mindee 4.13.0 → 5.0.0.rc1

data/lib/mindee/v2/product/extraction/extraction_inference.rb

@@ -1,22 +1,37 @@
 # frozen_string_literal: true
 
-require_relative '../../../parsing/v2/inference'
+require_relative 'extraction_response'
 require_relative 'extraction_result'
+require_relative 'params/extraction_parameters'
 
 module Mindee
   module V2
     module Product
       module Extraction
         # Extraction inference.
-        class ExtractionInference < Mindee::Parsing::V2::Inference
-          # @return [ExtractionResult] Parsed inference payload.
+        class ExtractionInference < Mindee::V2::Parsing::BaseInference
+          # @return [InferenceActiveOptions] Options which were activated during the inference.
+          attr_reader :active_options
+          # @return [ExtractionResult] Result contents.
           attr_reader :result
 
           # @param server_response [Hash] Hash representation of the JSON returned by the service.
           def initialize(server_response)
             super
 
-            @result = Mindee::V2::Product::Extraction::ExtractionResult.new(server_response['result'])
+            @active_options = V2::Parsing::InferenceActiveOptions.new(server_response['active_options'])
+            @result = ExtractionResult.new(server_response['result'])
+          end
+
+          # String representation.
+          # @return [String]
+          def to_s
+            [
+              super,
+              @active_options.to_s,
+              @result.to_s,
+              '',
+            ].join("\n")
           end
         end
       end

data/lib/mindee/v2/product/extraction/extraction_response.rb

@@ -1,6 +1,5 @@
 # frozen_string_literal: true
 
-require_relative '../../../parsing/v2/inference_response'
 require_relative 'extraction_inference'
 
 module Mindee
@@ -8,14 +7,24 @@ module Mindee
     module Product
       module Extraction
         # HTTP response wrapper that embeds a V2 Inference.
-        class ExtractionResponse < Mindee::Parsing::V2::InferenceResponse
+        class ExtractionResponse < Mindee::V2::Parsing::BaseResponse
           # @return [ExtractionInference] Parsed inference payload.
           attr_reader :inference
 
+          @slug = 'extraction'
+          @_params_type = Params::ExtractionParameters
+
           def initialize(server_response)
             super
+
             @inference = Mindee::V2::Product::Extraction::ExtractionInference.new(server_response['inference'])
           end
+
+          # String representation.
+          # @return [String]
+          def to_s
+            @inference.to_s
+          end
         end
       end
     end

data/lib/mindee/v2/product/extraction/extraction_result.rb

@@ -1,13 +1,42 @@
 # frozen_string_literal: true
 
-require_relative '../../../parsing/v2/inference_result'
+require_relative '../../parsing'
 
 module Mindee
   module V2
     module Product
       module Extraction
         # Result of an extraction utility inference.
-        class ExtractionResult < Mindee::Parsing::V2::InferenceResult
+        class ExtractionResult
+          # @return [Mindee::V2::Parsing::Field::InferenceFields] Fields produced by the model.
+          attr_reader :fields
+          # @return [Mindee::V2::Parsing::RawText, nil] Optional extra data.
+          attr_reader :raw_text
+          # @return [Mindee::V2::Parsing::RAGMetadata, nil] Optional RAG metadata.
+          attr_reader :rag
+
+          # @param server_response [Hash] Hash version of the JSON returned by the API.
+          def initialize(server_response)
+            raise ArgumentError, 'server_response must be a Hash' unless server_response.is_a?(Hash)
+
+            @fields = Parsing::Field::InferenceFields.new(server_response['fields'])
+
+            @raw_text = server_response['raw_text'] ? Parsing::RawText.new(server_response['raw_text']) : nil
+            return unless server_response.key?('rag') && server_response['rag']
+
+            @rag = Parsing::RAGMetadata.new(server_response['rag'])
+          end
+
+          # String representation.
+          # @return [String]
+          def to_s
+            parts = [
+              'Fields',
+              '======',
+              @fields.to_s,
+            ]
+            parts.join("\n")
+          end
         end
       end
     end

data/lib/mindee/v2/product/extraction/params/data_schema.rb

@@ -0,0 +1,51 @@
+# frozen_string_literal: true
+
+require_relative 'data_schema_replace'
+
+module Mindee
+  module V2
+    module Product
+      module Extraction
+        # Extraction parameters.
+        module Params
+          # Modify the Data Schema.
+          class DataSchema
+            # @return [Mindee::V2::Product::Extraction::Params::DataSchemaReplace]
+            attr_reader :replace
+
+            # @param data_schema [Hash, String]
+            def initialize(data_schema)
+              case data_schema
+              when String
+                parsed = JSON.parse(data_schema.to_s, object_class: Hash)
+                parsed.transform_keys!(&:to_sym)
+                @replace = DataSchemaReplace.new(parsed[:replace])
+              when Hash
+                data_schema.transform_keys!(&:to_sym)
+                @replace = if data_schema[:replace].is_a?(DataSchemaReplace)
+                             data_schema[:replace]
+                           else
+                             DataSchemaReplace.new(data_schema[:replace])
+                           end
+              when DataSchema
+                @replace = data_schema.replace
+              else
+                raise TypeError, 'Invalid Data Schema provided.'
+              end
+            end
+
+            # @return [Hash]
+            def to_hash
+              { replace: @replace.to_hash }
+            end
+
+            # @return [String]
+            def to_s
+              to_hash.to_json
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/lib/mindee/v2/product/extraction/params/data_schema_field.rb

@@ -0,0 +1,69 @@
+# frozen_string_literal: true
+
+module Mindee
+  module V2
+    module Product
+      module Extraction
+        module Params
+          # Data Schema Field.
+          class DataSchemaField
+            # @return [String] Display name for the field, also impacts inference results.
+            attr_reader :title
+            # @return [String] Name of the field in the data schema.
+            attr_reader :name
+            # @return [Boolean] Whether this field can contain multiple values.
+            attr_reader :is_array
+            # @return [String] Data type of the field.
+            attr_reader :type
+            # @return [Array<String>, nil] Allowed values when type is `classification`. Leave empty for other types.
+            attr_reader :classification_values
+            # @return [Boolean, nil] Whether to remove duplicate values in the array.
+            # Only applicable if `is_array` is True.
+            attr_reader :unique_values
+            # @return [String, nil] Detailed description of what this field represents.
+            attr_reader :description
+            # @return [String, nil] Optional extraction guidelines.
+            attr_reader :guidelines
+            # @return [Array<Hash>, nil] Nested fields.
+            attr_reader :nested_fields
+
+            # @param field [Hash]
+            def initialize(field)
+              field.transform_keys!(&:to_sym)
+              @name = field[:name]
+              @title = field[:title]
+              @is_array = field[:is_array]
+              @type = field[:type]
+              @classification_values = field[:classification_values]
+              @unique_values = field[:unique_values]
+              @description = field[:description]
+              @guidelines = field[:guidelines]
+              @nested_fields = field[:nested_fields]
+            end
+
+            # @return [Hash]
+            def to_hash
+              out = {
+                name: @name,
+                title: @title,
+                is_array: @is_array,
+                type: @type,
+              } # @type var out: Hash[Symbol, untyped]
+              out[:classification_values] = @classification_values unless @classification_values.nil?
+              out[:unique_values] = @unique_values unless @unique_values.nil?
+              out[:description] = @description unless @description.nil?
+              out[:guidelines] = @guidelines unless @guidelines.nil?
+              out[:nested_fields] = @nested_fields unless @nested_fields.nil?
+              out
+            end
+
+            # @return [String]
+            def to_s
+              to_hash.to_json
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/lib/mindee/v2/product/extraction/params/data_schema_replace.rb

@@ -0,0 +1,39 @@
+# frozen_string_literal: true
+
+require_relative 'data_schema_field'
+
+module Mindee
+  module V2
+    module Product
+      module Extraction
+        module Params
+          # The structure to completely replace the data schema of the model.
+          class DataSchemaReplace
+            # @return [Array<DataSchemaField>] Subfields when type is `nested_object`. Leave empty for other types.
+            attr_reader :fields
+
+            # @param data_schema_replace [Hash]
+            def initialize(data_schema_replace)
+              data_schema_replace.transform_keys!(&:to_sym)
+              fields_list = data_schema_replace[:fields]
+              raise Mindee::Error::MindeeError, 'Invalid Data Schema provided.' if fields_list.nil?
+              raise TypeError, 'Data Schema replacement fields cannot be empty.' if fields_list.empty?
+
+              @fields = fields_list.map { |field| DataSchemaField.new(field) }
+            end
+
+            # @return [Hash]
+            def to_hash
+              { fields: @fields.map(&:to_hash) }
+            end
+
+            # @return [String]
+            def to_s
+              to_hash.to_json
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/lib/mindee/v2/product/extraction/params/extraction_parameters.rb

@@ -1,6 +1,7 @@
 # frozen_string_literal: true
 
-require_relative '../../../../input/inference_parameters'
+require_relative '../../../../input/base_parameters'
+require_relative 'data_schema'
 
 module Mindee
   module V2
@@ -8,8 +9,114 @@ module Mindee
       module Extraction
         module Params
           # Parameters accepted by the extraction v2 endpoint.
-          # Currently a placeholder for the InferenceParameters class.
-          class ExtractionParameters < Input::InferenceParameters
+          class ExtractionParameters < Mindee::Input::BaseParameters
+            # @return [Boolean, nil] Enhance extraction accuracy with Retrieval-Augmented Generation.
+            attr_reader :rag
+
+            # @return [Boolean, nil] Extract the full text content from the document as strings,
+            #   and fill the raw_text` attribute.
+            attr_reader :raw_text
+
+            # @return [Boolean, nil] Calculate bounding box polygons for all fields,
+            #   and fill their `locations` attribute.
+            attr_reader :polygon
+
+            # @return [Boolean, nil] Boost the precision and accuracy of all extractions.
+            #   Calculate confidence scores for all fields, and fill their confidence attribute.
+            attr_reader :confidence
+
+            # @return [String, nil] Additional text context used by the model during inference.
+            #   Not recommended, for specific use only.
+            attr_reader :text_context
+
+            # @return [DataSchemaField]
+            attr_reader :data_schema
+
+            # @return [String] Slug for the endpoint.
+            def self.slug
+              'extraction'
+            end
+
+            # rubocop:disable Metrics/ParameterLists
+            # @param [String] model_id ID of the model
+            # @param [Boolean, nil] rag Whether to enable RAG.
+            # @param [Boolean, nil] raw_text Whether to enable rax text.
+            # @param [Boolean, nil] polygon Whether to enable polygons.
+            # @param [Boolean, nil] confidence Whether to enable confidence scores.
+            # @param [String, nil] file_alias File alias, if applicable.
+            # @param [Array<String>, nil] webhook_ids
+            # @param [String, nil] text_context
+            # @param [Hash, nil] polling_options
+            # @param [Boolean, nil] close_file
+            # @param [DataSchemaField, String, Hash nil] data_schema
+            def initialize(
+              model_id,
+              rag: nil,
+              raw_text: nil,
+              polygon: nil,
+              confidence: nil,
+              file_alias: nil,
+              webhook_ids: nil,
+              text_context: nil,
+              polling_options: nil,
+              close_file: true,
+              data_schema: nil
+            )
+              super(
+                model_id,
+                file_alias: file_alias,
+                webhook_ids: webhook_ids,
+                polling_options: polling_options,
+                close_file: close_file
+              )
+
+              @rag = rag
+              @raw_text = raw_text
+              @polygon = polygon
+              @confidence = confidence
+              @text_context = text_context
+              @data_schema = DataSchema.new(data_schema) unless data_schema.nil?
+              # rubocop:enable Metrics/ParameterLists
+            end
+
+            # Appends inference-specific form data to the provided array.
+            # @param [Array] form_data Array of form fields
+            # @return [Array]
+            def append_form_data(form_data)
+              new_form_data = super
+
+              new_form_data.push(['rag', @rag.to_s]) unless @rag.nil?
+              new_form_data.push(['raw_text', @raw_text.to_s]) unless @raw_text.nil?
+              new_form_data.push(['polygon', @polygon.to_s]) unless @polygon.nil?
+              new_form_data.push(['confidence', @confidence.to_s]) unless @confidence.nil?
+              new_form_data.push(['text_context', @text_context]) if @text_context
+              new_form_data.push(['data_schema', @data_schema.to_s]) if @data_schema
+
+              new_form_data
+            end
+
+            # Loads a prediction from a Hash.
+            # @param [Hash] params Parameters to provide as a hash.
+            # @return [ExtractionParameters]
+            def self.from_hash(params: {})
+              rag = params.fetch(:rag, nil)
+              raw_text = params.fetch(:raw_text, nil)
+              polygon = params.fetch(:polygon, nil)
+              confidence = params.fetch(:confidence, nil)
+              base_params = load_from_hash(params: params)
+              new_params = base_params.merge(rag: rag, raw_text: raw_text, polygon: polygon, confidence: confidence)
+              model_id = new_params.fetch(:model_id)
+
+              ExtractionParameters.new(
+                model_id, rag: rag,
+                          raw_text: raw_text,
+                          polygon: polygon,
+                          confidence: confidence,
+                          file_alias: params.fetch(:file_alias, nil),
+                          webhook_ids: params.fetch(:webhook_ids, nil),
+                          close_file: params.fetch(:close_file, true)
+              )
+            end
           end
         end
       end

data/lib/mindee/v2/product/ocr/ocr.rb

@@ -6,12 +6,13 @@ require_relative 'params/ocr_parameters'
 module Mindee
   module V2
     module Product
-      module Ocr
-        # Ocr product.
-        class Ocr < BaseProduct
+      # OCR module.
+      module OCR
+        # OCR product.
+        class OCR < BaseProduct
           @slug = 'ocr'
-          @params_type = Mindee::V2::Product::Ocr::Params::OcrParameters
-          @response_type = Mindee::V2::Product::Ocr::OcrResponse
+          @params_type = Mindee::V2::Product::OCR::Params::OCRParameters
+          @response_type = Mindee::V2::Product::OCR::OCRResponse
         end
       end
     end

data/lib/mindee/v2/product/ocr/ocr_inference.rb

@@ -5,17 +5,17 @@ require_relative 'ocr_result'
 module Mindee
   module V2
     module Product
-      module Ocr
+      module OCR
         # The inference result for an OCR utility request.
-        class OcrInference < Mindee::V2::Parsing::BaseInference
-          # @return [OcrResult] Parsed inference payload.
+        class OCRInference < Mindee::V2::Parsing::BaseInference
+          # @return [OCRResult] Parsed inference payload.
           attr_reader :result
 
           # @param server_response [Hash] Hash representation of the JSON returned by the service.
           def initialize(server_response)
             super
 
-            @result = OcrResult.new(server_response['result'])
+            @result = OCRResult.new(server_response['result'])
           end
 
           # String representation.

data/lib/mindee/v2/product/ocr/ocr_page.rb

@@ -5,17 +5,17 @@ require_relative 'ocr_word'
 module Mindee
   module V2
     module Product
-      module Ocr
+      module OCR
         # OCR result for a single page.
-        class OcrPage
-          # @return [Array<OcrWord>] List of words extracted from the document page.
+        class OCRPage
+          # @return [Array<OCRWord>] List of words extracted from the document page.
           attr_reader :words
           # @return [String] Full text content extracted from the document page.
           attr_reader :content
 
           # @param server_response [Hash] Hash representation of the JSON returned by the service.
           def initialize(server_response)
-            @words = server_response['words'].map { |word| OcrWord.new(word) }
+            @words = server_response['words'].map { |word| OCRWord.new(word) }
             @content = server_response['content']
           end
 
@@ -23,7 +23,7 @@ module Mindee
           # @return [String]
           def to_s
             ocr_words = "\n"
-            ocr_words += @words.map(&:to_s).join("\n\n") if @words&.any?
+            ocr_words += @words.join("\n\n") if @words&.any?
             "OCR Words\n======#{ocr_words}\n\n:Content: #{@content}"
           end
         end

data/lib/mindee/v2/product/ocr/ocr_response.rb

@@ -7,17 +7,17 @@ require_relative 'ocr_inference'
 module Mindee
   module V2
     module Product
-      module Ocr
+      module OCR
         # HTTP response wrapper that embeds a V2 Inference.
-        class OcrResponse < Mindee::V2::Parsing::BaseResponse
-          # @return [OcrInference] Parsed inference payload.
+        class OCRResponse < Mindee::V2::Parsing::BaseResponse
+          # @return [OCRInference] Parsed inference payload.
           attr_reader :inference
 
           # @param server_response [Hash] Hash parsed from the API JSON response.
           def initialize(server_response)
             super
 
-            @inference = OcrInference.new(server_response['inference'])
+            @inference = OCRInference.new(server_response['inference'])
           end
 
           # String representation.

data/lib/mindee/v2/product/ocr/ocr_result.rb

@@ -5,17 +5,17 @@ require_relative 'ocr_page'
 module Mindee
   module V2
     module Product
-      module Ocr
+      module OCR
         # Result of a ocr utility inference.
-        class OcrResult
-          # @return [Array<OcrPage>] List of OCR results for each page in the document.
+        class OCRResult
+          # @return [Array<OCRPage>] List of OCR results for each page in the document.
           attr_reader :pages
 
           # @param server_response [Hash] Hash representation of the JSON returned by the service.
           def initialize(server_response)
             @pages = if server_response.key?('pages')
                        server_response['pages'].map do |pages|
-                         OcrPage.new(pages)
+                         OCRPage.new(pages)
                        end
                      end
           end

data/lib/mindee/v2/product/ocr/ocr_word.rb

@@ -3,9 +3,9 @@
 module Mindee
   module V2
     module Product
-      module Ocr
+      module OCR
         # OCR result for a single word extracted from the document page.
-        class OcrWord
+        class OCRWord
           # @return [String] Text content of the word.
           attr_reader :content
           # @return [Mindee::Geometry::Polygon] Position information as a list of points in clockwise order.

data/lib/mindee/v2/product/ocr/params/ocr_parameters.rb

@@ -3,10 +3,11 @@
 module Mindee
   module V2
     module Product
-      module Ocr
+      module OCR
+        # OCR Parameters.
         module Params
           # Parameters accepted by the ocr utility v2 endpoint.
-          class OcrParameters < Mindee::Input::BaseParameters
+          class OCRParameters < Mindee::Input::BaseParameters
             # @return [String] Slug for the endpoint.
             def self.slug
               'ocr'
@@ -29,9 +30,9 @@ module Mindee
 
             # Loads the parameters from a Hash.
             # @param [Hash] params Parameters to provide as a hash.
-            # @return [OcrParameters]
+            # @return [OCRParameters]
             def self.from_hash(params: {})
-              OcrParameters.new(
+              OCRParameters.new(
                 params.fetch(:model_id),
                 file_alias: params.fetch(:file_alias, nil),
                 webhook_ids: params.fetch(:webhook_ids, nil),

data/lib/mindee/v2/product/split/params/split_parameters.rb

@@ -3,7 +3,9 @@
 module Mindee
   module V2
     module Product
+      # Split module.
       module Split
+        # Split parameters.
         module Params
           # Parameters accepted by the split utility v2 endpoint.
           class SplitParameters < Mindee::Input::BaseParameters

data/lib/mindee/v2/product/split/split_range.rb

@@ -23,6 +23,14 @@ module Mindee
           def to_s
             "* :Page Range: #{@page_range}\n  :Document Type: #{@document_type}"
           end
+
+          # Apply the split range inference to a file and return a single extracted PDF.
+          #
+          # @param input_source [Mindee::Input::Source::LocalInputSource] Local file to extract from
+          # @return [PDF::ExtractedPDF]
+          def extract_from_file(input_source)
+            FileOperation::Split.extract_single_split(input_source, @page_range)
+          end
         end
       end
     end

data/lib/mindee/v2/product/split/split_response.rb

@@ -25,6 +25,14 @@ module Mindee
           def to_s
             @inference.to_s
           end
+
+          # Splits the input PDF.
+          # @param input_source [Mindee::Input::Source::LocalInputSource] Path to the file or a File object.
+          # @return [FileOperation::SplitFiles]
+          def extract_from_file(input_source)
+            splits = @inference.result.splits.map(&:page_range)
+            FileOperation::Split.extract_splits(input_source, splits)
+          end
         end
       end
     end

data/lib/mindee/v2.rb

@@ -1,4 +1,7 @@
 # frozen_string_literal: true
 
-require_relative 'parsing'
-require_relative 'product'
+require_relative 'v2/client'
+require_relative 'v2/http'
+require_relative 'v2/file_operations' if Mindee::Dependencies.all_deps_available?
+require_relative 'v2/parsing'
+require_relative 'v2/product'

data/lib/mindee/version.rb

@@ -3,7 +3,7 @@
 # Mindee
 module Mindee
   # Current version.
-  VERSION = '4.13.0'
+  VERSION = '5.0.0.rc1'
 
   # Finds and return the current platform.
   # @return [Symbol, Hash[String | Symbol, Regexp], Nil?]