kreuzberg 4.3.5-aarch64-linux → 4.3.7-aarch64-linux

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4e7a8855affaf95c9e76424a8389a5c8b398022ee7fbfe0cd5f81acd42b4441f
-  data.tar.gz: 827920bb1c20a6885dd27c6f96a2144771309b97cde2c31b7c907eb279132584
+  metadata.gz: 8519525216a517ce9d2078552b1492c4209c2d1a8bff7b58eae88a8137a603d4
+  data.tar.gz: ab098d31da39a6965c57f65d4a6a282570333ed6a0a8d00886c22982b1bce76c
 SHA512:
-  metadata.gz: 6047eb8aa9fd0937981cd300410fa8fd64b4d78938af56c5c66adcb4ef550662d4a407b1b6f5921fff9754cf9af648efb19b623d11110241c6de0d7d266a7890
-  data.tar.gz: 9f98e886089fe5994359f5e46185dbb90cb2fc2b4469a3fd0c9b3722fa20fdca75348efcef4bdcdf4c8e9f1d0072728470a84d6af795f68468afbb70b5f85865
+  metadata.gz: 209c8c32c54f5b73e90954ae9db136aa87d13e9738c48de6e89282396986cc705b1f7c64841a7358f6357d34980f023890c8031a6e10a32eb20237564fad3d30
+  data.tar.gz: c92af45298595d9bd433dd9b01223b45870063149042bd3d36b26d5201a794b78fffa2f79cf6d9621a80e5d65751ea7fec52756fce44a28e53de1d6a7cf79e2f

data/Gemfile.lock

@@ -1,8 +1,9 @@
 PATH
   remote: .
   specs:
-    kreuzberg (4.3.5)
+    kreuzberg (4.3.7)
       rb_sys (~> 0.9.119)
+      sorbet-runtime (~> 0.5)
 
 GEM
   remote: https://rubygems.org/
@@ -123,7 +124,7 @@ GEM
       rubocop (~> 1.81)
     ruby-progressbar (1.13.0)
     securerandom (0.4.1)
-    sorbet-runtime (0.6.12942)
+    sorbet-runtime (0.6.12956)
     steep (1.10.0)
       activesupport (>= 5.1)
       concurrent-ruby (>= 1.1.10)
@@ -179,7 +180,6 @@ DEPENDENCIES
   rubocop (~> 1.66)
   rubocop-performance (~> 1.21)
   rubocop-rspec (~> 3.0)
-  sorbet-runtime (~> 0.5)
   steep (~> 1.8)
   yard (~> 0.9)
 
@@ -210,7 +210,7 @@ CHECKSUMS
   i18n (1.14.8) sha256=285778639134865c5e0f6269e0b818256017e8cde89993fdfcbfb64d088824a5
   io-console (0.8.2) sha256=d6e3ae7a7cc7574f4b8893b4fca2162e57a825b223a177b7afa236c5ef9814cc
   json (2.18.1) sha256=fe112755501b8d0466b5ada6cf50c8c3f41e897fa128ac5d263ec09eedc9f986
-  kreuzberg (4.3.5)
+  kreuzberg (4.3.7)
   language_server-protocol (3.17.0.5) sha256=fd1e39a51a28bf3eec959379985a72e296e9f9acfce46f6a79d31ca8760803cc
   lint_roller (1.1.0) sha256=2c0c845b632a7d172cb849cc90c1bce937a28c5c8ccccb50dfd46a485003cc87
   listen (3.10.0) sha256=c6e182db62143aeccc2e1960033bebe7445309c7272061979bb098d03760c9d2
@@ -245,7 +245,7 @@ CHECKSUMS
   rubocop-rspec (3.9.0) sha256=8fa70a3619408237d789aeecfb9beef40576acc855173e60939d63332fdb55e2
   ruby-progressbar (1.13.0) sha256=80fc9c47a9b640d6834e0dc7b3c94c9df37f08cb072b7761e4a71e22cff29b33
   securerandom (0.4.1) sha256=cc5193d414a4341b6e225f0cb4446aceca8e50d5e1888743fac16987638ea0b1
-  sorbet-runtime (0.6.12942) sha256=967bda04814d234e4239c4f883c1d0ee6de3e47bf8bafd2c0cc30d18df2ddd3a
+  sorbet-runtime (0.6.12956) sha256=fee716a62d0b1d94ebc8e6ba23e76a7654eeac66c1f5cc1e1bef78b8e9ff87c7
   steep (1.10.0) sha256=1b295b55f9aaff1b8d3ee42453ee55bc2a1078fda0268f288edb2dc014f4d7d1
   strscan (3.1.7) sha256=5f76462b94a3ea50b44973225b7d75b2cb96d4e1bee9ef1319b99ca117b72c8c
   terminal-table (4.0.0) sha256=f504793203f8251b2ea7c7068333053f0beeea26093ec9962e62ea79f94301d2

data/README.md

@@ -22,7 +22,7 @@
     <img src="https://img.shields.io/maven-central/v/dev.kreuzberg/kreuzberg?label=Java&color=007ec6" alt="Java">
   </a>
   <a href="https://github.com/kreuzberg-dev/kreuzberg/releases">
-    <img src="https://img.shields.io/github/v/tag/kreuzberg-dev/kreuzberg?label=Go&color=007ec6&filter=v4.3.5" alt="Go">
+    <img src="https://img.shields.io/github/v/tag/kreuzberg-dev/kreuzberg?label=Go&color=007ec6&filter=v4.3.7" alt="Go">
   </a>
   <a href="https://www.nuget.org/packages/Kreuzberg/">
     <img src="https://img.shields.io/nuget/v/Kreuzberg?label=C%23&color=007ec6" alt="C#">

data/kreuzberg.gemspec

@@ -241,7 +241,7 @@ Gem::Specification.new do |spec|
   spec.add_development_dependency 'rake', '~> 13.0'
   spec.add_development_dependency 'rake-compiler', '~> 1.2'
   spec.add_development_dependency 'rspec', '~> 3.12'
-  spec.add_development_dependency 'sorbet-runtime', '~> 0.5'
+  spec.add_dependency 'sorbet-runtime', '~> 0.5'
   unless Gem.win_platform?
     spec.add_development_dependency 'rbs', '~> 3.0'
     spec.add_development_dependency 'rubocop', '~> 1.66'

data/lib/kreuzberg/config.rb

@@ -391,14 +391,18 @@ module Kreuzberg
     #   pdf = PDF.new(extract_images: true, hierarchy: hierarchy)
     #
     class PDF
-      attr_reader :extract_images, :passwords, :extract_metadata, :font_config, :hierarchy
+      attr_reader :extract_images, :passwords, :extract_metadata, :font_config, :hierarchy,
+                  :extract_annotations, :top_margin_fraction, :bottom_margin_fraction
 
       def initialize(
         extract_images: false,
         passwords: nil,
         extract_metadata: true,
         font_config: nil,
-        hierarchy: nil
+        hierarchy: nil,
+        extract_annotations: false,
+        top_margin_fraction: nil,
+        bottom_margin_fraction: nil
       )
         @extract_images = extract_images ? true : false
         @passwords = if passwords.is_a?(Array)
@@ -409,6 +413,9 @@ module Kreuzberg
         @extract_metadata = extract_metadata ? true : false
         @font_config = normalize_font_config(font_config)
         @hierarchy = normalize_hierarchy(hierarchy)
+        @extract_annotations = extract_annotations ? true : false
+        @top_margin_fraction = top_margin_fraction&.to_f
+        @bottom_margin_fraction = bottom_margin_fraction&.to_f
       end
 
       def to_h
@@ -417,7 +424,10 @@ module Kreuzberg
           passwords: @passwords,
           extract_metadata: @extract_metadata,
           font_config: @font_config&.to_h,
-          hierarchy: @hierarchy&.to_h
+          hierarchy: @hierarchy&.to_h,
+          extract_annotations: @extract_annotations,
+          top_margin_fraction: @top_margin_fraction,
+          bottom_margin_fraction: @bottom_margin_fraction
         }.compact
       end
 

data/lib/kreuzberg/result.rb

@@ -14,7 +14,7 @@ module Kreuzberg
   class Result
     attr_reader :content, :mime_type, :metadata, :metadata_json, :tables,
                 :detected_languages, :chunks, :images, :pages, :elements, :ocr_elements, :djot_content,
-                :document, :extracted_keywords, :quality_score, :processing_warnings
+                :document, :extracted_keywords, :quality_score, :processing_warnings, :annotations
 
     # @!attribute [r] cells
     #   @return [Array<Array<String>>] Table cells (2D array)
@@ -339,6 +339,7 @@ module Kreuzberg
       @extracted_keywords = parse_extracted_keywords(get_value(hash, 'extracted_keywords'))
       @quality_score = get_value(hash, 'quality_score')
       @processing_warnings = parse_processing_warnings(get_value(hash, 'processing_warnings'))
+      @annotations = parse_annotations(get_value(hash, 'annotations'))
     end
     # rubocop:enable Metrics/AbcSize
 
@@ -346,6 +347,7 @@ module Kreuzberg
     #
     # @return [Hash] Hash representation
     #
+    # rubocop:disable Metrics/CyclomaticComplexity
     def to_h
       {
         content: @content,
@@ -362,9 +364,11 @@ module Kreuzberg
         document: @document&.to_h,
         extracted_keywords: @extracted_keywords&.map(&:to_h),
         quality_score: @quality_score,
-        processing_warnings: @processing_warnings.map(&:to_h)
+        processing_warnings: @processing_warnings.map(&:to_h),
+        annotations: @annotations&.map(&:to_h)
       }
     end
+    # rubocop:enable Metrics/CyclomaticComplexity
 
     # Convert to JSON
     #
@@ -707,6 +711,32 @@ module Kreuzberg
         )
       end
     end
+
+    def parse_annotations(annotations_data)
+      return nil if annotations_data.nil?
+
+      annotations_data.map { |a_hash| build_annotation(a_hash) }
+    end
+
+    def build_annotation(a_hash)
+      PdfAnnotation.new(
+        annotation_type: a_hash['annotation_type'] || '',
+        content: a_hash['content'],
+        page_number: a_hash['page_number']&.to_i,
+        bounding_box: build_annotation_bbox(a_hash['bounding_box'])
+      )
+    end
+
+    def build_annotation_bbox(bbox_data)
+      return nil if bbox_data.nil?
+
+      PdfAnnotationBoundingBox.new(
+        left: bbox_data['left']&.to_f,
+        top: bbox_data['top']&.to_f,
+        right: bbox_data['right']&.to_f,
+        bottom: bbox_data['bottom']&.to_f
+      )
+    end
   end
   # rubocop:enable Metrics/ClassLength
 end

data/lib/kreuzberg/types.rb

@@ -411,4 +411,24 @@ module Kreuzberg
 
     const :nodes, T::Array[DocumentNode]
   end
+
+  # Bounding box for a PDF annotation.
+  class PdfAnnotationBoundingBox < T::Struct
+    extend T::Sig
+
+    const :left, T.nilable(Float)
+    const :top, T.nilable(Float)
+    const :right, T.nilable(Float)
+    const :bottom, T.nilable(Float)
+  end
+
+  # A PDF annotation extracted from a document page.
+  class PdfAnnotation < T::Struct
+    extend T::Sig
+
+    const :annotation_type, String
+    const :content, T.nilable(String)
+    const :page_number, T.nilable(Integer)
+    const :bounding_box, T.nilable(PdfAnnotationBoundingBox)
+  end
 end

data/lib/kreuzberg/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Kreuzberg
-  VERSION = '4.3.5'
+  VERSION = '4.3.7'
 end

data/sig/kreuzberg.rbs

@@ -323,8 +323,11 @@ module Kreuzberg
       attr_reader extract_metadata: bool
       attr_reader font_config: FontConfig?
       attr_reader hierarchy: Hierarchy?
+      attr_reader extract_annotations: bool
+      attr_reader top_margin_fraction: Float?
+      attr_reader bottom_margin_fraction: Float?
 
-      def initialize: (?extract_images: bool, ?passwords: (Array[String] | String)?, ?extract_metadata: bool, ?font_config: (FontConfig | Hash[Symbol, untyped])?, ?hierarchy: (Hierarchy | Hash[Symbol, untyped])?) -> void
+      def initialize: (?extract_images: bool, ?passwords: (Array[String] | String)?, ?extract_metadata: bool, ?font_config: (FontConfig | Hash[Symbol, untyped])?, ?hierarchy: (Hierarchy | Hash[Symbol, untyped])?, ?extract_annotations: bool, ?top_margin_fraction: Float?, ?bottom_margin_fraction: Float?) -> void
       def to_h: () -> Hash[Symbol, untyped]
     end
 
@@ -525,6 +528,15 @@ module Kreuzberg
   end
 
   # Extraction result type
+  type pdf_annotation_type = 'text' | 'highlight' | 'link' | 'stamp' | 'underline' | 'strike_out' | 'other'
+
+  type pdf_annotation_hash = {
+    annotation_type: String,
+    content: String?,
+    page_number: Integer,
+    bounding_box: { x0: Float, y0: Float, x1: Float, y1: Float }?
+  }
+
   type extraction_result_hash = {
     content: String,
     mime_type: String,
@@ -541,7 +553,8 @@ module Kreuzberg
     document: document_structure_hash?,
     extracted_keywords: Array[extracted_keyword_hash]?,
     quality_score: Float?,
-    processing_warnings: Array[processing_warning_hash]?
+    processing_warnings: Array[processing_warning_hash]?,
+    annotations: Array[pdf_annotation_hash]?
   }
 
   type extracted_keyword_hash = {
@@ -1076,6 +1089,18 @@ module Kreuzberg
     attr_reader extracted_keywords: Array[ExtractedKeyword]?
     attr_reader quality_score: Float?
     attr_reader processing_warnings: Array[ProcessingWarning]?
+    attr_reader annotations: Array[PdfAnnotation]?
+
+    # PDF annotation extracted from a document page (Struct from result.rb)
+    class PdfAnnotation
+      attr_reader annotation_type: String
+      attr_reader content: String?
+      attr_reader page_number: Integer
+      attr_reader bounding_box: BoundingBox?
+
+      def initialize: (annotation_type: String, content: String?, page_number: Integer, bounding_box: BoundingBox?) -> void
+      def to_h: () -> pdf_annotation_hash
+    end
 
     def initialize: (extraction_result_hash hash) -> void
     def to_h: () -> Hash[Symbol, untyped]
@@ -1113,6 +1138,7 @@ module Kreuzberg
     def parse_ocr_geometry: (Hash[String, untyped]? data) -> OcrBoundingGeometry?
     def parse_ocr_confidence: (Hash[String, untyped]? data) -> OcrConfidence?
     def parse_ocr_rotation: (Hash[String, untyped]? data) -> OcrRotation?
+    def parse_annotations: (Array[pdf_annotation_hash]? annotations_data) -> Array[PdfAnnotation]?
   end
 
   # Module methods (extraction API)