⬆️ 💥 drop support for ruby < 3.2 and bump dependency

Author: sebastianMindee

.rubocop.yml

@@ -10,7 +10,7 @@ AllCops:
     - local_test/*
     - Steepfile
 
-  TargetRubyVersion: 3.0.0
+  TargetRubyVersion: 3.2
   SuggestExtensions: false
 
 Gemspec/DevelopmentDependencies:

lib/mindee/geometry/polygon.rb

@@ -7,7 +7,7 @@ module Geometry
     class Polygon < Array
       # @param server_response [Hash] Raw server response hash.
       def initialize(server_response)
-        points = []
+        points = [] # @type var points: Array[Mindee::Geometry::Point]
         server_response.map do |point|
           points << Point.new(point[0], point[1])
         end
@@ -30,7 +30,7 @@ def point_in_y?(point)
 
       # @return [String] Polygon as a string.
       def to_s
-        "(#{map(&:to_s).join(', ')})"
+        "(#{join(', ')})"
       end
     end
   end

lib/mindee/image/image_extractor.rb

@@ -47,7 +47,7 @@ def self.extract_multiple_images_from_source(input_source, page_id, polygons)
       # @param [Array<Geometry::Point, Geometry::Polygon, Geometry::Quadrilateral>] polygons
       # @return [Array<Image::ExtractedImage>] Extracted Images.
       def self.extract_images_from_polygons(input_source, pdf_stream, page_id, polygons)
-        extracted_elements = []
+        extracted_elements = [] # @type var extracted_elements: Array[Image::ExtractedImage]
 
         polygons.each_with_index do |polygon, element_id|
           polygon = ImageUtils.normalize_polygon(polygon)

lib/mindee/input/sources/path_input_source.rb

@@ -11,7 +11,7 @@ class PathInputSource < LocalInputSource
         # @param filepath [String]
         # @param repair_pdf [bool]
         def initialize(filepath, repair_pdf: false)
-          io_stream = File.open(filepath, 'rb')
+          io_stream = File.new(filepath, 'rb')
           super(io_stream, File.basename(filepath), repair_pdf: repair_pdf)
         end
       end

lib/mindee/page_options.rb

@@ -11,7 +11,7 @@ class PageOptions
     attr_accessor :page_indexes, :operation, :on_min_pages
 
     def initialize(params: {})
-      params ||= {}
+      params ||= {} # : Hash[Symbol, untyped]
       params = params.transform_keys(&:to_sym)
       @page_indexes = params.fetch(
         :page_indexes,

lib/mindee/pdf/extracted_pdf.rb

@@ -13,11 +13,17 @@ class ExtractedPDF
       # @return [String]
       attr_reader :filename
 
-      # @param pdf_bytes [StringIO]
+      # @param pdf_stream [StringIO, File]
       # @param filename [String]
-      def initialize(pdf_bytes, filename)
-        @pdf_bytes = pdf_bytes
+      def initialize(pdf_stream, filename)
         @filename = filename
+
+        if pdf_stream.is_a?(File)
+          pdf_stream.rewind
+          @pdf_bytes = StringIO.new(pdf_stream.read)
+        else
+          @pdf_bytes = pdf_stream
+        end
       end
 
       # Retrieves the page count for a given pdf.

lib/mindee/pdf/pdf_extractor.rb

@@ -40,7 +40,7 @@ def cut_pages(page_indexes)
       # @param page_indexes [Array<Array<Integer>>] List of page number to use for merging in the original Pdf.
       # @return [Array<Mindee::PDF::ExtractedPDF>] The buffer containing the new Pdf.
       def extract_sub_documents(page_indexes)
-        extracted_pdfs = []
+        extracted_pdfs = [] # @type var extracted_pdfs: Array[Mindee::PDF::ExtractedPDF]
         extension = File.extname(@filename)
         basename = File.basename(@filename, extension)
         page_indexes.each do |page_index_list|
@@ -54,7 +54,7 @@ def extract_sub_documents(page_indexes)
                     "Index #{page_index} is out of range."
             end
           end
-          formatted_max_index = format('%03d', page_index_list[page_index_list.length - 1] + 1).to_s
+          formatted_max_index = format('%03d', page_index_list[-1] + 1).to_s
           field_filename = "#{basename}_#{format('%03d',
                                                  page_index_list[0] + 1)}-#{formatted_max_index}#{extension}"
           extracted_pdf = Mindee::PDF::ExtractedPDF.new(cut_pages(page_index_list),
@@ -74,15 +74,15 @@ def extract_sub_documents(page_indexes)
       def extract_invoices(page_indexes, strict: false)
         raise Error::MindeePDFError, 'No indexes provided.' if page_indexes.empty?
 
-        if page_indexes[0].is_a?(Array) && page_indexes[0].all? { |i| i.is_a?(Integer) }
+        if page_indexes[0].is_a?(Array) && page_indexes[0].all?(Integer)
           page_indexes_as_array = page_indexes # @type var page_indexes : Array[Array[Integer]]
           return extract_sub_documents(page_indexes_as_array)
         end
         p_ids = page_indexes # @type var page_indexes: Product::InvoiceSplitter::InvoiceSplitterV1InvoicePageGroups
         return extract_sub_documents(p_ids.map(&:page_indexes)) unless strict
 
-        correct_page_indexes = []
-        current_list = []
+        correct_page_indexes = [] # @type var correct_page_indexes: Array[Array[Integer]]
+        current_list = [] # @type var current_list: Array[Integer]
         previous_confidence = nil
         p_ids.each_with_index do |p_i, i|
           page_index = p_i # @type var page_index: Product::InvoiceSplitter::InvoiceSplitterV1InvoicePageGroup

lib/mindee/pdf/pdf_processor.rb

@@ -1,6 +1,5 @@
 # frozen_string_literal: true
 
-require 'set'
 require 'origami'
 require_relative 'pdf_tools'
 

lib/mindee/v1/extraction/multi_receipts_extractor.rb

@@ -11,7 +11,7 @@ module Extraction
       # @param inference [Inference] Results of the inference.
       # @return [Array<ExtractedImage>] Individual extracted receipts as an array of ExtractedMultiReceiptsImage.
       def self.extract_receipts(input_source, inference)
-        images = []
+        images = [] # @type var images: Array[Image::ExtractedImage]
         unless inference.prediction.receipts
           raise Error::MindeeInputError,
                 'No possible receipts candidates found for Multi-Receipts extraction.'

lib/mindee/v1/parsing/common/extras/cropper_extra.rb

@@ -22,7 +22,7 @@ def initialize(raw_prediction, page_id = nil)
 
             # @return [String]
             def to_s
-              @croppings.map(&:to_s).join("\n           ")
+              @croppings.join("\n           ")
             end
           end
         end