mindee 1.2.0 → 2.0.0

data/lib/mindee/parsing/prediction/passport/passport_v1.rb

@@ -0,0 +1,184 @@
+# frozen_string_literal: true
+
+require 'mrz'
+
+require_relative '../common_fields'
+require_relative '../base'
+
+# We need to do this disgusting thing to avoid the following error message:
+# td3 line one does not match the required format (MRZ::InvalidFormatError)
+#
+# See:
+# https://github.com/streetspotr/mrz/issues/2
+# https://github.com/streetspotr/mrz/pull/3
+#
+MRZ::TD3Parser::FORMAT_ONE = %r{\A(.{2})(.{3})([^<]+)<(.*)\z}.freeze
+
+module Mindee
+  module Prediction
+    # Passport document.
+    class PassportV1 < Prediction
+      # The country of issue.
+      # @return [Mindee::TextField]
+      attr_reader :country
+      # The passport number.
+      # @return [Mindee::TextField]
+      attr_reader :id_number
+      # The expiration date of the passport.
+      # @return [Mindee::DateField]
+      attr_reader :expiry_date
+      # The issuance date of the passport.
+      # @return [Mindee::DateField]
+      attr_reader :issuance_date
+      # The surname (last name) of the passport holder.
+      # @return [Mindee::TextField]
+      attr_reader :surname
+      # List of first (given) names of the passport holder.
+      # @return [Mindee::TextField]
+      attr_reader :given_names
+      # The full name of the passport holder.
+      # @return [Array<Mindee::TextField>]
+      attr_reader :full_name
+      # The date of birth of the passport holder.
+      # @return [Mindee::DateField]
+      attr_reader :birth_date
+      # The place of birth of the passport holder.
+      # @return [Mindee::TextField]
+      attr_reader :birth_place
+      # The sex or gender of the passport holder.
+      # @return [Mindee::TextField]
+      attr_reader :gender
+      # The value of the first MRZ line.
+      # @return [Mindee::TextField]
+      attr_reader :mrz1
+      # The value of the second MRZ line.
+      # @return [Mindee::TextField]
+      attr_reader :mrz2
+      # All the MRZ values combined.
+      # @return [Mindee::TextField]
+      attr_reader :mrz
+
+      # @param prediction [Hash]
+      # @param page_id [Integer, nil]
+      def initialize(prediction, page_id)
+        super
+        @country = TextField.new(prediction['country'], page_id)
+        @id_number = TextField.new(prediction['id_number'], page_id)
+        @birth_date = DateField.new(prediction['birth_date'], page_id)
+        @expiry_date = DateField.new(prediction['expiry_date'], page_id)
+        @issuance_date = DateField.new(prediction['issuance_date'], page_id)
+        @birth_place = TextField.new(prediction['birth_place'], page_id)
+        @gender = TextField.new(prediction['gender'], page_id)
+        @surname = TextField.new(prediction['surname'], page_id)
+        @mrz1 = TextField.new(prediction['mrz1'], page_id)
+        @mrz2 = TextField.new(prediction['mrz2'], page_id)
+        @given_names = []
+        prediction['given_names'].each do |item|
+          @given_names.push(TextField.new(item, page_id))
+        end
+        @full_name = construct_full_name(page_id)
+        @mrz = construct_mrz(page_id)
+        check_mrz
+      end
+
+      def to_s
+        given_names = @given_names.join(' ')
+        out_str = String.new
+        out_str << "\n:Full name: #{@full_name}".rstrip
+        out_str << "\n:Given names: #{given_names}".rstrip
+        out_str << "\n:Surname: #{@surname}".rstrip
+        out_str << "\n:Country: #{@country}".rstrip
+        out_str << "\n:ID Number: #{@id_number}".rstrip
+        out_str << "\n:Issuance date: #{@issuance_date}".rstrip
+        out_str << "\n:Birth date: #{@birth_date}".rstrip
+        out_str << "\n:Expiry date: #{@expiry_date}".rstrip
+        out_str << "\n:MRZ 1: #{@mrz1}".rstrip
+        out_str << "\n:MRZ 2: #{@mrz2}".rstrip
+        out_str << "\n:MRZ: #{@mrz}".rstrip
+        out_str[1..].to_s
+      end
+
+      # @return [Boolean]
+      def expired?
+        return true unless @expiry_date.date_object
+
+        @expiry_date.date_object < Date.today
+      end
+
+      private
+
+      def check_mrz
+        return if @mrz1.value.nil? || @mrz2.value.nil?
+
+        mrz = MRZ.parse([@mrz1.value, @mrz2.value])
+        checks = {
+          mrz_valid: valid_mrz?(mrz),
+          mrz_valid_birth_date: valid_birth_date?(mrz),
+          mrz_valid_expiry_date: valid_expiry_date?(mrz),
+          mrz_valid_id_number: valid_id_number?(mrz),
+          mrz_valid_surname: valid_surname?(mrz),
+          mrz_valid_country: valid_country?(mrz),
+        }
+        @checklist.merge!(checks)
+      end
+
+      def valid_mrz?(mrz)
+        check = mrz.valid?
+        @mrz.confidence = 1.0 if check
+        check
+      end
+
+      def valid_birth_date?(mrz)
+        check = mrz.valid_birth_date? && mrz.birth_date == @birth_date.date_object
+        @birth_date.confidence = 1.0 if check
+        check
+      end
+
+      def valid_expiry_date?(mrz)
+        check = mrz.valid_expiration_date? && mrz.expiration_date == @expiry_date.date_object
+        @expiry_date.confidence = 1.0 if check
+        check
+      end
+
+      def valid_id_number?(mrz)
+        check = mrz.valid_document_number? && mrz.document_number == @id_number.value
+        @id_number.confidence = 1.0 if check
+        check
+      end
+
+      def valid_surname?(mrz)
+        check = mrz.last_name == @surname.value
+        @surname.confidence = 1.0 if check
+        check
+      end
+
+      def valid_country?(mrz)
+        check = mrz.nationality == @country.value
+        @country.confidence = 1.0 if check
+        check
+      end
+
+      def construct_full_name(page_id)
+        return unless @surname.value &&
+                      !@given_names.empty? &&
+                      @given_names[0].value
+
+        full_name = {
+          'value' => "#{@given_names[0].value} #{@surname.value}",
+          'confidence' => TextField.array_confidence([@surname, @given_names[0]]),
+        }
+        TextField.new(full_name, page_id, reconstructed: true)
+      end
+
+      def construct_mrz(page_id)
+        return if @mrz1.value.nil? || @mrz2.value.nil?
+
+        mrz = {
+          'value' => @mrz1.value + @mrz2.value,
+          'confidence' => Mindee::TextField.array_confidence([@mrz1, @mrz2]),
+        }
+        TextField.new(mrz, page_id, reconstructed: true)
+      end
+    end
+  end
+end

data/lib/mindee/parsing/prediction/receipt/receipt_v4.rb

@@ -0,0 +1,84 @@
+# frozen_string_literal: true
+
+require_relative '../common_fields'
+require_relative '../base'
+
+module Mindee
+  module Prediction
+    # Receipt document.
+    class ReceiptV4 < Prediction
+      # Where the purchase was made, the language, and the currency.
+      # @return [Mindee::Locale]
+      attr_reader :locale
+      # Total including taxes
+      # @return [Mindee::AmountField]
+      attr_reader :total_amount
+      # Total amount of the purchase excluding taxes.
+      # @return [Mindee::AmountField]
+      attr_reader :total_net
+      # Total tax amount of the purchase.
+      # @return [Mindee::AmountField]
+      attr_reader :total_tax
+      # The purchase date.
+      # @return [Mindee::DateField]
+      attr_reader :date
+      # The name of the supplier or merchant, as seen on the receipt.
+      # @return [Mindee::TextField]
+      attr_reader :supplier
+      # List of taxes detected on the receipt.
+      # @return [Array<Mindee::TaxField>]
+      attr_reader :taxes
+      # Time as seen on the receipt in HH:MM format.
+      # @return [Mindee::TextField]
+      attr_reader :time
+      # The receipt category among predefined classes.
+      # @return [Mindee::TextField]
+      attr_reader :category
+      # The receipt sub-category among predefined classes.
+      # @return [Mindee::TextField]
+      attr_reader :subcategory
+      # Whether the document is an expense receipt or a credit card receipt.
+      # @return [Mindee::TextField]
+      attr_reader :document_type
+
+      # @param prediction [Hash]
+      # @param page_id [Integer, nil]
+      def initialize(prediction, page_id)
+        super
+        @locale = Locale.new(prediction['locale'])
+        @total_amount = AmountField.new(prediction['total_amount'], page_id)
+        @total_net = AmountField.new(prediction['total_net'], page_id)
+        @total_tax = AmountField.new(prediction['total_tax'], page_id)
+        @tip = AmountField.new(prediction['tip'], page_id)
+        @date = DateField.new(prediction['date'], page_id)
+        @category = TextField.new(prediction['category'], page_id)
+        @subcategory = TextField.new(prediction['subcategory'], page_id)
+        @document_type = TextField.new(prediction['document_type'], page_id)
+        @supplier = TextField.new(prediction['supplier'], page_id)
+        @time = TextField.new(prediction['time'], page_id)
+        @taxes = []
+        prediction['taxes'].each do |item|
+          @taxes.push(TaxField.new(item, page_id))
+        end
+      end
+
+      def to_s
+        taxes = @taxes.join("\n       ")
+        out_str = String.new
+        out_str << "\n:Locale: #{@locale}".rstrip
+        out_str << "\n:Date: #{@date}".rstrip
+        out_str << "\n:Category: #{@category}".rstrip
+        out_str << "\n:Subcategory: #{@subcategory}".rstrip
+        out_str << "\n:Document type: #{@document_type}".rstrip
+        out_str << "\n:Time: #{@time}".rstrip
+        out_str << "\n:Supplier name: #{@supplier}".rstrip
+        out_str << "\n:Taxes: #{taxes}".rstrip
+        out_str << "\n:Total net: #{@total_net}".rstrip
+        out_str << "\n:Total taxes: #{@total_tax}".rstrip
+        out_str << "\n:Tip: #{@tip}".rstrip
+        out_str << "\n:Total amount: #{@total_amount}".rstrip
+        out_str[1..].to_s
+      end
+    end
+  end
+end

data/lib/mindee/parsing/prediction/shipping_container/shipping_container_v1.rb

@@ -0,0 +1,38 @@
+# frozen_string_literal: true
+
+require_relative '../common_fields'
+require_relative '../base'
+
+module Mindee
+  module Prediction
+    # License plate prediction.
+    class ShippingContainerV1 < Prediction
+      # ISO 6346 code for container owner prefix + equipment identifier.
+      # @return [Array<Mindee::TextField>]
+      attr_reader :owner
+      # ISO 6346 code for container serial number (6+1 digits).
+      # @return [Array<Mindee::TextField>]
+      attr_reader :serial_number
+      # ISO 6346 code for container length, height and type.
+      # @return [Array<Mindee::TextField>]
+      attr_reader :size_type
+
+      # @param prediction [Hash]
+      # @param page_id [Integer, nil]
+      def initialize(prediction, page_id)
+        super
+        @owner = TextField.new(prediction['owner'], page_id)
+        @serial_number = TextField.new(prediction['serial_number'], page_id)
+        @size_type = TextField.new(prediction['size_type'], page_id)
+      end
+
+      def to_s
+        out_str = String.new
+        out_str << "\n:Owner: #{@owner}".rstrip
+        out_str << "\n:Serial number: #{@serial_number}".rstrip
+        out_str << "\n:Size and type: #{@size_type}".rstrip
+        out_str[1..].to_s
+      end
+    end
+  end
+end

data/lib/mindee/parsing/prediction/us/bank_check/bank_check_v1.rb

@@ -0,0 +1,70 @@
+# frozen_string_literal: true
+
+require_relative '../../common_fields'
+require_relative '../../base'
+
+module Mindee
+  module Prediction
+    module US
+      # License plate prediction.
+      class BankCheckV1 < Prediction
+        # Payer's bank account number.
+        # @return [Array<Mindee::TextField>]
+        attr_reader :account_number
+        # Amount to be paid.
+        # @return [Array<Mindee::AmountField>]
+        attr_reader :amount
+        # The check number.
+        # @return [Mindee::TextField]
+        attr_reader :check_number
+        # Payer's bank account routing number.
+        # @return [Mindee::TextField]
+        attr_reader :routing_number
+        # Date the check was issued.
+        # @return [Mindee::DateField]
+        attr_reader :date
+        # The positions of the signatures on the image.
+        # @return [Array<Mindee::PositionField>]
+        attr_reader :signatures_positions
+        # Check's position in the image.
+        # @return [Mindee::PositionField]
+        attr_reader :check_position
+        # List of payees (full name or company name).
+        # @return [Array<Mindee::TextField>]
+        attr_reader :payees
+
+        # @param prediction [Hash]
+        # @param page_id [Integer, nil]
+        def initialize(prediction, page_id)
+          super
+          @account_number = TextField.new(prediction['account_number'], page_id)
+          @amount = AmountField.new(prediction['amount'], page_id)
+          @check_number = TextField.new(prediction['check_number'], page_id)
+          @date = DateField.new(prediction['date'], page_id)
+          @check_position = PositionField.new(prediction['check_position'], page_id)
+          @routing_number = TextField.new(prediction['routing_number'], page_id)
+          @signatures_positions = []
+          prediction['signatures_positions'].each do |item|
+            @signatures_positions.push(PositionField.new(item, page_id))
+          end
+          @payees = []
+          prediction['payees'].each do |item|
+            @payees.push(TextField.new(item, page_id))
+          end
+        end
+
+        def to_s
+          payees = @payees.map(&:value).join(', ')
+          out_str = String.new
+          out_str << "\n:Routing number: #{@routing_number}".rstrip
+          out_str << "\n:Account number: #{@account_number}".rstrip
+          out_str << "\n:Check number: #{@check_number}".rstrip
+          out_str << "\n:Date: #{@date}".rstrip
+          out_str << "\n:Amount: #{@amount}".rstrip
+          out_str << "\n:Payees: #{payees}".rstrip
+          out_str[1..].to_s
+        end
+      end
+    end
+  end
+end

data/lib/mindee/parsing/prediction.rb

@@ -0,0 +1,12 @@
+# frozen_string_literal: true
+
+require_relative 'prediction/custom/custom_v1'
+require_relative 'prediction/invoice/invoice_v4'
+require_relative 'prediction/passport/passport_v1'
+require_relative 'prediction/receipt/receipt_v4'
+require_relative 'prediction/eu/license_plate/license_plate_v1'
+require_relative 'prediction/shipping_container/shipping_container_v1'
+require_relative 'prediction/us/bank_check/bank_check_v1'
+require_relative 'prediction/fr/bank_account_details/bank_account_details_v1'
+require_relative 'prediction/fr/carte_vitale/carte_vitale_v1'
+require_relative 'prediction/fr/id_card/id_card_v1'

data/lib/mindee/parsing.rb

@@ -0,0 +1,4 @@
+# frozen_string_literal: true
+
+require_relative 'parsing/prediction'
+require_relative 'parsing/document'

data/lib/mindee/version.rb

@@ -2,7 +2,7 @@
 
 # Mindee
 module Mindee
-  VERSION = '1.2.0'
+  VERSION = '2.0.0'
 
   def self.find_platform
     host = RbConfig::CONFIG['host_os']

data/mindee.gemspec

@@ -31,9 +31,10 @@ Gem::Specification.new do |spec|
 
   spec.add_runtime_dependency 'marcel', '~> 1.0.2'
   spec.add_runtime_dependency 'mrz', '~> 0.2'
-  spec.add_runtime_dependency 'origamindee', '~> 3.0'
+  spec.add_runtime_dependency 'origamindee', '~> 3.1'
 
   spec.add_development_dependency 'rake', '~> 12.3.1'
   spec.add_development_dependency 'rspec', '~> 3.11'
   spec.add_development_dependency 'rubocop', '~> 1.41'
+  spec.add_development_dependency 'yard', '0.9.28'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: mindee
 version: !ruby/object:Gem::Version
-  version: 1.2.0
+  version: 2.0.0
 platform: ruby
 authors:
 - Mindee, SA
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2022-12-26 00:00:00.000000000 Z
+date: 2023-01-13 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: marcel
@@ -44,14 +44,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '3.0'
+        version: '3.1'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '3.0'
+        version: '3.1'
 - !ruby/object:Gem::Dependency
   name: rake
   requirement: !ruby/object:Gem::Requirement
@@ -94,6 +94,20 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '1.41'
+- !ruby/object:Gem::Dependency
+  name: yard
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '='
+      - !ruby/object:Gem::Version
+        version: 0.9.28
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '='
+      - !ruby/object:Gem::Version
+        version: 0.9.28
 description: Quickly and easily connect to Mindee's API services using Ruby.
 email:
 - devrel@mindee.co
@@ -107,6 +121,7 @@ files:
 - ".gitignore"
 - ".gitmodules"
 - ".rubocop.yml"
+- ".yardopts"
 - CHANGELOG.md
 - CODE_OF_CONDUCT.md
 - CONTRIBUTING.md
@@ -116,30 +131,48 @@ files:
 - Rakefile
 - bin/console
 - bin/mindee.rb
+- docs/ruby-api-builder.md
+- docs/ruby-getting-started.md
+- docs/ruby-invoice-ocr.md
+- docs/ruby-passport-ocr.md
+- docs/ruby-receipt-ocr.md
 - lib/mindee.rb
 - lib/mindee/client.rb
 - lib/mindee/document_config.rb
-- lib/mindee/documents.rb
-- lib/mindee/documents/base.rb
-- lib/mindee/documents/custom.rb
-- lib/mindee/documents/financial_doc.rb
-- lib/mindee/documents/invoice.rb
-- lib/mindee/documents/passport.rb
-- lib/mindee/documents/receipt.rb
-- lib/mindee/endpoint.rb
-- lib/mindee/fields.rb
-- lib/mindee/fields/amount.rb
-- lib/mindee/fields/base.rb
-- lib/mindee/fields/company_registration.rb
-- lib/mindee/fields/custom_docs.rb
-- lib/mindee/fields/datefield.rb
-- lib/mindee/fields/locale.rb
-- lib/mindee/fields/orientation.rb
-- lib/mindee/fields/payment_details.rb
-- lib/mindee/fields/tax.rb
 - lib/mindee/geometry.rb
-- lib/mindee/inputs.rb
-- lib/mindee/response.rb
+- lib/mindee/http/endpoint.rb
+- lib/mindee/input.rb
+- lib/mindee/input/pdf_processing.rb
+- lib/mindee/input/sources.rb
+- lib/mindee/parsing.rb
+- lib/mindee/parsing/document.rb
+- lib/mindee/parsing/error.rb
+- lib/mindee/parsing/inference.rb
+- lib/mindee/parsing/page.rb
+- lib/mindee/parsing/prediction.rb
+- lib/mindee/parsing/prediction/base.rb
+- lib/mindee/parsing/prediction/common_fields.rb
+- lib/mindee/parsing/prediction/common_fields/amount.rb
+- lib/mindee/parsing/prediction/common_fields/base.rb
+- lib/mindee/parsing/prediction/common_fields/company_registration.rb
+- lib/mindee/parsing/prediction/common_fields/date.rb
+- lib/mindee/parsing/prediction/common_fields/locale.rb
+- lib/mindee/parsing/prediction/common_fields/payment_details.rb
+- lib/mindee/parsing/prediction/common_fields/position.rb
+- lib/mindee/parsing/prediction/common_fields/tax.rb
+- lib/mindee/parsing/prediction/common_fields/text.rb
+- lib/mindee/parsing/prediction/custom/custom_v1.rb
+- lib/mindee/parsing/prediction/custom/fields.rb
+- lib/mindee/parsing/prediction/eu/license_plate/license_plate_v1.rb
+- lib/mindee/parsing/prediction/fr/bank_account_details/bank_account_details_v1.rb
+- lib/mindee/parsing/prediction/fr/carte_vitale/carte_vitale_v1.rb
+- lib/mindee/parsing/prediction/fr/id_card/id_card_v1.rb
+- lib/mindee/parsing/prediction/invoice/invoice_line_item.rb
+- lib/mindee/parsing/prediction/invoice/invoice_v4.rb
+- lib/mindee/parsing/prediction/passport/passport_v1.rb
+- lib/mindee/parsing/prediction/receipt/receipt_v4.rb
+- lib/mindee/parsing/prediction/shipping_container/shipping_container_v1.rb
+- lib/mindee/parsing/prediction/us/bank_check/bank_check_v1.rb
 - lib/mindee/version.rb
 - mindee.gemspec
 homepage: https://github.com/mindee/mindee-api-ruby

data/lib/mindee/documents/base.rb

@@ -1,35 +0,0 @@
-# frozen_string_literal: true
-
-module Mindee
-  # Base document object.
-  class Document
-    # document type
-    # @return [String]
-    attr_reader :document_type
-    # Validation checks for the document
-    # @return [Hash<Symbol, Boolean>]
-    attr_reader :checklist
-    # Original filename of the document
-    # @return [String, nil]
-    attr_reader :filename
-    # Detected MIME type of the document
-    # @return [String, nil]
-    attr_reader :file_mimetype
-
-    # @param input_file [Mindee::InputDocument, nil]
-    # @param document_type [String]
-    def initialize(document_type, input_file: nil)
-      @document_type = document_type
-      unless input_file.nil?
-        @filename = input_file.filename
-        @file_mimetype = input_file.file_mimetype
-      end
-      @checklist = {}
-    end
-
-    # @return [Boolean]
-    def all_checks
-      @checklist.all? { |_, value| value == true }
-    end
-  end
-end

data/lib/mindee/documents/custom.rb

@@ -1,65 +0,0 @@
-# frozen_string_literal: true
-
-require_relative '../fields'
-require_relative 'base'
-
-module Mindee
-  # Custom document object.
-  class CustomDocument < Document
-    # All value fields in the document
-    # @return [Hash<Symbol, Mindee::ListField>]
-    attr_reader :fields
-    # All classifications in the document
-    # @return [Hash<Symbol, Mindee::ClassificationField>]
-    attr_reader :classifications
-
-    # @param document_type [String]
-    # @param prediction [Hash]
-    # @param input_file [Mindee::InputDocument, nil]
-    # @param page_id [Integer, nil]
-    def initialize(document_type, prediction, input_file: nil, page_id: nil)
-      super(document_type, input_file: input_file)
-      @fields = {}
-      @classifications = {}
-      prediction.each do |field_name, field_prediction|
-        field_sym = field_name.to_sym
-        complete_field = set_field(field_sym, field_prediction, page_id)
-
-        # Create a dynamic accessor function for the field
-        singleton_class.module_eval { attr_accessor field_sym }
-        send("#{field_sym}=", complete_field)
-      end
-    end
-
-    def to_s
-      out_str = String.new
-      out_str << "----- #{@document_type} -----"
-      out_str << "\nFilename: #{@filename}".rstrip
-      @classifications.each do |name, info|
-        out_str << "\n#{name}: #{info}".rstrip
-      end
-      @fields.each do |name, info|
-        out_str << "\n#{name}: #{info}".rstrip
-      end
-      out_str << "\n----------------------"
-      out_str
-    end
-
-    private
-
-    # @param field_prediction [Hash]
-    def set_field(field_sym, field_prediction, page_id)
-      # Currently two types of fields possible in a custom API response:
-      # fields having a list of values, and classification fields.
-      # Here we use the fact that only value lists have the 'values' attribute.
-
-      if field_prediction.key? 'values'
-        @fields[field_sym] = ListField.new(field_prediction, page_id)
-      elsif field_prediction.key? 'value'
-        @classifications[field_sym] = ClassificationField.new(field_prediction)
-      else
-        throw 'Unknown API field type'
-      end
-    end
-  end
-end