search_flip 3.5.0 → 4.0.0.beta

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: b534f2309a3433637b7e8c661d098bcbb04aa6c047655ad3c4ab944cbb860bf7
-  data.tar.gz: ec32a8f3ef761bb9d1f4dd99e1e31448b5eceb88a40643c63cada7bc4bedeaa7
+  metadata.gz: b023bc6411647eaecfcb8f5b0fe535379974a0824d49699dd74ef867a67c69eb
+  data.tar.gz: 5dc4a0439f9442811f3966e86b0c6b09d09c54450a04234c3738ae0ab7935900
 SHA512:
-  metadata.gz: ad28e9a932ad4a572bb401593e1d0c9551d2aeb7e128ffad6d1c237f8e78f7a5385a2e94e4a113ae47bcc4576c07bfc74ab24bcbc2d65e8e7f2751951b5d09da
-  data.tar.gz: 73c88dc39c4709dc716f3a45428771c08ce83ec51b278bcbd1285663e5cdafc2c66181b61ad78ee741feb2963bde607916ddaab8ffe8ec35124e695d37f287b7
+  metadata.gz: a5fc637e8c016531df22af678adcae60a4f4169e5f1d2215d52be0a0dff1111b849cb4d79ca36d6cbf315c8516bc51d543852ab5026339d59cb1f5d08ba356f7
+  data.tar.gz: 7cf9d55a9aca5d6d63401750d40bc656fcc345101d8ce92a300c24649e38810d399d43ff225611bbdd91a257d6cb15eee26a6c811974b987e92e72e67eded704

data/.github/workflows/test.yml

@@ -11,11 +11,11 @@ jobs:
         - elasticsearch:5.4
         - docker.elastic.co/elasticsearch/elasticsearch:6.7.0
         - docker.elastic.co/elasticsearch/elasticsearch:7.0.0
-        - docker.elastic.co/elasticsearch/elasticsearch:7.11.2
+        - docker.elastic.co/elasticsearch/elasticsearch:7.9.0
         ruby:
+        - 2.5
         - 2.6
         - 2.7
-        - 3.0
     services:
       elasticsearch:
         image: ${{ matrix.elasticsearch }}
@@ -28,7 +28,6 @@ jobs:
     - uses: actions/setup-ruby@v1
       with:
         ruby-version: ${{ matrix.ruby }}
-    - run: gem install bundler
     - run: bundle
     - run: sleep 10
     - run: bundle exec rspec

data/.rubocop.yml

@@ -1,14 +1,13 @@
 AllCops:
   NewCops: enable
-  TargetRubyVersion: 2.5
-  SuggestExtensions: false
-
-Layout/EmptyLineBetweenDefs:
-  EmptyLineBetweenClassDefs: false
+  TargetRubyVersion: 2.4
 
 Gemspec/RequiredRubyVersion:
   Enabled: false
 
+Style/CaseLikeIf:
+  Enabled: false
+
 Style/ExplicitBlockArgument:
   Enabled: false
 

data/CHANGELOG.md

@@ -1,29 +1,13 @@
 
 # CHANGELOG
 
-## v3.5.0
+## v4.0.0
 
-* Add `SearchFlip::Criteria#http_timeout` to allow specifying timeouts on
-  a query level
-
-## v3.4.0
-
-* Expose `Http#timeout` via `SearchFlip::HTTPClient`
-
-## v3.3.0
-
-* Update httprb
-* Changed oj default options
-* Allow to set oj json options
-
-## v3.2.1
-
-* Fix `refresh` having a empty body breaking in elasticsearch 7.11
-
-## v3.2.0
-
-* Fix `index_scope` not being passed in `each_record` without block
-* Added `SearchFlip::Criteria#match_none`
+* [BREAKING] For performance reasons, `SearchFlip::Result` now no longer
+  inherits `Hashie::Mash`
+  * It no longer supports symbol based access like `result[:id]`
+  * It no longer supports question mark methods like `result.title?`
+  * It no longer supports method based assignment like `result.some_key = "value"`
 
 ## v3.1.2
 

data/README.md

@@ -51,7 +51,7 @@ CommentIndex.search("hello world").where(available: true).sort(id: "desc").aggre
 
 ```
 
-Finally, SearchFlip comes with a minimal set of dependencies.
+Finally, SearchFlip comes with a minimal set of dependencies (http-rb and oj only).
 
 ## Reference Docs
 
@@ -418,14 +418,6 @@ Simply matches all documents:
 CommentIndex.match_all
 ```
 
-* `match_none`
-
-Simply matches none documents at all:
-
-```ruby
-CommentIndex.match_none
-```
-
 * `all`
 
 Simply returns the criteria as is or an empty criteria when called on the index
@@ -482,8 +474,8 @@ end
 ```
 
 Generally, aggregation results returned by Elasticsearch are returned as a
-`SearchFlip::Result`, which basically is a `Hashie::Mash`, such that you can
-access them via:
+`SearchFlip::Result`, which basically is a Hash with method-like access, such
+that you can access them via:
 
 ```ruby
 query.aggregations(:username)["mrkamel"].revenue.value
@@ -698,14 +690,6 @@ Specify a timeout to limit query processing time:
 CommentIndex.timeout("3s").execute
 ```
 
-* `http_timeout`
-
-Specify a http timeout for the request which will be send to Elasticsearch:
-
-```ruby
-CommentIndex.http_timeout(3).execute
-```
-
 * `terminate_after`
 
 Activate early query termination to stop query processing after the specified
@@ -764,7 +748,7 @@ end
 This allows to use different clusters per index e.g. when migrating indices to
 new versions of Elasticsearch.
 
-You can specify basic auth, additional headers, request timeouts, etc via:
+You can specify basic auth, additional headers, etc via:
 
 ```ruby
 http_client = SearchFlip::HTTPClient.new
@@ -781,9 +765,6 @@ http_client = http_client.via("proxy.host", 8080)
 # Custom headers
 http_client = http_client.headers(key: "value")
 
-# Timeouts
-http_client = http_client.timeout(20)
-
 SearchFlip::Connection.new(base_url: "...", http_client: http_client)
 ```
 
@@ -893,52 +874,57 @@ Thus, if your ORM supports `.find_each`, `#id` and `#where` you are already
 good to go. Otherwise, simply add your custom implementation of those methods
 that work with whatever ORM you use.
 
-## JSON
+## Date and Timestamps in JSON
 
-SearchFlip is using the [Oj gem](https://github.com/ohler55/oj) to generate
-JSON. More concretely, SearchFlip is using:
+Elasticsearch requires dates and timestamps to have one of the formats listed
+here: [https://www.elastic.co/guide/en/elasticsearch/reference/current/mapping-date-format.html#strict-date-time](https://www.elastic.co/guide/en/elasticsearch/reference/current/mapping-date-format.html#strict-date-time).
+
+However, `JSON.generate` in ruby by default outputs something like:
 
 ```ruby
-Oj.dump({ key: "value" }, mode: :custom, use_to_json: true, time_format: :xmlschema, bigdecimal_as_decimal: false)
+JSON.generate(time: Time.now.utc)
+# => "{\"time\":\"2018-02-22 18:19:33 UTC\"}"
 ```
 
-The `use_to_json` option is used for maximum compatibility, most importantly
-when using rails `ActiveSupport::TimeWithZone` timestamps, which `oj` can not
-serialize natively. However, `use_to_json` adds performance overhead. You can
-change the json options via:
+This format is not compatible with Elasticsearch by default. If you're on
+Rails, ActiveSupport adds its own `#to_json` methods to `Time`, `Date`, etc.
+However, ActiveSupport checks whether they are used in combination with
+`JSON.generate` or not and adapt:
 
 ```ruby
-SearchFlip::Config[:json_options] = {
-  mode: :custom,
-  use_to_json: false,
-  time_format: :xmlschema,
-  bigdecimal_as_decimal: false
-}
+Time.now.utc.to_json
+=> "\"2018-02-22T18:18:22.088Z\""
+
+JSON.generate(time: Time.now.utc)
+=> "{\"time\":\"2018-02-22 18:18:59 UTC\"}"
 ```
 
-However, you then have to convert timestamps manually for indexation via e.g.:
+SearchFlip is using the [Oj gem](https://github.com/ohler55/oj) to generate
+JSON. More concretely, SearchFlip is using:
 
 ```ruby
-class MyIndex
-  # ...
+Oj.dump({ key: "value" }, mode: :custom, use_to_json: true)
+```
 
-  def self.serialize(model)
-    {
-      # ...
+This mitigates the issues if you're on Rails:
 
-      created_at: model.created_at.to_time
-    }
-  end
-end
+```ruby
+Oj.dump(Time.now, mode: :custom, use_to_json: true)
+# => "\"2018-02-22T18:21:21.064Z\""
 ```
 
-Please check out the oj docs for more details.
+However, if you're not on Rails, you need to add `#to_json` methods to `Time`,
+`Date` and `DateTime` to get proper serialization. You can either add them on
+your own, via other libraries or by simply using:
+
+```ruby
+require "search_flip/to_json"
+```
 
 ## Feature Support
 
 * for Elasticsearch 2.x, the delete-by-query plugin is required to delete
   records via queries
-* `#match_none` is only available with Elasticsearch >= 5
 * `#track_total_hits` is only available with Elasticsearch >= 7
 
 ## Keeping your Models and Indices in Sync

data/UPDATING.md

@@ -1,6 +1,46 @@
 
 # Updating from previous SearchFlip versions
 
+## Update 3.x to 4.x
+
+**[BREAKING]** For performance reasons, `SearchFlip::Result` no longer
+inherits `Hashie::Mash`
+
+* It no longer supports symbol based access like `result[:id]`
+
+2.x:
+
+```ruby
+CommentIndex.match_all.results.first[:id]
+CommentIndex.aggregate(:tags).aggregations(:tags).values.first[:doc_count]
+```
+
+3.x
+
+```ruby
+CommentIndex.match_all.results.first["id"] # or .id
+CommentIndex.aggregate(:tags).aggregations(:tags).values.first["doc_count"] # or .doc_count
+```
+
+* It no longer supports question mark methods like `result.title?`
+
+2.x:
+
+```ruby
+CommentIndex.match_all.results.first.is_published?
+```
+
+3.x
+
+```ruby
+CommentIndex.match_all.results.first.is_published == true
+```
+
+* It no longer supports method based assignment like `result.some_key = "value"`.
+
+However, this should not have any practical implications, as changing the
+results is not considered to be a common use case.
+
 ## Update 2.x to 3.x
 
 * **[BREAKING]**  No longer pass multiple arguments to `#must`, `#must_not`,

data/lib/search_flip/bulk.rb

@@ -143,7 +143,7 @@ module SearchFlip
 
       return if options[:raise] == false
 
-      parsed_response = SearchFlip::JSON.parse(response.to_s)
+      parsed_response = response.parse
 
       return unless parsed_response["errors"]
 

data/lib/search_flip/config.rb

@@ -5,12 +5,6 @@ module SearchFlip
     bulk_limit: 1_000,
     bulk_max_mb: 100,
     auto_refresh: false,
-    instrumenter: NullInstrumenter.new,
-    json_options: {
-      mode: :custom,
-      use_to_json: true,
-      time_format: :xmlschema,
-      bigdecimal_as_decimal: false
-    }
+    instrumenter: NullInstrumenter.new
   }
 end

data/lib/search_flip/connection.rb

@@ -28,11 +28,7 @@ module SearchFlip
 
     def version
       @version_mutex.synchronize do
-        @version ||= begin
-          response = http_client.headers(accept: "application/json").get("#{base_url}/")
-
-          SearchFlip::JSON.parse(response.to_s)["version"]["number"]
-        end
+        @version ||= http_client.headers(accept: "application/json").get("#{base_url}/").parse["version"]["number"]
       end
     end
 
@@ -44,9 +40,7 @@ module SearchFlip
     # @return [Hash] The raw response
 
     def cluster_health
-      response = http_client.headers(accept: "application/json").get("#{base_url}/_cluster/health")
-
-      SearchFlip::JSON.parse(response.to_s)
+      http_client.headers(accept: "application/json").get("#{base_url}/_cluster/health").parse
     end
 
     # Uses the Elasticsearch Multi Search API to execute multiple search requests
@@ -96,11 +90,10 @@ module SearchFlip
     # @return [Hash] The raw response
 
     def update_aliases(payload)
-      response = http_client
+      http_client
         .headers(accept: "application/json", content_type: "application/json")
         .post("#{base_url}/_aliases", body: SearchFlip::JSON.generate(payload))
-
-      SearchFlip::JSON.parse(response.to_s)
+        .parse
     end
 
     # Sends an analyze request to Elasticsearch. Raises
@@ -112,11 +105,10 @@ module SearchFlip
     # @return [Hash] The raw response
 
     def analyze(request, params = {})
-      response = http_client
+      http_client
         .headers(accept: "application/json")
         .post("#{base_url}/_analyze", json: request, params: params)
-
-      SearchFlip::JSON.parse(response.to_s)
+        .parse
     end
 
     # Fetches information about the specified index aliases. Raises
@@ -132,11 +124,10 @@ module SearchFlip
     # @return [Hash] The raw response
 
     def get_aliases(index_name: "*", alias_name: "*")
-      response = http_client
+      http_client
         .headers(accept: "application/json", content_type: "application/json")
         .get("#{base_url}/#{index_name}/_alias/#{alias_name}")
-
-      SearchFlip::JSON.parse(response.to_s)
+        .parse
     end
 
     # Returns whether or not the associated Elasticsearch alias already
@@ -168,11 +159,10 @@ module SearchFlip
     # @return [Array] The raw response
 
     def get_indices(name = "*", params: {})
-      response = http_client
+      http_client
         .headers(accept: "application/json", content_type: "application/json")
         .get("#{base_url}/_cat/indices/#{name}", params: params)
-
-      SearchFlip::JSON.parse(response.to_s)
+        .parse
     end
 
     alias_method :cat_indices, :get_indices
@@ -269,11 +259,10 @@ module SearchFlip
     # @return [Hash] The index settings
 
     def get_index_settings(index_name)
-      response = http_client
+      http_client
         .headers(accept: "application/json")
         .get("#{index_url(index_name)}/_settings")
-
-      SearchFlip::JSON.parse(response.to_s)
+        .parse
     end
 
     # Sends a refresh request to Elasticsearch. Raises
@@ -283,7 +272,7 @@ module SearchFlip
     # @return [Boolean] Returns true or raises SearchFlip::ResponseError
 
     def refresh(index_names = nil)
-      http_client.post("#{index_names ? index_url(Array(index_names).join(",")) : base_url}/_refresh")
+      http_client.post("#{index_names ? index_url(Array(index_names).join(",")) : base_url}/_refresh", json: {})
 
       true
     end
@@ -321,9 +310,7 @@ module SearchFlip
       url = type_name ? type_url(index_name, type_name) : index_url(index_name)
       params = type_name && version.to_f >= 6.7 ? { include_type_name: true } : {}
 
-      response = http_client.headers(accept: "application/json").get("#{url}/_mapping", params: params)
-
-      SearchFlip::JSON.parse(response.to_s)
+      http_client.headers(accept: "application/json").get("#{url}/_mapping", params: params).parse
     end
 
     # Deletes the specified index from Elasticsearch. Raises

data/lib/search_flip/criteria.rb

@@ -26,8 +26,7 @@ module SearchFlip
 
     attr_accessor :target, :profile_value, :source_value, :suggest_values, :includes_values,
       :eager_load_values, :preload_values, :failsafe_value, :scroll_args, :terminate_after_value,
-      :timeout_value, :preference_value, :search_type_value, :routing_value, :track_total_hits_value,
-      :http_timeout_value
+      :timeout_value, :preference_value, :search_type_value, :routing_value, :track_total_hits_value
 
     # Creates a new criteria while merging the attributes (constraints,
     # settings, etc) of the current criteria with the attributes of another one
@@ -48,7 +47,7 @@ module SearchFlip
         [
           :profile_value, :failsafe_value, :terminate_after_value, :timeout_value, :offset_value,
           :limit_value, :scroll_args, :source_value, :preference_value, :search_type_value,
-          :routing_value, :track_total_hits_value, :explain_value, :http_timeout_value
+          :routing_value, :track_total_hits_value, :explain_value
         ].each do |name|
           criteria.send(:"#{name}=", other.send(name)) unless other.send(name).nil?
         end
@@ -149,22 +148,6 @@ module SearchFlip
       end
     end
 
-    # Specifies a http timeout, such that a SearchFlip::TimeoutError will be
-    # thrown when the request times out.
-    #
-    # @example
-    #   ProductIndex.http_timeout(3).search("hello world")
-    #
-    # @param value [Fixnum] The timeout value
-    #
-    # @return [SearchFlip::Criteria] A newly created extended criteria
-
-    def http_timeout(value)
-      fresh.tap do |criteria|
-        criteria.http_timeout_value = value
-      end
-    end
-
     # Specifies early query termination, such that the processing will be
     # stopped after the specified number of results has been accumulated.
     #
@@ -347,13 +330,10 @@ module SearchFlip
       dupped_request.delete(:from)
       dupped_request.delete(:size)
 
-      http_request = connection.http_client
-      http_request = http_request.timeout(http_timeout_value) if http_timeout_value
-
       if connection.version.to_i >= 5
-        http_request.post("#{target.type_url}/_delete_by_query", params: request_params.merge(params), json: dupped_request)
+        connection.http_client.post("#{target.type_url}/_delete_by_query", params: request_params.merge(params), json: dupped_request)
       else
-        http_request.delete("#{target.type_url}/_query", params: request_params.merge(params), json: dupped_request)
+        connection.http_client.delete("#{target.type_url}/_query", params: request_params.merge(params), json: dupped_request)
       end
 
       target.refresh if SearchFlip::Config[:auto_refresh]
@@ -521,8 +501,8 @@ module SearchFlip
     end
 
     # Executes the search request for the current criteria, ie sends the
-    # request to Elasticsearch and returns the response. Connection, timeout
-    # and response errors will be rescued if you specify the criteria to be
+    # request to Elasticsearch and returns the response. Connection and
+    # response errors will be rescued if you specify the criteria to be
     # #failsafe, such that an empty response is returned instead.
     #
     # @example
@@ -610,7 +590,6 @@ module SearchFlip
 
     def execute!
       http_request = connection.http_client.headers(accept: "application/json")
-      http_request = http_request.timeout(http_timeout_value) if http_timeout_value
 
       http_response =
         if scroll_args && scroll_args[:id]
@@ -630,7 +609,7 @@ module SearchFlip
         end
 
       SearchFlip::Response.new(self, SearchFlip::JSON.parse(http_response.to_s))
-    rescue SearchFlip::ConnectionError, SearchFlip::TimeoutError, SearchFlip::ResponseError => e
+    rescue SearchFlip::ConnectionError, SearchFlip::ResponseError => e
       raise e unless failsafe_value
 
       SearchFlip::Response.new(self, "took" => 0, "hits" => { "total" => 0, "hits" => [] })

data/lib/search_flip/filterable.rb

@@ -235,22 +235,6 @@ module SearchFlip
       filter(match_all: options)
     end
 
-    # Adds a match none filter to the criteria, which simply matches none
-    # documents at all. Check out the Elasticsearch docs for further details.
-    #
-    # @example Basic usage
-    #   CommentIndex.match_none
-    #
-    # @example Filter chaining
-    #   query = CommentIndex.search("...")
-    #   query = query.match_none unless current_user.admin?
-    #
-    # @return [SearchFlip::Criteria] A newly created extended criteria
-
-    def match_none
-      filter(match_none: {})
-    end
-
     # Adds an exists filter to the criteria, which selects all documents for
     # which the specified field has a non-null value.
     #

data/lib/search_flip/http_client.rb

@@ -1,28 +1,7 @@
 module SearchFlip
-  # The SearchFlip::HTTPClient class wraps the http gem and responsible for the
-  # http request/response handling, ie communicating with Elasticsearch. You
-  # only need to use it directly if you need authentication to communicate with
-  # Elasticsearch or if you want to set some custom http settings.
-  #
-  # @example
-  #   http_client = SearchFlip::HTTPClient.new
-  #
-  #   # Basic Auth
-  #   http_client = http_client.basic_auth(user: "username", pass: "password")
-  #
-  #   # Raw Auth Header
-  #   http_client = http_client.auth("Bearer VGhlIEhUVFAgR2VtLCBST0NLUw")
-  #
-  #   # Proxy Settings
-  #   http_client = http_client.via("proxy.host", 8080)
-  #
-  #   # Custom headers
-  #   http_client = http_client.headers(key: "value")
-  #
-  #   # Timeouts
-  #   http_client = http_client.timeout(20)
-  #
-  #   SearchFlip::Connection.new(base_url: "...", http_client: http_client)
+  # The SearchFlip::HTTPClient class wraps the http gem, is for internal use
+  # and responsible for the http request/response handling, ie communicating
+  # with Elasticsearch.
 
   class HTTPClient
     attr_accessor :request, :plugins
@@ -35,11 +14,11 @@ module SearchFlip
     class << self
       extend Forwardable
 
-      def_delegators :new, :headers, :via, :basic_auth, :auth, :timeout
+      def_delegators :new, :headers, :via, :basic_auth, :auth
       def_delegators :new, :get, :post, :put, :delete, :head
     end
 
-    [:headers, :via, :basic_auth, :auth, :timeout].each do |method|
+    [:headers, :via, :basic_auth, :auth].each do |method|
       define_method method do |*args|
         dup.tap do |client|
           client.request = request.send(method, *args)
@@ -66,10 +45,6 @@ module SearchFlip
       response
     rescue HTTP::ConnectionError => e
       raise SearchFlip::ConnectionError, e.message
-    rescue HTTP::TimeoutError => e
-      raise SearchFlip::TimeoutError, e.message
-    rescue HTTP::Error => e
-      raise SearchFlip::HttpError, e.message
     end
   end
 end

data/lib/search_flip/index.rb

@@ -153,7 +153,7 @@ module SearchFlip
       #   scope to be applied to the scope
 
       def each_record(scope, index_scope: false)
-        return enum_for(:each_record, scope, index_scope: index_scope) unless block_given?
+        return enum_for(:each_record, scope) unless block_given?
 
         if scope.respond_to?(:find_each)
           (index_scope ? self.index_scope(scope) : scope).find_each do |record|
@@ -247,14 +247,14 @@ module SearchFlip
         SearchFlip::Criteria.new(target: self)
       end
 
-      def_delegators :criteria, :all, :profile, :where, :where_not, :filter, :range, :match_all, :match_none,
-        :exists, :exists_not, :post_where, :post_where_not, :post_range, :post_exists, :post_exists_not,
+      def_delegators :criteria, :all, :profile, :where, :where_not, :filter, :range, :match_all, :exists,
+        :exists_not, :post_where, :post_where_not, :post_range, :post_exists, :post_exists_not,
         :post_filter, :post_must, :post_must_not, :post_should, :aggregate, :scroll, :source,
         :includes, :eager_load, :preload, :sort, :resort, :order, :reorder, :offset, :limit, :paginate,
         :page, :per, :search, :highlight, :suggest, :custom, :find_in_batches, :find_results_in_batches,
         :find_each, :find_each_result, :failsafe, :total_entries, :total_count, :timeout, :terminate_after,
         :records, :results, :must, :must_not, :should, :preference, :search_type, :routing,
-        :track_total_hits, :explain, :http_timeout
+        :track_total_hits, :explain
 
       # Override to specify the type name used within Elasticsearch. Recap,
       # this gem uses an individual index for each index class, because
@@ -487,9 +487,7 @@ module SearchFlip
       # @return [Hash] The raw response
 
       def analyze(request, params = {})
-        response = connection.http_client.headers(accept: "application/json").post("#{index_url}/_analyze", json: request, params: params)
-
-        SearchFlip::JSON.parse(response.to_s)
+        connection.http_client.headers(accept: "application/json").post("#{index_url}/_analyze", json: request, params: params).parse
       end
 
       # Sends a index refresh request to Elasticsearch. Raises

data/lib/search_flip/json.rb

@@ -1,11 +1,11 @@
 module SearchFlip
   class JSON
     def self.generate(obj)
-      Oj.dump(obj, SearchFlip::Config[:json_options])
+      Oj.dump(obj, mode: :custom, use_to_json: true)
     end
 
-    def self.parse(json)
-      ::JSON.parse(json)
+    def self.parse(str)
+      Oj.load(str)
     end
   end
 end