facets
       self.results = raw.hits.hits.collect {|r|
-        r['_source'].merge({"_id" => r['_id']})
+        (r.has_key?('_source') ? r['_source'] : r['fields']).merge({"_id" => r['_id']})
       }
     end
   end

data/lib/stretcher/server.rb

@@ -1,5 +1,5 @@
 module Stretcher
-  class Server
+  class Server < EsComponent
     attr_reader :uri, :http, :logger
 
     # Instantiate a new instance in a manner convenient for using the block syntax.
@@ -9,7 +9,7 @@ module Stretcher
       s = self.new(*args)
       yield s
     end
-    
+
     # Represents a Server context in elastic search.
     # Use +with_server+ when you want to use the block syntax.
     # The options hash takes an optional instance of Logger under :logger.
@@ -18,25 +18,25 @@ module Stretcher
     def initialize(uri='http://localhost:9200', options={})
       @uri = uri
 
-      @http = Faraday.new(:url => @uri) do |builder|        
+      @http = Faraday.new(:url => @uri) do |builder|
         builder.response :mashify
         builder.response :json, :content_type => /\bjson$/
-        
+
         builder.request :json
-        
+
         builder.adapter :net_http_persistent
-        
+
         builder.options[:read_timeout] = 4
         builder.options[:open_timeout] = 2
       end
-      
+
       if options[:logger]
         @logger = options[:logger]
       else
         @logger = Logger.new(STDOUT)
         @logger.level = Logger::WARN
       end
-      
+
       @logger.formatter = proc do |severity, datetime, progname, msg|
         "[Stretcher][#{severity}]: #{msg}\n"
       end
@@ -45,11 +45,11 @@ module Stretcher
     # Returns a Stretcher::Index object for the index +name+.
     # Optionally takes a block, which will be passed a single arg with the Index obj
     # The block syntax returns the evaluated value of the block
-    # 
+    #
     #    my_server.index(:foo) # => #<Stretcher::Index ...>
     #    my_server.index(:foo) {|idx| 1+1} # => 2
     def index(name, &block)
-      idx = Index.new(self, name, logger: logger)
+      idx = Index.new(self, name, :logger => logger)
       block ? block.call(idx) : idx
     end
 
@@ -68,13 +68,13 @@ module Stretcher
     def status
       request :get, path_uri("/_status")
     end
-    
+
     # Returns true if the server is currently reachable, raises an error otherwise
     def up?
       request(:get, path_uri)
       true
     end
-    
+
     # Takes an array of msearch data as per
     # http://www.elasticsearch.org/guide/reference/api/multi-search.html
     # Should look something like:
@@ -92,13 +92,13 @@ module Stretcher
       res = request(:get, path_uri("/_msearch")) {|req|
         req.body = fmt_body
       }
-      
+
       errors = res.responses.select {|r| r[:error]}.map(&:error)
       if !errors.empty?
-        raise RequestError.new(res), "Could not msearch #{errors.inspect}" 
+        raise RequestError.new(res), "Could not msearch #{errors.inspect}"
       end
-      
-      res['responses'].map {|r| SearchResults.new(raw: r)}
+
+      res['responses'].map {|r| SearchResults.new(:raw => r)}
     end
 
     # Retrieves multiple documents, possibly from multiple indexes
@@ -108,9 +108,9 @@ module Stretcher
         req.body = body
       }
     end
-    
+
     # Implements the Analyze API
-    # EX: 
+    # EX:
     #    server.analyze("Candles", analyzer: :snowball)
     #    # => #<Hashie::Mash tokens=[#<Hashie::Mash end_offset=7 position=1 start_offset=0 token="candl" type="<ALPHANUM>">]>
     # as per: http://www.elasticsearch.org/guide/reference/api/admin-indices-analyze.html
@@ -128,7 +128,7 @@ module Stretcher
     # Handy way to query the server, returning *only* the body
     # Will raise an exception when the status is not in the 2xx range
     def request(method, *args, &block)
-      logger.info { "Stretcher: Issuing Request #{method.upcase}, #{args}" }
+      logger.info { "Stretcher: Issuing Request #{method.to_s.upcase}, #{args}" }
       res = if block
               http.send(method, *args) do |req|
                 # Elastic search does mostly deal with JSON

data/lib/stretcher/version.rb

@@ -1,3 +1,3 @@
 module Stretcher
-  VERSION = "1.3.2"
+  VERSION = "1.3.3"
 end

data/spec/lib/stretcher_index_spec.rb

@@ -2,19 +2,19 @@ require 'spec_helper'
 
 describe Stretcher::Index do
   let(:server) {
-    Stretcher::Server.new(ES_URL, logger: DEBUG_LOGGER)
+    Stretcher::Server.new(ES_URL, :logger => DEBUG_LOGGER)
   }
   let(:index) { server.index('foo') }
   let(:corpus) {
     [
-     {text: "Foo", "_type" => 'tweet', "_id" => 'fooid'},
-     {text: "Bar", "_type" => 'tweet', "_id" => 'barid'},
-     {text: "Baz", "_type" => 'tweet', "id" => 'bazid'} # Note we support both _id and id
+     {:text => "Foo", "_type" => 'tweet', "_id" => 'fooid'},
+     {:text => "Bar", "_type" => 'tweet', "_id" => 'barid'},
+     {:text => "Baz", "_type" => 'tweet', "id" => 'bazid'} # Note we support both _id and id
     ]
   }
 
   def create_tweet_mapping
-    mdata = {tweet:{properties: {text: {type: :string}}}}
+    mdata = {:tweet => {:properties => {:text => {:type => :string}}}}
     index.type('tweet').put_mapping(mdata)
   end
 
@@ -22,7 +22,7 @@ describe Stretcher::Index do
     create_tweet_mapping
     index.bulk_index(corpus)
   end
-  
+
   it "should work on an existential level" do
     index.delete rescue nil
     index.exists?.should be_false
@@ -67,26 +67,26 @@ describe Stretcher::Index do
     seed_corpus
     match_text = corpus.first[:text]
     sleep 1 # ES needs time to update!
-    res = index.search({}, {query: {match: {text: match_text}}})
+    res = index.search({}, {:query => {:match => {:text => match_text}}})
     res.results.first.text.should == match_text
   end
 
   # TODO: Actually use two indexes
   it "should msearch across the index returning all docs" do
     seed_corpus
-    res = index.msearch([{query: {match_all: {}}}])
+    res = index.msearch([{:query => {:match_all => {}}}])
     res.length.should == 1
     res[0].class.should == Stretcher::SearchResults
   end
 
   it "execute the analysis API and return an expected result" do
-    analyzed = index.analyze("Candles", analyzer: :snowball)
+    analyzed = index.analyze("Candles", :analyzer => :snowball)
     analyzed.tokens.first.token.should == 'candl'
   end
 
   it "should raise an exception when msearching a non-existant index" do
     lambda {
-      res = server.index(:does_not_exist).msearch([{query: {match_all: {}}}])
+      res = server.index(:does_not_exist).msearch([{:query => {:match_all => {}}}])
     }.should raise_exception(Stretcher::RequestError)
   end
 end

data/spec/lib/stretcher_index_type_spec.rb

@@ -16,27 +16,32 @@ describe Stretcher::Index do
 
   describe "searching" do
     before do
-      @doc = {message: "hello"}
+      @doc = {:message => "hello"}
     end
 
     it "should search and find the right doc" do
       match_text = 'hello'
       type.put(123123, @doc)
       sleep 1
-      res = type.search({}, {query: {match: {message: match_text}}})
+      res = type.search({}, {:query => {:match => {:message => match_text}}})
+      res.results.first.message.should == @doc[:message]
+    end
+
+    it "should build results when _source is not included in loaded fields" do
+      res = type.search({}, {query: {match_all: {}}, fields: ['message']})
       res.results.first.message.should == @doc[:message]
     end
   end
 
   describe "put/get" do
     before do
-      @doc = {message: "hello!"}
+      @doc = {:message => "hello!"}
     end
-    
+
     it "should put correctly" do
       type.put(987, @doc).should_not be_nil
     end
-    
+
     it "should get individual documents correctly" do
       type.get(987).message.should == @doc[:message]
     end
@@ -48,14 +53,14 @@ describe Stretcher::Index do
     end
 
     it "should update individual docs correctly" do
-      type.update(987, script: "ctx._source.message = 'Updated!'")
+      type.update(987, :script => "ctx._source.message = 'Updated!'")
       type.get(987).message.should == 'Updated!'
     end
-    
+
     it "should delete individual docs correctly" do
       type.exists?(987).should be_true
       type.delete(987)
-      type.exists?(987).should be_false      
+      type.exists?(987).should be_false
     end
   end
 end

data/spec/lib/stretcher_server_spec.rb

@@ -2,7 +2,7 @@ require 'spec_helper'
 
 describe Stretcher::Server do
   let(:server) { Stretcher::Server.new(ES_URL) }
-  
+
   it "should initialize cleanly" do
     server.class.should == Stretcher::Server
   end
@@ -28,7 +28,7 @@ describe Stretcher::Server do
   it "should check the status w/o error" do
     server.status.ok.should be_true
   end
-  
+
   it "should beget an index object cleanly" do
     server.index('foo').class.should == Stretcher::Index
   end
@@ -44,7 +44,7 @@ describe Stretcher::Server do
   end
 
   it "execute the analysis API and return an expected result" do
-    analyzed = server.analyze("Candles", analyzer: :snowball)
+    analyzed = server.analyze("Candles", :analyzer => :snowball)
     analyzed.tokens.first.token.should == 'candl'
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: stretcher
 version: !ruby/object:Gem::Version
-  version: 1.3.2
+  version: 1.3.3
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-02-25 00:00:00.000000000 Z
+date: 2013-03-03 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: faraday
@@ -152,6 +152,7 @@ files:
 - README.md
 - Rakefile
 - lib/stretcher.rb
+- lib/stretcher/es_component.rb
 - lib/stretcher/index.rb
 - lib/stretcher/index_type.rb
 - lib/stretcher/request_error.rb