kreuzberg 4.6.1-aarch64-linux → 4.6.3-aarch64-linux

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 7d632985bd042b73c0e51c84e292d9626267f205992bf49c9c62bcef78d9656f
-  data.tar.gz: ded364d7414b8fb4eafbda15af492698338799eb86749f9708f3f595ab7cad6d
+  metadata.gz: adae55dc7f30e68a211cc0493985d0d1687b3988e76509fbffd87f91fee45207
+  data.tar.gz: 6071b7d76b01dc15b47a11fc5eaeb4292fbb07630d20c3ac113751bbded3de0f
 SHA512:
-  metadata.gz: 5714ba27e6e41a99db90b829cd0c02986eab8c522673f62635ec0c5457d3d7367d290ef1146486fcccdda44fdb55751a99b24758454fdd58c90d93b83b77656e
-  data.tar.gz: ff4e355557401115eb34c227ddc4cd573e6b877f830b5b1a13b749508d202cd691866122e2be4132a883e7c80a4e7bae9f388296c79df7890adbd9aabcb822ca
+  metadata.gz: 1e90683694a29205d479b3cda7fb367658bf311520884f7c1faa1b4ec1d6be69dad491b620e33d521b8dea893e062c620738b0974157a096670e825a0fd1a434
+  data.tar.gz: ff29c19cb5b0085b84ba1a3ad9f97602ed83211e934d5eff1b8a630868a2ff7e919bba5937d7c49cb7a424fbc2239b726750cdca5c8893fdeb6cb4d98540f5fa

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=v4.6.1" alt="Go">
+    <img src="https://img.shields.io/github/v/tag/kreuzberg-dev/kreuzberg?label=Go&color=007ec6&filter=v4.6.3" 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/lib/kreuzberg/config.rb

@@ -837,6 +837,25 @@ module Kreuzberg
       end
     end
 
+    # Email extraction configuration
+    #
+    # @example With fallback codepage
+    #   email = Email.new(msg_fallback_codepage: 1251)
+    #
+    class Email
+      attr_reader :msg_fallback_codepage
+
+      def initialize(msg_fallback_codepage: nil)
+        @msg_fallback_codepage = msg_fallback_codepage&.to_i
+      end
+
+      def to_h
+        h = {}
+        h[:msg_fallback_codepage] = @msg_fallback_codepage unless @msg_fallback_codepage.nil?
+        h
+      end
+    end
+
     # Layout detection configuration
     #
     # @example Basic usage with fast preset
@@ -933,7 +952,8 @@ module Kreuzberg
                   :token_reduction, :keywords, :html_options, :pages,
                   :max_concurrent_extractions, :output_format, :result_format,
                   :security_limits, :layout, :concurrency,
-                  :cache_namespace, :cache_ttl_secs, :extraction_timeout_secs
+                  :cache_namespace, :cache_ttl_secs, :extraction_timeout_secs,
+                  :max_archive_depth, :acceleration, :email
 
       # Alias for backward compatibility - image_extraction is the canonical name
       alias image_extraction images
@@ -959,6 +979,7 @@ module Kreuzberg
         postprocessor token_reduction keywords html_options pages
         max_concurrent_extractions output_format result_format
         security_limits layout concurrency cache_namespace cache_ttl_secs extraction_timeout_secs
+        max_archive_depth acceleration email
       ].freeze
 
       # Aliases for backward compatibility
@@ -1015,7 +1036,7 @@ module Kreuzberg
         new(**normalize_hash_keys(hash))
       end
 
-      def initialize(hash = nil,
+      def initialize(hash = nil, # rubocop:disable Metrics/MethodLength
                      use_cache: true,
                      enable_quality_processing: true,
                      force_ocr: false,
@@ -1039,7 +1060,10 @@ module Kreuzberg
                      concurrency: nil,
                      cache_namespace: nil,
                      cache_ttl_secs: nil,
-                     extraction_timeout_secs: nil)
+                     extraction_timeout_secs: nil,
+                     max_archive_depth: 3,
+                     acceleration: nil,
+                     email: nil)
         kwargs = {
           use_cache: use_cache, enable_quality_processing: enable_quality_processing,
           force_ocr: force_ocr, force_ocr_pages: force_ocr_pages,
@@ -1054,7 +1078,10 @@ module Kreuzberg
           concurrency: concurrency,
           cache_namespace: cache_namespace,
           cache_ttl_secs: cache_ttl_secs,
-          extraction_timeout_secs: extraction_timeout_secs
+          extraction_timeout_secs: extraction_timeout_secs,
+          max_archive_depth: max_archive_depth,
+          acceleration: acceleration,
+          email: email
         }
         extracted = extract_from_hash(hash, kwargs)
 
@@ -1086,7 +1113,10 @@ module Kreuzberg
         @pages = normalize_config(params[:pages], PageConfig)
         @layout = normalize_config(params[:layout], LayoutDetection)
         @concurrency = normalize_config(params[:concurrency], Concurrency)
+        @acceleration = normalize_config(params[:acceleration], Acceleration)
+        @email = normalize_config(params[:email], Email)
         @max_concurrent_extractions = params[:max_concurrent_extractions]&.to_i
+        @max_archive_depth = params[:max_archive_depth]&.to_i || 3
         @output_format = validate_output_format(params[:output_format])
         @result_format = validate_result_format(params[:result_format])
         @cache_namespace = params[:cache_namespace]
@@ -1127,6 +1157,7 @@ module Kreuzberg
           force_ocr_pages: @force_ocr_pages,
           include_document_structure: @include_document_structure,
           max_concurrent_extractions: @max_concurrent_extractions,
+          max_archive_depth: @max_archive_depth,
           output_format: @output_format,
           result_format: @result_format,
           cache_namespace: @cache_namespace,
@@ -1142,7 +1173,8 @@ module Kreuzberg
           image_extraction: @images&.to_h, postprocessor: @postprocessor&.to_h,
           token_reduction: @token_reduction&.to_h, keywords: @keywords&.to_h,
           html_options: @html_options&.to_h, pages: @pages&.to_h,
-          layout: @layout&.to_h, concurrency: @concurrency&.to_h
+          layout: @layout&.to_h, concurrency: @concurrency&.to_h,
+          acceleration: @acceleration&.to_h, email: @email&.to_h
         }
       end
 
@@ -1286,6 +1318,12 @@ module Kreuzberg
           @layout = normalize_config(value, LayoutDetection)
         when :concurrency
           @concurrency = normalize_config(value, Concurrency)
+        when :acceleration
+          @acceleration = normalize_config(value, Acceleration)
+        when :email
+          @email = normalize_config(value, Email)
+        when :max_archive_depth
+          @max_archive_depth = value&.to_i || 3
         when :max_concurrent_extractions
           @max_concurrent_extractions = value&.to_i
         when :output_format
@@ -1373,6 +1411,9 @@ module Kreuzberg
         @html_options = merged.html_options
         @pages = merged.pages
         @layout = merged.layout
+        @acceleration = merged.acceleration
+        @email = merged.email
+        @max_archive_depth = merged.max_archive_depth
       end
 
       def update_output_options(merged)

data/lib/kreuzberg/extraction_api.rb

@@ -319,6 +319,43 @@ module Kreuzberg
       results
     end
 
+    # Render a single PDF page as a PNG image.
+    #
+    # @param path [String, Pathname] Path to the PDF file
+    # @param page_index [Integer] Zero-based page index
+    # @param dpi [Integer] Rendering resolution (default 150)
+    # @return [String] PNG-encoded binary string
+    # @raise [Errors::IOError] If the file cannot be read
+    # @raise [Errors::ParsingError] If rendering fails
+    def render_pdf_page(path, page_index, dpi: 150)
+      path_str = path.to_s
+      raise ArgumentError, 'page_index must be non-negative' if page_index.negative?
+      raise Errors::IOError, "File not found: #{path_str}" unless File.exist?(path_str)
+
+      native_render_pdf_page(path_str, page_index, dpi)
+    end
+
+    # Iterate over pages of a PDF lazily, yielding each page as it is rendered.
+    #
+    # Each page is rendered via the native FFI iterator, so only one page is in
+    # memory at a time.
+    #
+    # @param path [String, Pathname] Path to the PDF file
+    # @param dpi [Integer] Rendering resolution (default 150)
+    # @yieldparam page_index [Integer] Zero-based page index
+    # @yieldparam png_bytes [String] PNG-encoded binary string for the page
+    # @return [Enumerator] if no block is given
+    # @raise [Errors::IOError] If the file cannot be read
+    # @raise [Errors::ParsingError] If rendering fails
+    def render_pdf_pages_iter(path, dpi: 150, &block)
+      path_str = path.to_s
+      raise Errors::IOError, "File not found: #{path_str}" unless File.exist?(path_str)
+
+      return enum_for(:render_pdf_pages_iter, path, dpi: dpi) unless block
+
+      native_render_pdf_pages_iter(path_str, dpi, &block)
+    end
+
     def normalize_config(config)
       return {} if config.nil?
       return config if config.is_a?(Hash)

data/lib/kreuzberg/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Kreuzberg
-  VERSION = '4.6.1'
+  VERSION = '4.6.3'
 end

data/sig/kreuzberg.rbs

@@ -459,6 +459,21 @@ module Kreuzberg
       def to_h: () -> Hash[Symbol, untyped]
     end
 
+    class Acceleration
+      attr_reader provider: String
+      attr_reader device_id: Integer
+
+      def initialize: (?provider: String, ?device_id: Integer) -> void
+      def to_h: () -> Hash[Symbol, untyped]
+    end
+
+    class Email
+      attr_reader msg_fallback_codepage: Integer?
+
+      def initialize: (?msg_fallback_codepage: Integer?) -> void
+      def to_h: () -> Hash[Symbol, untyped]
+    end
+
     class LayoutDetection
       attr_reader preset: String
       attr_reader confidence_threshold: Float?
@@ -497,7 +512,10 @@ module Kreuzberg
       attr_reader pages: PageConfig?
       attr_reader layout: LayoutDetection?
       attr_reader concurrency: Concurrency?
+      attr_reader acceleration: Acceleration?
+      attr_reader email: Email?
       attr_reader max_concurrent_extractions: Integer?
+      attr_reader max_archive_depth: Integer
       attr_reader output_format: String?
       attr_reader result_format: String?
       attr_reader security_limits: Hash[String, Integer]?
@@ -524,7 +542,10 @@ module Kreuzberg
         ?pages: (PageConfig | Hash[Symbol, untyped])?,
         ?layout: (LayoutDetection | Hash[Symbol, untyped])?,
         ?concurrency: (Concurrency | Hash[Symbol, untyped])?,
+        ?acceleration: (Acceleration | Hash[Symbol, untyped])?,
+        ?email: (Email | Hash[Symbol, untyped])?,
         ?max_concurrent_extractions: Integer?,
+        ?max_archive_depth: Integer,
         ?output_format: String?,
         ?result_format: String?,
         ?cache_namespace: String?,

data/spec/binding/render_spec.rb

@@ -0,0 +1,91 @@
+# Hand-written binding-specific edge case tests for PDF rendering.
+# Happy-path render tests are auto-generated from fixtures in e2e/.
+# These tests cover error handling, validation, and lifecycle patterns
+# that vary per language and can't be generated uniformly.
+
+# frozen_string_literal: true
+
+require 'spec_helper'
+
+RSpec.describe 'PDF Rendering' do
+  it 'exposes rendering methods' do
+    expect(Kreuzberg).to respond_to(:render_pdf_page)
+    expect(Kreuzberg).to respond_to(:render_pdf_pages_iter)
+  end
+
+  describe '.render_pdf_page' do
+    it 'raises an error for a nonexistent file' do
+      expect do
+        Kreuzberg.render_pdf_page('/nonexistent/path/to/document.pdf', 0)
+      end.to raise_error(Kreuzberg::Errors::IOError)
+    end
+
+    it 'raises an error for an out-of-bounds page index' do
+      pdf_path = test_document_path('pdf/tiny.pdf')
+      skip 'Test PDF not available' unless File.exist?(pdf_path)
+
+      expect do
+        Kreuzberg.render_pdf_page(pdf_path, 9999)
+      end.to raise_error(StandardError)
+    end
+  end
+
+  describe '.render_pdf_page with negative index' do
+    it 'raises ArgumentError for a negative page index' do
+      pdf_path = test_document_path('pdf/tiny.pdf')
+      skip 'Test PDF not available' unless File.exist?(pdf_path)
+
+      expect do
+        Kreuzberg.render_pdf_page(pdf_path, -1)
+      end.to raise_error(ArgumentError)
+    end
+  end
+
+  describe '.render_pdf_pages_iter' do
+    it 'raises an error for a nonexistent file' do
+      expect do
+        Kreuzberg.render_pdf_pages_iter('/nonexistent/path/to/document.pdf') { |_, _| nil }
+      end.to raise_error(Kreuzberg::Errors::IOError)
+    end
+  end
+
+  describe '.render_pdf_page with empty path' do
+    it 'raises an error for an empty path' do
+      expect do
+        Kreuzberg.render_pdf_page('', 0)
+      end.to raise_error(StandardError)
+    end
+  end
+
+  describe '.render_pdf_pages_iter cleanup' do
+    it 'handles iterator cleanup without fully consuming' do
+      pdf_path = test_document_path('pdf/tiny.pdf')
+      skip 'Test PDF not available' unless File.exist?(pdf_path)
+
+      # Iterate but stop immediately — no crash
+      Kreuzberg.render_pdf_pages_iter(pdf_path) do |_page_index, _png_data|
+        break
+      end
+    end
+  end
+
+  describe '.render_pdf_pages_iter early termination' do
+    it 'returns valid PNG for the first page then stops' do
+      pdf_path = test_document_path('pdf/tiny.pdf')
+      skip 'Test PDF not available' unless File.exist?(pdf_path)
+
+      first_png = nil
+      Kreuzberg.render_pdf_pages_iter(pdf_path) do |page_index, png_data|
+        expect(page_index).to eq(0)
+        expect(png_data).to be_a(String)
+        expect(png_data.bytesize).to be > 8
+        # PNG magic bytes
+        expect(png_data.bytes[0..3]).to eq([0x89, 0x50, 0x4E, 0x47])
+        first_png = png_data
+        break
+      end
+
+      expect(first_png).not_to be_nil
+    end
+  end
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: kreuzberg
 version: !ruby/object:Gem::Version
-  version: 4.6.1
+  version: 4.6.3
 platform: aarch64-linux
 authors:
 - Na'aman Hirschfeld
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-03-25 00:00:00.000000000 Z
+date: 2026-03-27 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -220,6 +220,7 @@ files:
 - spec/binding/plugins/ocr_backend_spec.rb
 - spec/binding/plugins/postprocessor_spec.rb
 - spec/binding/plugins/validator_spec.rb
+- spec/binding/render_spec.rb
 - spec/binding/tables_spec.rb
 - spec/serialization_spec.rb
 - spec/smoke/package_spec.rb