mindee 3.11.0 → 3.12.0

data/lib/mindee/image_extraction/common/image_extractor.rb

@@ -0,0 +1,191 @@
+# frozen_string_literal: true
+
+require 'mini_magick'
+require 'origami'
+require 'stringio'
+require 'tempfile'
+require_relative '../../input/sources'
+require_relative 'extracted_image'
+
+module Mindee
+  # Image Extraction Module.
+  module ImageExtraction
+    def attach_image_as_new_file(input_buffer)
+      # Attaches an image as a new page in a PdfDocument object.
+      #
+      # @param [StringIO] input_buffer Input buffer. Only supports JPEG.
+      # @return [Origami::PDF] A PdfDocument handle.
+
+      magick_image = MiniMagick::Image.read(input_buffer)
+      # NOTE: some jpeg images get rendered as three different versions of themselves per output if the format isn't
+      # converted.
+      magick_image.format('jpg')
+      original_density = magick_image.resolution
+      scale_factor = original_density[0].to_f / 4.166666 # No clue why bit the resolution needs to be reduced for
+      # the pdf otherwise the resulting image shrinks.
+      magick_image.format('pdf', 0, { density: scale_factor.to_s })
+      io_buffer = StringIO.new
+      magick_image.write(io_buffer)
+      Origami::PDF.read(io_buffer)
+    end
+
+    # Extracts multiple images from a given local input source.
+    #
+    # @param [Mindee::Input::Source::LocalInputSource] input_source
+    # @param [Integer] page_id ID of the Page to extract from.
+    # @param [Array<Array<Mindee::Geometry::Point>>, Array<Mindee::Geometry::Quadrangle>] polygons List of coordinates
+    # to extract.
+    # @return [Array<Mindee::ImageExtraction::ExtractedImage>] Extracted Images.
+    def extract_multiple_images_from_source(input_source, page_id, polygons)
+      new_stream = load_doc(input_source, page_id)
+      new_stream.seek(0)
+
+      extract_images_from_polygons(input_source, new_stream, page_id, polygons)
+    end
+
+    # Retrieves a PDF document's page.
+    #
+    # @param [Origami::PDF] pdf_doc Origami PDF handle.
+    # @param [Integer] page_id Page ID.
+    def get_page(pdf_doc, page_id)
+      stream = StringIO.new
+      pdf_doc.save(stream)
+
+      options = {
+        page_indexes: [page_id - 1],
+      }
+
+      Mindee::PDF::PdfProcessor.parse(stream, options)
+    end
+
+    # Extracts images from their positions on a file (as polygons).
+    #
+    # @param [Mindee::Input::Source::LocalInputSource] input_source Local input source.
+    # @param [StringIO] pdf_stream Buffer of the PDF.
+    # @param [Integer] page_id Page ID.
+    # @param [Array<Mindee::Geometry::Point, Mindee::Geometry::Polygon, Mindee::Geometry::Quadrangle>] polygons
+    # @return [Array<Mindee::ImageExtraction::ExtractedImage>] Extracted Images.
+    def extract_images_from_polygons(input_source, pdf_stream, page_id, polygons)
+      extracted_elements = []
+
+      polygons.each_with_index do |polygon, element_id|
+        polygon = normalize_polygon(polygon)
+        page_content = read_page_content(pdf_stream)
+
+        min_max_x = Geometry.get_min_max_x([
+                                             polygon.top_left,
+                                             polygon.bottom_right,
+                                             polygon.top_right,
+                                             polygon.bottom_left,
+                                           ])
+        min_max_y = Geometry.get_min_max_y([
+                                             polygon.top_left,
+                                             polygon.bottom_right,
+                                             polygon.top_right,
+                                             polygon.bottom_left,
+                                           ])
+        file_extension = determine_file_extension(input_source)
+        cropped_image = crop_image(page_content, min_max_x, min_max_y)
+        if file_extension == 'pdf'
+          cropped_image.format('jpg')
+        else
+          cropped_image.format(file_extension)
+        end
+
+        buffer = StringIO.new
+        write_image_to_buffer(cropped_image, buffer)
+        file_name = "#{input_source.filename}_page#{page_id}-#{element_id}.#{file_extension}"
+
+        extracted_elements << create_extracted_image(buffer, file_name, page_id, element_id)
+      end
+
+      extracted_elements
+    end
+
+    # Retrieves the bounding box of a polygon.
+    #
+    # @param [Array<Point>, Mindee::Geometry::Polygon] polygon
+    def normalize_polygon(polygon)
+      if polygon.is_a?(Mindee::Geometry::Polygon)
+        Mindee::Geometry.get_bounding_box(polygon)
+      else
+        polygon
+      end
+    end
+
+    # Loads a buffer into a MiniMagick Image.
+    #
+    # @param [StringIO] pdf_stream Buffer containg the PDF
+    # @return [MiniMagick::Image] a valid MiniMagick image handle.
+    def read_page_content(pdf_stream)
+      pdf_stream.rewind
+      MiniMagick::Image.read(pdf_stream)
+    end
+
+    # Crops a MiniMagick Image from a the given bounding box.
+    #
+    # @param [MiniMagick::Image] image Input Image.
+    # @param [Mindee::Geometry::MinMax] min_max_x minimum & maximum values for the x coordinates.
+    # @param [Mindee::Geometry::MinMax] min_max_y minimum & maximum values for the y coordinates.
+    def crop_image(image, min_max_x, min_max_y)
+      width = image[:width].to_i
+      height = image[:height].to_i
+
+      image.format('jpg')
+      new_width = (min_max_x.max - min_max_x.min) * width
+      new_height = (min_max_y.max - min_max_y.min) * height
+      image.crop("#{new_width}x#{new_height}+#{min_max_x.min * width}+#{min_max_y.min * height}")
+
+      image
+    end
+
+    # Writes a MiniMagick::Image to a buffer.
+    #
+    # @param [MiniMagick::Image] image a valid MiniMagick image.
+    # @param [StringIO] buffer
+    def write_image_to_buffer(image, buffer)
+      image.write(buffer)
+    end
+
+    # Retrieves the file extension from the main file to apply it to the extracted images. Note: coerces pdf as jpg.
+    #
+    # @param [Mindee::Input::Source::LocalInputSource] input_source Local input source.
+    # @return [String] A valid file extension.
+    def determine_file_extension(input_source)
+      if input_source.pdf? || input_source.filename.downcase.end_with?('pdf')
+        'jpg'
+      else
+        File.extname(input_source.filename).strip.downcase[1..]
+      end
+    end
+
+    # Generates an ExtractedImage.
+    #
+    # @param [StringIO] buffer Buffer containing the image.
+    # @param [String] file_name Name for the file.
+    # @param [Object] page_id ID of the page the file was generated from.
+    # @param [Object] element_id ID of the element of a given page.
+    def create_extracted_image(buffer, file_name, page_id, element_id)
+      buffer.rewind
+      ExtractedImage.new(
+        Mindee::Input::Source::BytesInputSource.new(buffer.read, file_name),
+        page_id,
+        element_id
+      )
+    end
+
+    # Loads a single_page from an image file or a pdf document.
+    #
+    # @param input_file [LocalInputSource] Local input.
+    # @param [Integer] page_id Page ID.
+    # @return [MiniMagick::Image] A valid PdfDocument handle.
+    def load_doc(input_file, page_id)
+      input_file.io_stream.rewind
+      if input_file.pdf?
+        get_page(Origami::PDF.read(input_file.io_stream), page_id)
+      else
+        input_file.io_stream
+      end
+    end
+  end
+end

data/lib/mindee/image_extraction/common.rb

@@ -0,0 +1,3 @@
+# frozen_string_literal: true
+
+require_relative 'common/image_extractor'

data/lib/mindee/image_extraction/multi_receipts_extractor/multi_receipts_extractor.rb

@@ -0,0 +1,26 @@
+# frozen_string_literal: true
+
+module Mindee
+  # Image Extraction Module.
+  module ImageExtraction
+    def extract_receipts(input_source, inference)
+      # Extracts individual receipts from multi-receipts documents.
+      #
+      # @param input_source [LocalInputSource] Local Input Source to extract sub-receipts from.
+      # @param inference [Inference] Results of the inference.
+      # @return [Array<ExtractedImage>] Individual extracted receipts as an array of ExtractedMultiReceiptsImage.
+
+      images = []
+      raise 'No possible receipts candidates found for MultiReceipts extraction.' unless inference.prediction.receipts
+
+      (0...input_source.count_pdf_pages).each do |page_id|
+        receipt_positions = inference.pages[page_id].prediction.receipts.map(&:bounding_box)
+        images.concat(
+          extract_multiple_images_from_source(input_source, page_id + 1, receipt_positions)
+        )
+      end
+
+      images
+    end
+  end
+end

data/lib/mindee/image_extraction/multi_receipts_extractor.rb

@@ -0,0 +1,3 @@
+# frozen_string_literal: true
+
+require_relative 'multi_receipts_extractor/multi_receipts_extractor'

data/lib/mindee/image_extraction.rb

@@ -0,0 +1,4 @@
+# frozen_string_literal: true
+
+require_relative 'image_extraction/common'
+require_relative 'image_extraction/multi_receipts_extractor'

data/lib/mindee/input/sources.rb

@@ -118,6 +118,14 @@ module Mindee
           @io_stream.close if close
           ['document', data, { filename: Mindee::Input::Source.convert_to_unicode_escape(@filename) }]
         end
+
+        def count_pdf_pages
+          return 1 unless pdf?
+
+          @io_stream.seek(0)
+          pdf_processor = Mindee::PDF::PdfProcessor.open_pdf(@io_stream)
+          pdf_processor.pages.size
+        end
       end
 
       # Load a document from a path.

data/lib/mindee/parsing/standard/company_registration_field.rb

@@ -15,6 +15,23 @@ module Mindee
           super
           @type = prediction['type']
         end
+
+        def to_table_line
+          printable = printable_values
+          format('| %<type>-15s | %<value>-20s ', type: printable['type'], value: printable['value'])
+        end
+
+        def to_s
+          printable = printable_values
+          format('Type: %<type>s, Value: %<value>s', type: printable['type'], value: printable['value'])
+        end
+
+        def printable_values
+          printable = {}
+          printable['type'] = type
+          printable['value'] = value
+          printable
+        end
       end
     end
   end

data/lib/mindee/product/financial_document/financial_document_v1_document.rb

@@ -6,7 +6,7 @@ require_relative 'financial_document_v1_line_item'
 module Mindee
   module Product
     module FinancialDocument
-      # Financial Document API version 1.7 document data.
+      # Financial Document API version 1.9 document data.
       class FinancialDocumentV1Document < Mindee::Parsing::Common::Prediction
         include Mindee::Parsing::Standard
         # The customer's address used for billing.
@@ -206,6 +206,7 @@ module Mindee
           out_str << "+#{char * 12}"
           out_str << "+#{char * 14}"
           out_str << "+#{char * 14}"
+          out_str << "+#{char * 17}"
           out_str << "+#{char * 12}"
           out_str << '+'
           out_str
@@ -225,6 +226,7 @@ module Mindee
           out_str << ' Tax Amount |'
           out_str << ' Tax Rate (%) |'
           out_str << ' Total Amount |'
+          out_str << ' Unit of measure |'
           out_str << ' Unit Price |'
           out_str << "\n#{line_items_separator('=')}"
           out_str << "\n  #{line_items}"

data/lib/mindee/product/financial_document/financial_document_v1_line_item.rb

@@ -26,6 +26,9 @@ module Mindee
         # The item total amount.
         # @return [Float]
         attr_reader :total_amount
+        # The item unit of measure.
+        # @return [String]
+        attr_reader :unit_measure
         # The item unit price.
         # @return [Float]
         attr_reader :unit_price
@@ -40,6 +43,7 @@ module Mindee
           @tax_amount = prediction['tax_amount']
           @tax_rate = prediction['tax_rate']
           @total_amount = prediction['total_amount']
+          @unit_measure = prediction['unit_measure']
           @unit_price = prediction['unit_price']
           @page_id = page_id
         end
@@ -53,6 +57,7 @@ module Mindee
           printable[:tax_amount] = @tax_amount.nil? ? '' : Field.float_to_string(@tax_amount)
           printable[:tax_rate] = @tax_rate.nil? ? '' : Field.float_to_string(@tax_rate)
           printable[:total_amount] = @total_amount.nil? ? '' : Field.float_to_string(@total_amount)
+          printable[:unit_measure] = format_for_display(@unit_measure, nil)
           printable[:unit_price] = @unit_price.nil? ? '' : Field.float_to_string(@unit_price)
           printable
         end
@@ -67,6 +72,7 @@ module Mindee
           out_str << format('| %- 11s', printable[:tax_amount])
           out_str << format('| %- 13s', printable[:tax_rate])
           out_str << format('| %- 13s', printable[:total_amount])
+          out_str << format('| %- 16s', printable[:unit_measure])
           out_str << format('| %- 11s', printable[:unit_price])
           out_str << '|'
         end
@@ -81,6 +87,7 @@ module Mindee
           out_str << "\n  :Tax Amount: #{printable[:tax_amount]}"
           out_str << "\n  :Tax Rate (%): #{printable[:tax_rate]}"
           out_str << "\n  :Total Amount: #{printable[:total_amount]}"
+          out_str << "\n  :Unit of measure: #{printable[:unit_measure]}"
           out_str << "\n  :Unit Price: #{printable[:unit_price]}"
           out_str
         end

data/lib/mindee/product/financial_document/financial_document_v1_page.rb

@@ -6,7 +6,7 @@ require_relative 'financial_document_v1_document'
 module Mindee
   module Product
     module FinancialDocument
-      # Financial Document API version 1.7 page data.
+      # Financial Document API version 1.9 page data.
       class FinancialDocumentV1Page < Mindee::Parsing::Common::Page
         # @param prediction [Hash]
         def initialize(prediction)

data/lib/mindee/product/international_id/international_id_v2_document.rb

@@ -5,7 +5,7 @@ require_relative '../../parsing'
 module Mindee
   module Product
     module InternationalId
-      # International ID API version 2.0 document data.
+      # International ID API version 2.1 document data.
       class InternationalIdV2Document < Mindee::Parsing::Common::Prediction
         include Mindee::Parsing::Standard
         # The physical address of the document holder.

data/lib/mindee/product/international_id/international_id_v2_page.rb

@@ -6,7 +6,7 @@ require_relative 'international_id_v2_document'
 module Mindee
   module Product
     module InternationalId
-      # International ID API version 2.0 page data.
+      # International ID API version 2.1 page data.
       class InternationalIdV2Page < Mindee::Parsing::Common::Page
         # @param prediction [Hash]
         def initialize(prediction)

data/lib/mindee/product/invoice/invoice_v4_document.rb

@@ -6,7 +6,7 @@ require_relative 'invoice_v4_line_item'
 module Mindee
   module Product
     module Invoice
-      # Invoice API version 4.6 document data.
+      # Invoice API version 4.7 document data.
       class InvoiceV4Document < Mindee::Parsing::Common::Prediction
         include Mindee::Parsing::Standard
         # The customer's address used for billing.
@@ -176,6 +176,7 @@ module Mindee
           out_str << "+#{char * 12}"
           out_str << "+#{char * 14}"
           out_str << "+#{char * 14}"
+          out_str << "+#{char * 17}"
           out_str << "+#{char * 12}"
           out_str << '+'
           out_str
@@ -195,6 +196,7 @@ module Mindee
           out_str << ' Tax Amount |'
           out_str << ' Tax Rate (%) |'
           out_str << ' Total Amount |'
+          out_str << ' Unit of measure |'
           out_str << ' Unit Price |'
           out_str << "\n#{line_items_separator('=')}"
           out_str << "\n  #{line_items}"

data/lib/mindee/product/invoice/invoice_v4_line_item.rb

@@ -26,6 +26,9 @@ module Mindee
         # The item total amount.
         # @return [Float]
         attr_reader :total_amount
+        # The item unit of measure.
+        # @return [String]
+        attr_reader :unit_measure
         # The item unit price.
         # @return [Float]
         attr_reader :unit_price
@@ -40,6 +43,7 @@ module Mindee
           @tax_amount = prediction['tax_amount']
           @tax_rate = prediction['tax_rate']
           @total_amount = prediction['total_amount']
+          @unit_measure = prediction['unit_measure']
           @unit_price = prediction['unit_price']
           @page_id = page_id
         end
@@ -53,6 +57,7 @@ module Mindee
           printable[:tax_amount] = @tax_amount.nil? ? '' : Field.float_to_string(@tax_amount)
           printable[:tax_rate] = @tax_rate.nil? ? '' : Field.float_to_string(@tax_rate)
           printable[:total_amount] = @total_amount.nil? ? '' : Field.float_to_string(@total_amount)
+          printable[:unit_measure] = format_for_display(@unit_measure, nil)
           printable[:unit_price] = @unit_price.nil? ? '' : Field.float_to_string(@unit_price)
           printable
         end
@@ -67,6 +72,7 @@ module Mindee
           out_str << format('| %- 11s', printable[:tax_amount])
           out_str << format('| %- 13s', printable[:tax_rate])
           out_str << format('| %- 13s', printable[:total_amount])
+          out_str << format('| %- 16s', printable[:unit_measure])
           out_str << format('| %- 11s', printable[:unit_price])
           out_str << '|'
         end
@@ -81,6 +87,7 @@ module Mindee
           out_str << "\n  :Tax Amount: #{printable[:tax_amount]}"
           out_str << "\n  :Tax Rate (%): #{printable[:tax_rate]}"
           out_str << "\n  :Total Amount: #{printable[:total_amount]}"
+          out_str << "\n  :Unit of measure: #{printable[:unit_measure]}"
           out_str << "\n  :Unit Price: #{printable[:unit_price]}"
           out_str
         end

data/lib/mindee/product/invoice/invoice_v4_page.rb

@@ -6,7 +6,7 @@ require_relative 'invoice_v4_document'
 module Mindee
   module Product
     module Invoice
-      # Invoice API version 4.6 page data.
+      # Invoice API version 4.7 page data.
       class InvoiceV4Page < Mindee::Parsing::Common::Page
         # @param prediction [Hash]
         def initialize(prediction)

data/lib/mindee/product/multi_receipts_detector/multi_receipts_detector_v1_document.rb

@@ -5,7 +5,7 @@ require_relative '../../parsing'
 module Mindee
   module Product
     module MultiReceiptsDetector
-      # Multi Receipts Detector API version 1.0 document data.
+      # Multi Receipts Detector API version 1.1 document data.
       class MultiReceiptsDetectorV1Document < Mindee::Parsing::Common::Prediction
         include Mindee::Parsing::Standard
         # Positions of the receipts on the document.

data/lib/mindee/product/multi_receipts_detector/multi_receipts_detector_v1_page.rb

@@ -6,7 +6,7 @@ require_relative 'multi_receipts_detector_v1_document'
 module Mindee
   module Product
     module MultiReceiptsDetector
-      # Multi Receipts Detector API version 1.0 page data.
+      # Multi Receipts Detector API version 1.1 page data.
       class MultiReceiptsDetectorV1Page < Mindee::Parsing::Common::Page
         # @param prediction [Hash]
         def initialize(prediction)

data/lib/mindee/product/receipt/receipt_v5_document.rb

@@ -6,7 +6,7 @@ require_relative 'receipt_v5_line_item'
 module Mindee
   module Product
     module Receipt
-      # Receipt API version 5.2 document data.
+      # Receipt API version 5.3 document data.
       class ReceiptV5Document < Mindee::Parsing::Common::Prediction
         include Mindee::Parsing::Standard
         # The purchase category among predefined classes.

data/lib/mindee/product/receipt/receipt_v5_page.rb

@@ -6,7 +6,7 @@ require_relative 'receipt_v5_document'
 module Mindee
   module Product
     module Receipt
-      # Receipt API version 5.2 page data.
+      # Receipt API version 5.3 page data.
       class ReceiptV5Page < Mindee::Parsing::Common::Page
         # @param prediction [Hash]
         def initialize(prediction)

data/lib/mindee/product/us/healthcare_card/healthcare_card_v1.rb

@@ -0,0 +1,41 @@
+# frozen_string_literal: true
+
+require_relative '../../../parsing'
+require_relative 'healthcare_card_v1_document'
+require_relative 'healthcare_card_v1_page'
+
+module Mindee
+  module Product
+    module US
+      # Healthcare Card module.
+      module HealthcareCard
+        # Healthcare Card API version 1 inference prediction.
+        class HealthcareCardV1 < Mindee::Parsing::Common::Inference
+          @endpoint_name = 'us_healthcare_cards'
+          @endpoint_version = '1'
+
+          # @param prediction [Hash]
+          def initialize(prediction)
+            super
+            @prediction = HealthcareCardV1Document.new(prediction['prediction'], nil)
+            @pages = []
+            prediction['pages'].each do |page|
+              if page.key?('prediction') && !page['prediction'].nil? && !page['prediction'].empty?
+                @pages.push(HealthcareCardV1Page.new(page))
+              end
+            end
+          end
+
+          class << self
+            # Name of the endpoint for this product.
+            # @return [String]
+            attr_reader :endpoint_name
+            # Version for this product.
+            # @return [String]
+            attr_reader :endpoint_version
+          end
+        end
+      end
+    end
+  end
+end

data/lib/mindee/product/us/healthcare_card/healthcare_card_v1_copay.rb

@@ -0,0 +1,57 @@
+# frozen_string_literal: true
+
+require_relative '../../../parsing'
+
+module Mindee
+  module Product
+    module US
+      module HealthcareCard
+        # Is a fixed amount for a covered service.
+        class HealthcareCardV1Copay < Mindee::Parsing::Standard::FeatureField
+          include Mindee::Parsing::Standard
+          # The price of service.
+          # @return [Float]
+          attr_reader :service_fees
+          # The name of service of the copay.
+          # @return [String]
+          attr_reader :service_name
+
+          # @param prediction [Hash]
+          # @param page_id [Integer, nil]
+          def initialize(prediction, page_id)
+            super(prediction, page_id)
+            @service_fees = prediction['service_fees']
+            @service_name = prediction['service_name']
+            @page_id = page_id
+          end
+
+          # @return [Hash]
+          def printable_values
+            printable = {}
+            printable[:service_fees] = @service_fees.nil? ? '' : Field.float_to_string(@service_fees)
+            printable[:service_name] = format_for_display(@service_name, nil)
+            printable
+          end
+
+          # @return [String]
+          def to_table_line
+            printable = printable_values
+            out_str = String.new
+            out_str << format('| %- 13s', printable[:service_fees])
+            out_str << format('| %- 13s', printable[:service_name])
+            out_str << '|'
+          end
+
+          # @return [String]
+          def to_s
+            printable = printable_values
+            out_str = String.new
+            out_str << "\n  :Service Fees: #{printable[:service_fees]}"
+            out_str << "\n  :Service Name: #{printable[:service_name]}"
+            out_str
+          end
+        end
+      end
+    end
+  end
+end