kreuzberg 4.0.0.rc2 → 4.0.0

data/spec/binding/cli_spec.rb

@@ -1,55 +1,55 @@
-# frozen_string_literal: true
-
-RSpec.describe Kreuzberg::CLI do
-  describe '.extract' do
-    it 'extracts content from a file' do
-      path = test_document_path('documents/simple.odt')
-      output = described_class.extract(path)
-
-      expect(output).to be_a(String)
-      expect(output).not_to be_empty
-    end
-
-    it 'accepts output format option' do
-      path = test_document_path('documents/simple.odt')
-      output = described_class.extract(path, output: 'json')
-
-      expect(output).to be_a(String)
-      expect(output).not_to be_empty
-    end
-
-    it 'accepts OCR option' do
-      path = test_document_path('pdfs/100_g_networking_technology_overview_slides_toronto_august_2016.pdf')
-      output = described_class.extract(path, ocr: false)
-
-      expect(output).to be_a(String)
-      expect(output).not_to be_empty
-    end
-  end
-
-  describe '.detect' do
-    it 'detects MIME type' do
-      path = test_document_path('documents/simple.odt')
-      mime_type = described_class.detect(path)
-
-      expect(mime_type).to be_a(String)
-      expect(mime_type).not_to be_empty
-    end
-  end
-
-  describe '.version' do
-    it 'returns version string' do
-      version = described_class.version
-      expect(version).to be_a(String)
-      expect(version).to match(/\d+\.\d+/)
-    end
-  end
-
-  describe '.help' do
-    it 'returns help text' do
-      help_text = described_class.help
-      expect(help_text).to be_a(String)
-      expect(help_text).to include('kreuzberg')
-    end
-  end
-end
+# frozen_string_literal: true
+
+RSpec.describe Kreuzberg::CLI do
+  describe '.extract' do
+    it 'extracts content from a file' do
+      path = test_document_path('documents/simple.odt')
+      output = described_class.extract(path)
+
+      expect(output).to be_a(String)
+      expect(output).not_to be_empty
+    end
+
+    it 'accepts output format option' do
+      path = test_document_path('documents/simple.odt')
+      output = described_class.extract(path, output: 'json')
+
+      expect(output).to be_a(String)
+      expect(output).not_to be_empty
+    end
+
+    it 'accepts OCR option' do
+      path = test_document_path('pdfs/100_g_networking_technology_overview_slides_toronto_august_2016.pdf')
+      output = described_class.extract(path, ocr: false)
+
+      expect(output).to be_a(String)
+      expect(output).not_to be_empty
+    end
+  end
+
+  describe '.detect' do
+    it 'detects MIME type' do
+      path = test_document_path('documents/simple.odt')
+      mime_type = described_class.detect(path)
+
+      expect(mime_type).to be_a(String)
+      expect(mime_type).not_to be_empty
+    end
+  end
+
+  describe '.version' do
+    it 'returns version string' do
+      version = described_class.version
+      expect(version).to be_a(String)
+      expect(version).to match(/\d+\.\d+/)
+    end
+  end
+
+  describe '.help' do
+    it 'returns help text' do
+      help_text = described_class.help
+      expect(help_text).to be_a(String)
+      expect(help_text).to include('kreuzberg')
+    end
+  end
+end

data/spec/binding/config_result_spec.rb

@@ -0,0 +1,377 @@
+# frozen_string_literal: true
+
+require 'spec_helper'
+
+RSpec.describe 'Phase 1 FFI Config and Result Methods' do
+  let(:fixture_dir) { File.expand_path('../fixtures', __dir__) }
+
+  describe 'Kreuzberg::Config::Extraction' do
+    describe '#to_json' do
+      it 'serializes a basic config to JSON' do
+        config = Kreuzberg::Config::Extraction.new(use_cache: true)
+        json = config.to_json
+        expect(json).to be_a(String)
+        parsed = JSON.parse(json)
+        expect(parsed['use_cache']).to be true
+      end
+
+      it 'serializes complex nested config to JSON' do
+        ocr = Kreuzberg::Config::OCR.new(backend: 'tesseract', language: 'deu')
+        chunking = Kreuzberg::Config::Chunking.new(max_chars: 500, max_overlap: 50)
+        config = Kreuzberg::Config::Extraction.new(
+          use_cache: true,
+          force_ocr: false,
+          ocr: ocr,
+          chunking: chunking
+        )
+
+        json = config.to_json
+        parsed = JSON.parse(json)
+
+        expect(parsed['use_cache']).to be true
+        expect(parsed['force_ocr']).to be false
+        expect(parsed['ocr']['backend']).to eq('tesseract')
+        expect(parsed['ocr']['language']).to eq('deu')
+        expect(parsed['chunking']['max_chars']).to eq(500)
+        expect(parsed['chunking']['max_overlap']).to eq(50)
+      end
+
+      it 'handles minimal config' do
+        config = Kreuzberg::Config::Extraction.new
+        json = config.to_json
+        expect(json).to be_a(String)
+        parsed = JSON.parse(json)
+        expect(parsed).to be_a(Hash)
+      end
+    end
+
+    describe '#get_field' do
+      let(:config) do
+        ocr = Kreuzberg::Config::OCR.new(backend: 'tesseract', language: 'eng')
+        Kreuzberg::Config::Extraction.new(
+          use_cache: true,
+          force_ocr: false,
+          ocr: ocr
+        )
+      end
+
+      it 'gets a top-level field' do
+        value = config.get_field('use_cache')
+        expect(value).to be true
+      end
+
+      it 'gets a nested field with dot notation' do
+        value = config.get_field('ocr.backend')
+        expect(value).to eq('tesseract')
+      end
+
+      it 'gets another nested field' do
+        value = config.get_field('ocr.language')
+        expect(value).to eq('eng')
+      end
+
+      it 'returns nil for non-existent field' do
+        value = config.get_field('nonexistent')
+        expect(value).to be_nil
+      end
+
+      it 'returns nil for non-existent nested field' do
+        value = config.get_field('ocr.nonexistent')
+        expect(value).to be_nil
+      end
+
+      it 'supports symbol field names' do
+        value = config.get_field(:use_cache)
+        expect(value).to be true
+      end
+
+      it 'gets boolean fields correctly' do
+        value = config.get_field('force_ocr')
+        expect(value).to be false
+      end
+    end
+
+    describe '#merge' do
+      let(:base_config) do
+        Kreuzberg::Config::Extraction.new(
+          use_cache: true,
+          force_ocr: false,
+          enable_quality_processing: false
+        )
+      end
+
+      let(:override_config) do
+        Kreuzberg::Config::Extraction.new(
+          force_ocr: true,
+          enable_quality_processing: true
+        )
+      end
+
+      it 'merges two configs without modifying original' do
+        merged = base_config.merge(override_config)
+
+        expect(base_config.use_cache).to be true
+        expect(base_config.force_ocr).to be false
+        expect(base_config.enable_quality_processing).to be false
+
+        expect(merged.use_cache).to be true
+        expect(merged.force_ocr).to be true
+        expect(merged.enable_quality_processing).to be true
+      end
+
+      it 'returns a new Extraction instance' do
+        merged = base_config.merge(override_config)
+        expect(merged).to be_a(Kreuzberg::Config::Extraction)
+        expect(merged).not_to be(base_config)
+      end
+
+      it 'merges with a Hash' do
+        merged = base_config.merge(force_ocr: true)
+
+        expect(merged.use_cache).to be true
+        expect(merged.force_ocr).to be true
+      end
+
+      it 'handles nested config merging' do
+        ocr1 = Kreuzberg::Config::OCR.new(backend: 'tesseract', language: 'eng')
+        base = Kreuzberg::Config::Extraction.new(ocr: ocr1, use_cache: true)
+
+        ocr2 = Kreuzberg::Config::OCR.new(backend: 'easyocr')
+        override = Kreuzberg::Config::Extraction.new(ocr: ocr2)
+
+        merged = base.merge(override)
+
+        expect(merged.use_cache).to be true
+        expect(merged.ocr.backend).to eq('easyocr')
+      end
+    end
+
+    describe '#merge!' do
+      let(:base_config) do
+        Kreuzberg::Config::Extraction.new(
+          use_cache: true,
+          force_ocr: false
+        )
+      end
+
+      let(:override_config) do
+        Kreuzberg::Config::Extraction.new(
+          force_ocr: true
+        )
+      end
+
+      it 'modifies the original config in place' do
+        original_object_id = base_config.object_id
+        result = base_config.merge!(override_config)
+
+        expect(result.object_id).to eq(original_object_id)
+        expect(base_config.use_cache).to be true
+        expect(base_config.force_ocr).to be true
+      end
+
+      it 'returns self' do
+        result = base_config.merge!(override_config)
+        expect(result).to be(base_config)
+      end
+
+      it 'works with Hash argument' do
+        base_config.merge!(force_ocr: true, use_cache: false)
+
+        expect(base_config.force_ocr).to be true
+        expect(base_config.use_cache).to be false
+      end
+
+      it 'updates all fields correctly' do
+        ocr = Kreuzberg::Config::OCR.new(backend: 'easyocr')
+        base_config.merge!(ocr: ocr, enable_quality_processing: true)
+
+        expect(base_config.ocr.backend).to eq('easyocr')
+        expect(base_config.enable_quality_processing).to be true
+      end
+    end
+  end
+
+  describe 'Kreuzberg::Result' do
+    let(:sample_result_hash) do
+      {
+        'content' => 'Sample document content',
+        'mime_type' => 'application/pdf',
+        'metadata_json' => {
+          'title' => 'Test Document',
+          'language' => 'en',
+          'pages' => {
+            'total_count' => 10,
+            'unit_type' => 'Page'
+          },
+          'format' => {
+            'name' => 'PDF',
+            'pages' => 10
+          }
+        }.to_json,
+        'tables' => [],
+        'detected_languages' => %w[en de],
+        'chunks' => [
+          {
+            'content' => 'Chunk 1',
+            'byte_start' => 0,
+            'byte_end' => 7,
+            'token_count' => 2,
+            'chunk_index' => 0,
+            'total_chunks' => 2,
+            'first_page' => 1,
+            'last_page' => 1,
+            'embedding' => nil
+          },
+          {
+            'content' => 'Chunk 2',
+            'byte_start' => 8,
+            'byte_end' => 15,
+            'token_count' => 2,
+            'chunk_index' => 1,
+            'total_chunks' => 2,
+            'first_page' => 2,
+            'last_page' => 2,
+            'embedding' => nil
+          }
+        ]
+      }
+    end
+
+    let(:result) { Kreuzberg::Result.new(sample_result_hash) }
+
+    describe '#page_count' do
+      it 'returns the total page count' do
+        expect(result.page_count).to eq(10)
+      end
+
+      it 'returns 0 for result without page info' do
+        minimal_result = Kreuzberg::Result.new(
+          'content' => 'Test',
+          'mime_type' => 'text/plain',
+          'metadata_json' => '{}'
+        )
+        expect(minimal_result.page_count).to eq(0)
+      end
+
+      it 'returns 0 when metadata has no pages info' do
+        result_no_pages = Kreuzberg::Result.new(
+          'content' => 'Test',
+          'mime_type' => 'text/plain',
+          'metadata_json' => '{"title": "Test"}'
+        )
+        expect(result_no_pages.page_count).to eq(0)
+      end
+    end
+
+    describe '#chunk_count' do
+      it 'returns the total number of chunks' do
+        expect(result.chunk_count).to eq(2)
+      end
+
+      it 'returns 0 for result without chunks' do
+        no_chunks_result = Kreuzberg::Result.new(
+          'content' => 'Test',
+          'mime_type' => 'text/plain',
+          'metadata_json' => '{}'
+        )
+        expect(no_chunks_result.chunk_count).to eq(0)
+      end
+
+      it 'returns 0 for empty chunks array' do
+        empty_chunks_result = Kreuzberg::Result.new(
+          'content' => 'Test',
+          'mime_type' => 'text/plain',
+          'metadata_json' => '{}',
+          'chunks' => []
+        )
+        expect(empty_chunks_result.chunk_count).to eq(0)
+      end
+    end
+
+    describe '#detected_language' do
+      it 'returns the primary detected language from metadata' do
+        expect(result.detected_language).to eq('en')
+      end
+
+      it 'returns the first detected language if metadata language is not set' do
+        result_with_detected = Kreuzberg::Result.new(
+          'content' => 'Test',
+          'mime_type' => 'text/plain',
+          'metadata_json' => '{}',
+          'detected_languages' => %w[fr de]
+        )
+        expect(result_with_detected.detected_language).to eq('fr')
+      end
+
+      it 'returns nil when no language is detected' do
+        no_lang_result = Kreuzberg::Result.new(
+          'content' => 'Test',
+          'mime_type' => 'text/plain',
+          'metadata_json' => '{}'
+        )
+        expect(no_lang_result.detected_language).to be_nil
+      end
+
+      it 'returns nil for empty detected languages array' do
+        empty_langs_result = Kreuzberg::Result.new(
+          'content' => 'Test',
+          'mime_type' => 'text/plain',
+          'metadata_json' => '{}',
+          'detected_languages' => []
+        )
+        expect(empty_langs_result.detected_language).to be_nil
+      end
+    end
+
+    describe '#metadata_field' do
+      it 'gets a top-level metadata field' do
+        value = result.metadata_field('title')
+        expect(value).to eq('Test Document')
+      end
+
+      it 'gets a nested metadata field with dot notation' do
+        value = result.metadata_field('pages.total_count')
+        expect(value).to eq(10)
+      end
+
+      it 'gets another nested field' do
+        value = result.metadata_field('format.name')
+        expect(value).to eq('PDF')
+      end
+
+      it 'returns nil for non-existent field' do
+        value = result.metadata_field('nonexistent')
+        expect(value).to be_nil
+      end
+
+      it 'returns nil for non-existent nested field' do
+        value = result.metadata_field('format.nonexistent')
+        expect(value).to be_nil
+      end
+
+      it 'supports symbol field names' do
+        value = result.metadata_field(:title)
+        expect(value).to eq('Test Document')
+      end
+
+      it 'returns nil when trying to access nested field on non-hash value' do
+        value = result.metadata_field('title.nested')
+        expect(value).to be_nil
+      end
+
+      it 'handles deeply nested fields' do
+        value = result.metadata_field('format.pages')
+        expect(value).to eq(10)
+      end
+
+      it 'returns nil for result without metadata' do
+        no_metadata = Kreuzberg::Result.new(
+          'content' => 'Test',
+          'mime_type' => 'text/plain',
+          'metadata_json' => 'invalid json'
+        )
+        expect(no_metadata.metadata_field('title')).to be_nil
+      end
+    end
+  end
+end