meilisearch 0.1.0 → 0.11.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: c31820aca925f28c1e53011587a9f6f2d9dfd210425ee3e1bce4c6e2728cb655
-  data.tar.gz: a74b7e2540093dc1f7fc65f5da4259cd2c07e17460ce2497be66899559cd0137
+  metadata.gz: 9a36df5bca77a8d8487a6cfdf12ef055db52aec77d25d0a522e4fecb455d8714
+  data.tar.gz: e6a5b012fa2d4099be7ce43e5790ee032d612d68ce7fbcf697a6928458a23e51
 SHA512:
-  metadata.gz: 9b07704bb782d34b98b68b39e6a243f2882e22bc088e1ea3d66ca89f5371efa6843aa7a9b58251d3a9d0acb68825b65e61037a70f9b1ca0cd0e7e6a614aad5a3
-  data.tar.gz: c3a1018f981315a8e55135074ac88c688e94c8fab6ce0d2544c0db587d3ba3e9a91e24771455ddf32716c283931092c751ce4d7a413b45e9a26deaf4894f153e
+  metadata.gz: 37cba34617de2798ff77d3f5bd711862a84a14cbea9dca44349897dacff37f004184e6b71fad314ddefc3ce6a1cbcb124c5c9ed85781d7ee2028385bc20d3418
+  data.tar.gz: ce410e17029be8a779f10921aef8b28200dd3ae2fc1cb3924bbb5085596a1adc70c2a5e4aed5492751dee3d6f1ab4dbbdc7186673fe1041d2b4875819ab3ab65

data/lib/meilisearch.rb

@@ -2,6 +2,7 @@
 
 require 'meilisearch/version'
 require 'meilisearch/client'
+require 'meilisearch/index'
 
 module MeiliSearch
 end

data/lib/meilisearch/client.rb

@@ -1,86 +1,87 @@
 # frozen_string_literal: true
 
-require 'httparty'
-
-require 'meilisearch/error'
-require 'meilisearch/client/keys'
-require 'meilisearch/client/stats'
-require 'meilisearch/client/health'
-require 'meilisearch/client/indexes'
-require 'meilisearch/client/documents'
-require 'meilisearch/client/prepare'
-require 'meilisearch/client/search'
+require 'meilisearch/http_request'
 
 module MeiliSearch
-  class Client
-    include HTTParty
-
-    include MeiliSearch::Client::Keys
-    include MeiliSearch::Client::Stats
-    include MeiliSearch::Client::Health
-    include MeiliSearch::Client::Indexes
-    include MeiliSearch::Client::Documents
-    include MeiliSearch::Client::Prepare
-    include MeiliSearch::Client::Search
-
-    def initialize(url, api_key = nil)
-      # api_key is is for basic api authorization
-      @headers = {}
-      @headers['X-Meili-Api-Key'] = api_key if api_key
-      @headers['Content-Type'] = 'application/json'
-      @base_url = url
+  class Client < HTTPRequest
+    ### INDEXES
+
+    def indexes
+      http_get '/indexes'
     end
 
-    def get(path = '', query = {})
-      response = self.class.get(
-        (@base_url + path),
-        query: query,
-        headers: @headers
-      )
-      validate(response)
+    def show_index(index_uid)
+      index_object(index_uid).show
     end
 
-    def post(path = '', body = nil)
-      if body.nil?
-        response = self.class.post(
-          (@base_url + path),
-          headers: @headers
-        )
-      else
-        response = self.class.post(
-          (@base_url + path),
-          body: body.to_json,
-          headers: @headers
-        )
-      end
-      validate(response)
+    # Usage:
+    # client.create_index('indexUID')
+    # client.create_index('indexUID', primaryKey: 'id')
+    def create_index(index_uid, options = {})
+      body = { uid: index_uid }.merge(options)
+      res = http_post '/indexes', body
+      index_object(res['uid'])
     end
 
-    def put(path = '', body = {})
-      response = self.class.put(
-        (@base_url + path),
-        body: body.to_json,
-        headers: @headers
-      )
-      validate(response)
+    def delete_index(index_uid)
+      index_object(index_uid).delete
     end
 
-    def delete(path = '')
-      response = self.class.delete(
-        (@base_url + path),
-        headers: @headers
-      )
-      validate(response)
+    # Usage:
+    # client.index('indexUID')
+    def index(index_uid)
+      raise IndexUidError if index_uid.nil?
+
+      index_object(index_uid)
     end
+    alias get_index index
 
-    private
+    ### KEYS
+
+    def keys
+      http_get '/keys'
+    end
+    alias get_keys keys
 
-    def validate(response)
-      unless response.success?
-        raise ClientError, "#{response.code}: #{response.message}\n#{response.body}"
-      end
+    ### HEALTH
+
+    def healthy?
+      http_get '/health'
+      true
+    rescue StandardError
+      false
+    end
+
+    def health
+      http_get '/health'
+    end
+
+    def update_health(bool)
+      http_put '/health', health: bool
+    end
+
+    ### STATS
+
+    def version
+      http_get '/version'
+    end
+
+    def sysinfo
+      http_get '/sys-info'
+    end
+
+    def pretty_sysinfo
+      http_get '/sys-info/pretty'
+    end
+
+    def stats
+      http_get '/stats'
+    end
+
+    private
 
-      response.parsed_response
+    def index_object(uid)
+      Index.new(uid, @base_url, @api_key)
     end
   end
 end

data/lib/meilisearch/error.rb

@@ -2,5 +2,44 @@
 
 module MeiliSearch
   class MeiliSearchError < StandardError; end
-  class ClientError < MeiliSearchError; end
+  class IndexUidError < MeiliSearchError; end
+  class MeiliSearchTimeoutError < MeiliSearchError
+    attr_reader :message
+
+    def initialize
+      @message = "MeiliSearchTimeoutError: update wasn't processed in the expected time"
+    end
+
+    def to_s
+      "#{@message}."
+    end
+  end
+
+  class HTTPError < MeiliSearchError
+    attr_reader :status
+    attr_reader :message
+    attr_reader :http_body
+    attr_reader :http_body_message
+    attr_reader :details
+
+    alias code         status
+    alias body         http_body
+    alias body_message http_body_message
+
+    def initialize(status, message, http_body, details = nil)
+      @status = status
+      unless http_body.nil? || http_body.empty?
+        @http_body = JSON.parse(http_body)
+        @http_body_message = @http_body['message']
+      end
+      @message = message.capitalize
+      @message = "#{@message} - #{@http_body_message.capitalize}" unless @http_body_message.nil?
+      @details = details
+    end
+
+    def to_s
+      final_message = @details.nil? ? @message : "#{@message}. #{@details}"
+      "#{@status}: #{final_message}."
+    end
+  end
 end

data/lib/meilisearch/http_request.rb

@@ -0,0 +1,74 @@
+# frozen_string_literal: true
+
+require 'httparty'
+require 'meilisearch/error'
+
+module MeiliSearch
+  class HTTPRequest
+    include HTTParty
+
+    def initialize(url, api_key = nil)
+      @base_url = url
+      @api_key = api_key
+      @headers = {
+        'Content-Type' => 'application/json',
+        'X-Meili-API-Key' => api_key
+      }.compact
+    end
+
+    def http_get(path = '', query_params = {})
+      response = self.class.get(
+        @base_url + path,
+        query: query_params,
+        headers: @headers,
+        timeout: 1
+      )
+      validate(response)
+    end
+
+    def http_post(path = '', body = nil, query_params = nil)
+      body = body.to_json unless body.nil?
+      response = self.class.post(
+        @base_url + path,
+        {
+          body: body,
+          query: query_params,
+          headers: @headers,
+          timeout: 1
+        }.compact
+      )
+      validate(response)
+    end
+
+    def http_put(path = '', body = nil, query_params = nil)
+      body = body.to_json unless body.nil?
+      response = self.class.put(
+        @base_url + path,
+        {
+          body: body,
+          query: query_params,
+          headers: @headers,
+          timeout: 1
+        }.compact
+      )
+      validate(response)
+    end
+
+    def http_delete(path = '')
+      response = self.class.delete(
+        @base_url + path,
+        headers: @headers,
+        timeout: 1
+      )
+      validate(response)
+    end
+
+    private
+
+    def validate(response)
+      raise HTTPError.new(response.code, response.message, response.body) unless response.success?
+
+      response.parsed_response
+    end
+  end
+end

data/lib/meilisearch/index.rb

@@ -0,0 +1,273 @@
+# frozen_string_literal: true
+
+require 'meilisearch/http_request'
+require 'timeout'
+
+module MeiliSearch
+  class Index < HTTPRequest
+    attr_reader :uid
+
+    def initialize(index_uid, url, api_key = nil)
+      @uid = index_uid
+      super(url, api_key)
+    end
+
+    def show
+      http_get "/indexes/#{@uid}"
+    end
+    alias show_index show
+
+    def update(body)
+      http_put "/indexes/#{@uid}", body
+    end
+    alias update_index update
+
+    def delete
+      http_delete "/indexes/#{@uid}"
+    end
+    alias delete_index delete
+
+    def primary_key
+      show['primaryKey']
+    end
+    alias get_primary_key primary_key
+
+    ### DOCUMENTS
+
+    def document(document_id)
+      encode_document = URI.encode_www_form_component(document_id)
+      http_get "/indexes/#{@uid}/documents/#{encode_document}"
+    end
+    alias get_document document
+    alias get_one_document document
+
+    def documents(options = {})
+      http_get "/indexes/#{@uid}/documents", options
+    end
+    alias get_documents documents
+
+    def add_documents(documents, primary_key = nil)
+      documents = [documents] if documents.is_a?(Hash)
+      http_post "/indexes/#{@uid}/documents", documents, { primaryKey: primary_key }.compact
+    end
+    alias replace_documents add_documents
+    alias add_or_replace_documents add_documents
+
+    def update_documents(documents, primary_key = nil)
+      documents = [documents] if documents.is_a?(Hash)
+      http_put "/indexes/#{@uid}/documents", documents, { primaryKey: primary_key }.compact
+    end
+    alias add_or_update_documents update_documents
+
+    def delete_documents(documents_ids)
+      if documents_ids.is_a?(Array)
+        http_post "/indexes/#{@uid}/documents/delete-batch", documents_ids
+      else
+        delete_document(documents_ids)
+      end
+    end
+    alias delete_multiple_documents delete_documents
+
+    def delete_document(document_id)
+      encode_document = URI.encode_www_form_component(document_id)
+      http_delete "/indexes/#{@uid}/documents/#{encode_document}"
+    end
+    alias delete_one_document delete_document
+
+    def delete_all_documents
+      http_delete "/indexes/#{@uid}/documents"
+    end
+
+    ### SEARCH
+
+    def search(query, options = {})
+      parsed_options = options.transform_keys(&:to_sym).map do |k, v|
+        if [:facetFilters, :facetsDistribution].include?(k)
+          [k, v.inspect]
+        elsif v.is_a?(Array)
+          [k, v.join(',')]
+        else
+          [k, v]
+        end
+      end.to_h
+      http_get "/indexes/#{@uid}/search", { q: query }.merge(parsed_options)
+    end
+
+    ### UPDATES
+
+    def get_update_status(update_id)
+      http_get "/indexes/#{@uid}/updates/#{update_id}"
+    end
+
+    def get_all_update_status
+      http_get "/indexes/#{@uid}/updates"
+    end
+
+    def wait_for_pending_update(update_id, timeout_in_ms = 5000, interval_in_ms = 50)
+      Timeout.timeout(timeout_in_ms.to_f / 1000) do
+        loop do
+          get_update = get_update_status(update_id)
+          return get_update if get_update['status'] != 'enqueued'
+
+          sleep interval_in_ms.to_f / 1000
+        end
+      end
+    rescue Timeout::Error
+      raise MeiliSearch::MeiliSearchTimeoutError
+    end
+
+    ### STATS
+
+    def stats
+      http_get "/indexes/#{@uid}/stats"
+    end
+
+    def number_of_documents
+      stats['numberOfDocuments']
+    end
+
+    def indexing?
+      stats['isIndexing']
+    end
+
+    def last_update
+      stats['lastUpdate']
+    end
+
+    def fields_distribution
+      stats['fieldsDistribution']
+    end
+
+    ### SETTINGS - GENERAL
+
+    def settings
+      http_get "/indexes/#{@uid}/settings"
+    end
+    alias get_settings settings
+
+    def update_settings(settings)
+      http_post "/indexes/#{@uid}/settings", settings
+    end
+
+    def reset_settings
+      http_delete "/indexes/#{@uid}/settings"
+    end
+
+    ### SETTINGS - RANKING RULES
+
+    def ranking_rules
+      http_get "/indexes/#{@uid}/settings/ranking-rules"
+    end
+    alias get_ranking_rules ranking_rules
+
+    def update_ranking_rules(ranking_rules)
+      http_post "/indexes/#{@uid}/settings/ranking-rules", ranking_rules
+    end
+
+    def reset_ranking_rules
+      http_delete "/indexes/#{@uid}/settings/ranking-rules"
+    end
+
+    ### SETTINGS - SYNONYMS
+
+    def synonyms
+      http_get "/indexes/#{@uid}/settings/synonyms"
+    end
+    alias get_synonyms synonyms
+
+    def update_synonyms(synonyms)
+      http_post "/indexes/#{@uid}/settings/synonyms", synonyms
+    end
+
+    def reset_synonyms
+      http_delete "/indexes/#{@uid}/settings/synonyms"
+    end
+
+    ### SETTINGS - STOP-WORDS
+
+    def stop_words
+      http_get "/indexes/#{@uid}/settings/stop-words"
+    end
+    alias get_stop_words stop_words
+
+    def update_stop_words(stop_words)
+      body = stop_words.is_a?(Array) ? stop_words : [stop_words]
+      http_post "/indexes/#{@uid}/settings/stop-words", body
+    end
+
+    def reset_stop_words
+      http_delete "/indexes/#{@uid}/settings/stop-words"
+    end
+
+    ### SETTINGS - DINSTINCT ATTRIBUTE
+
+    def distinct_attribute
+      http_get "/indexes/#{@uid}/settings/distinct-attribute"
+    end
+    alias get_distinct_attribute distinct_attribute
+
+    def update_distinct_attribute(distinct_attribute)
+      http_post "/indexes/#{@uid}/settings/distinct-attribute", distinct_attribute
+    end
+
+    def reset_distinct_attribute
+      http_delete "/indexes/#{@uid}/settings/distinct-attribute"
+    end
+
+    ### SETTINGS - SEARCHABLE ATTRIBUTES
+
+    def searchable_attributes
+      http_get "/indexes/#{@uid}/settings/searchable-attributes"
+    end
+    alias get_searchable_attributes searchable_attributes
+
+    def update_searchable_attributes(searchable_attributes)
+      http_post "/indexes/#{@uid}/settings/searchable-attributes", searchable_attributes
+    end
+
+    def reset_searchable_attributes
+      http_delete "/indexes/#{@uid}/settings/searchable-attributes"
+    end
+
+    ### SETTINGS - DISPLAYED ATTRIBUTES
+
+    def displayed_attributes
+      http_get "/indexes/#{@uid}/settings/displayed-attributes"
+    end
+    alias get_displayed_attributes displayed_attributes
+
+    def update_displayed_attributes(displayed_attributes)
+      http_post "/indexes/#{@uid}/settings/displayed-attributes", displayed_attributes
+    end
+
+    def reset_displayed_attributes
+      http_delete "/indexes/#{@uid}/settings/displayed-attributes"
+    end
+
+    ### SETTINGS - ACCEPT NEW FIELS VALUE
+
+    def accept_new_fields
+      http_get "/indexes/#{@uid}/settings/accept-new-fields"
+    end
+    alias get_accept_new_fields accept_new_fields
+
+    def update_accept_new_fields(accept_new_fields)
+      http_post "/indexes/#{@uid}/settings/accept-new-fields", accept_new_fields
+    end
+
+    ### SETTINGS - ATTRIBUTES FOR FACETING
+
+    def attributes_for_faceting
+      http_get "/indexes/#{@uid}/settings/attributes-for-faceting"
+    end
+    alias get_attributes_for_faceting attributes_for_faceting
+
+    def update_attributes_for_faceting(attributes_for_faceting)
+      http_post "/indexes/#{@uid}/settings/attributes-for-faceting", attributes_for_faceting
+    end
+
+    def reset_attributes_for_faceting
+      http_delete "/indexes/#{@uid}/settings/attributes-for-faceting"
+    end
+  end
+end

data/lib/meilisearch/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module MeiliSearch
-  VERSION = '0.1.0'
+  VERSION = '0.11.0'
 end

metadata

@@ -1,30 +1,36 @@
 --- !ruby/object:Gem::Specification
 name: meilisearch
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.11.0
 platform: ruby
 authors:
 - Meili
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-11-20 00:00:00.000000000 Z
+date: 2020-06-15 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: httparty
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
         version: 0.17.1
+    - - "<"
+      - !ruby/object:Gem::Version
+        version: 0.19.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
         version: 0.17.1
-description: A simple ruby client for Meilisearch API. See https://github.com/meilisearch/MeiliDB
+    - - "<"
+      - !ruby/object:Gem::Version
+        version: 0.19.0
+description: An easy-to-use ruby client for Meilisearch API. See https://github.com/meilisearch/MeiliSearch
 email: bonjour@meilisearch.com
 executables: []
 extensions: []
@@ -32,18 +38,11 @@ extra_rdoc_files: []
 files:
 - lib/meilisearch.rb
 - lib/meilisearch/client.rb
-- lib/meilisearch/client/documents.rb
-- lib/meilisearch/client/health.rb
-- lib/meilisearch/client/indexes.rb
-- lib/meilisearch/client/keys.rb
-- lib/meilisearch/client/objects.rb
-- lib/meilisearch/client/prepare.rb
-- lib/meilisearch/client/search.rb
-- lib/meilisearch/client/stats.rb
-- lib/meilisearch/dataset.rb
 - lib/meilisearch/error.rb
+- lib/meilisearch/http_request.rb
+- lib/meilisearch/index.rb
 - lib/meilisearch/version.rb
-homepage: https://github.com/meilisearch/ruby-meili-api
+homepage: https://github.com/meilisearch/meilisearch-ruby
 licenses:
 - MIT
 metadata: {}
@@ -65,5 +64,5 @@ requirements: []
 rubygems_version: 3.0.3
 signing_key: 
 specification_version: 4
-summary: A simple ruby client for Meilisearch API
+summary: An easy-to-use ruby client for Meilisearch API
 test_files: []

data/lib/meilisearch/client/documents.rb

@@ -1,41 +0,0 @@
-# frozen_string_literal: true
-
-module MeiliSearch
-  class Client
-    module Documents
-      def add_documents(index_name, documents)
-        documents.each_slice(1000).map do |slice|
-          post "/indexes/#{index_name}/documents", slice
-        end
-      end
-
-      def document(index_name, document_uid)
-        get "/indexes/#{index_name}/documents/#{document_uid}"
-      end
-
-      def get_all_documents(index_name)
-        get "/indexes/#{index_name}/documents"
-      end
-
-      def batch_documents
-        raise NotImplementedError
-      end
-
-      def update_documents
-        raise NotImplementedError
-      end
-
-      def delete_one_document(index_name, document_uid)
-        delete "/indexes/#{index_name}/documents/#{document_uid}"
-      end
-
-      def delete_multiple_documents(index_name, document_uids)
-        post "/indexes/#{index_name}/documents/delete", document_uids
-      end
-
-      def clear_all_documents(index_name)
-        delete "/indexes/#{index_name}/documents"
-      end
-    end
-  end
-end

data/lib/meilisearch/client/health.rb

@@ -1,22 +0,0 @@
-# frozen_string_literal: true
-
-module MeiliSearch
-  class Client
-    module Health
-      def is_healthy?
-        get '/health'
-        true
-      rescue StandardError
-        false
-      end
-
-      def health
-        get '/health'
-      end
-
-      def update_health(bool)
-        put '/health', health: bool
-      end
-    end
-  end
-end

data/lib/meilisearch/client/indexes.rb

@@ -1,31 +0,0 @@
-# frozen_string_literal: true
-
-module MeiliSearch
-  class Client
-    module Indexes
-      def indexes
-        get '/indexes'
-      end
-
-      def index(index_uid)
-        get "/indexes/#{index_uid}"
-      end
-
-      def create_index(index_uid, schema = nil)
-        if schema.nil?
-          post "/indexes/#{index_uid}"
-        else
-          post "/indexes/#{index_uid}", schema
-        end
-      end
-
-      def update_index(index_uid, schema = nil)
-        put "/indexes/#{index_uid}", schema
-      end
-
-      def delete_index(index_uid)
-        delete "/indexes/#{index_uid}"
-      end
-    end
-  end
-end

data/lib/meilisearch/client/keys.rb

@@ -1,27 +0,0 @@
-# frozen_string_literal: true
-
-module MeiliSearch
-  class Client
-    module Keys
-      def keys
-        get '/keys'
-      end
-
-      def key(key_hash)
-        get "/keys/#{key_hash}"
-      end
-
-      def create_key(options = {})
-        post '/keys', options
-      end
-
-      def update_key(key_hash, options = {})
-        put "/keys/#{key_hash}", options
-      end
-
-      def delete_key(key_hash)
-        delete "/keys/#{key_hash}"
-      end
-    end
-  end
-end

data/lib/meilisearch/client/objects.rb

@@ -1,37 +0,0 @@
-# frozen_string_literal: true
-
-module MeiliSearch
-  class Client
-    module Objects
-      def add_objects(index_name, objects)
-        if objects.size <= 1000
-          post "/indexes/#{index_name}/objects", objects
-        else
-          objects.each_slice(1000) do |slice|
-            post "/indexes/#{index_name}/objects", slice
-          end
-        end
-      end
-
-      def object(index_name, object_identifier)
-        get "/indexes/#{index_name}/objects/#{object_identifier}"
-      end
-
-      def browse(index_name)
-        get "/indexes/#{index_name}/objects"
-      end
-
-      def batch_objects
-        raise NotImplementedError
-      end
-
-      def update_objects
-        raise NotImplementedError
-      end
-
-      def delete_objects(index_name, object_ids)
-        delete "/indexes/#{index_name}/objects", object_ids
-      end
-    end
-  end
-end

data/lib/meilisearch/client/prepare.rb

@@ -1,11 +0,0 @@
-# frozen_string_literal: true
-
-module MeiliSearch
-  class Client
-    module Prepare
-      def rollout
-        put '/prepare/rollout'
-      end
-    end
-  end
-end

data/lib/meilisearch/client/search.rb

@@ -1,11 +0,0 @@
-# frozen_string_literal: true
-
-module MeiliSearch
-  class Client
-    module Search
-      def search(index_uid, query, options = {})
-        get "/indexes/#{index_uid}/search", { q: query }.merge(options)
-      end
-    end
-  end
-end

data/lib/meilisearch/client/stats.rb

@@ -1,39 +0,0 @@
-# frozen_string_literal: true
-
-module MeiliSearch
-  class Client
-    module Stats
-      def version
-        get '/version'
-      end
-
-      def sysinfo
-        get '/sys-info'
-      end
-
-      def stats
-        get '/stats'
-      end
-
-      def stats_index(index_uid)
-        get "/stats/#{index_uid}"
-      end
-
-      def number_of_documents_in_index(index_uid)
-        stats_index(index_uid)['numberOfDocuments']
-      end
-
-      def index_is_indexing?(index_uid)
-        stats_index(index_uid)['isIndexing']
-      end
-
-      def index_last_update(index_uid)
-        stats_index(index_uid)['lastUpdate']
-      end
-
-      def index_fields_frequency(index_uid)
-        stats_index(index_uid)['fieldsFrequency']
-      end
-    end
-  end
-end

data/lib/meilisearch/dataset.rb

@@ -1,76 +0,0 @@
-# frozen_string_literal: true
-
-module MeiliSearch
-  # This class represents the dataset we can index
-  class Dataset
-    class NotEnumerable < StandardError; end
-    class CannotDetermineHeader < StandardError; end
-    class ExtensionNotRecognized < StandardError; end
-
-    attr_accessor :data
-
-    def self.new_from_s3(s3_file)
-      filename = s3_file.filename.to_s
-      if filename.match?(/.csv$/)
-        data = CSV.parse(s3_file.download, headers: true)
-      elsif filename.match?(/.json$/)
-        data = JSON.parse(s3_file.download)
-      else
-        raise Dataset::ExtensionNotRecognized
-      end
-      new(data)
-    end
-
-    def initialize(data)
-      @data = data
-      raise Dataset::NotEnumerable unless enumerable?
-    end
-
-    def clean
-      encode_to_utf8
-      replace_nil_values
-      @data
-    end
-
-    def headers
-      if @data.class == Array
-        @data.first.keys
-      elsif @data.class == CSV::Table
-        @data.headers
-      else
-        raise CannotDetermineHeader
-      end
-    end
-
-    def schema
-      schema = headers.map do |attribute|
-        [attribute, [:indexed, :displayed]]
-      end.to_h
-
-      # Find first attribute containing id
-      identifier = headers.detect { |attribute| attribute[/id/i] }
-
-      # Then give it the :identifier attribute
-      schema[identifier].push :identifier
-      schema
-    end
-
-    private
-
-    def replace_nil_values
-      @data = @data.map do |record|
-        record.each do |key, value|
-          record[key] = '' if value.nil?
-        end
-      end
-    end
-
-    def encode_to_utf8
-      @data = @data.map(&:to_h).map(&:to_utf8)
-    end
-
-    def enumerable?
-      @data.respond_to? :each
-    end
-  end
-end