kreuzberg 4.0.0.pre.rc.6 → 4.0.0.rc1

data/sig/kreuzberg.rbs

@@ -3,16 +3,6 @@
 module Kreuzberg
   VERSION: String
 
-  # Error code constants
-  ERROR_CODE_SUCCESS: Integer
-  ERROR_CODE_GENERIC: Integer
-  ERROR_CODE_PANIC: Integer
-  ERROR_CODE_INVALID_ARGUMENT: Integer
-  ERROR_CODE_IO: Integer
-  ERROR_CODE_PARSING: Integer
-  ERROR_CODE_OCR: Integer
-  ERROR_CODE_MISSING_DEPENDENCY: Integer
-
   # Config namespace (defined in lib/kreuzberg/config.rb)
   module Config
     class OCR
@@ -401,10 +391,6 @@ module Kreuzberg
   # Config loading (native method)
   def self._config_from_file_native: (String path) -> Hash[Symbol, untyped]
 
-  # Error introspection (native methods)
-  def self._last_error_code_native: () -> Integer
-  def self._last_panic_context_json_native: () -> String?
-
   # Plugin registration
   def self.register_post_processor: (String name, _PostProcessor processor, ?stage: Symbol?) -> void
   def self.unregister_post_processor: (String name) -> void
@@ -427,63 +413,25 @@ module Kreuzberg
     def extract_text: (String file_path_or_bytes, Hash[Symbol, untyped] config) -> String
   end
 
-  module ErrorContext
-    def self.last_error_code: () -> Integer
-    def self.last_panic_context: () -> Errors::PanicContext?
-    def self.last_panic_context_json: () -> String?
-  end
-
   module Errors
-    # Panic context information from FFI error introspection
-    class PanicContext
-      attr_reader file: String
-      attr_reader line: Integer
-      attr_reader function: String
-      attr_reader message: String
-      attr_reader timestamp_secs: Integer
-
-      def initialize: (
-        file: String,
-        line: Integer,
-        function: String,
-        message: String,
-        timestamp_secs: Integer
-      ) -> void
-      def to_s: () -> String
-      def to_h: () -> Hash[Symbol, String | Integer]
-      def self.from_json: (String) -> PanicContext?
-
-      private
-
-      def self.with_defaults: (Hash[Symbol, untyped] sliced) -> {file: String, line: Integer, function: String, message: String, timestamp_secs: Integer}
-    end
-
     class Error < StandardError
-      attr_reader panic_context: PanicContext?
-      attr_reader error_code: Integer?
-
-      def initialize: (String message, ?panic_context: PanicContext?, ?error_code: Integer?) -> void
     end
 
     class ValidationError < Error
     end
 
     class ParsingError < Error
-      attr_reader context: Hash[untyped, untyped]?
-
-      def initialize: (String message, ?context: Hash[untyped, untyped]?, ?panic_context: PanicContext?, ?error_code: Integer?) -> void
+      def initialize: (String message, ?context: Hash[untyped, untyped]?) -> void
     end
 
     class OCRError < Error
-      attr_reader context: Hash[untyped, untyped]?
-
-      def initialize: (String message, ?context: Hash[untyped, untyped]?, ?panic_context: PanicContext?, ?error_code: Integer?) -> void
+      def initialize: (String message, ?context: Hash[untyped, untyped]?) -> void
     end
 
     class MissingDependencyError < Error
       attr_reader dependency: String?
 
-      def initialize: (String message, ?dependency: String?, ?panic_context: PanicContext?, ?error_code: Integer?) -> void
+      def initialize: (String message, ?dependency: String?) -> void
     end
 
     class IOError < Error

data/spec/binding/cli_proxy_spec.rb

@@ -3,7 +3,8 @@
 RSpec.describe Kreuzberg::CLIProxy do
   describe '.find_cli_binary' do
     context 'when binary exists' do
-      it 'finds the binary in search paths' do
+      it 'finds the binary in search paths', :skip do
+        # Skip in CI/test environments where binary might not be built
         binary = described_class.find_cli_binary
         expect(binary).to be_a(Pathname)
         expect(binary.file?).to be true
@@ -24,8 +25,9 @@ RSpec.describe Kreuzberg::CLIProxy do
   end
 
   describe '.call' do
-    context 'when binary is available' do
+    context 'when binary is available', :skip do
       it 'executes CLI command successfully' do
+        # Skip in environments without built binary
         output = described_class.call(['--version'])
         expect(output).to be_a(String)
         expect(output).not_to be_empty

data/spec/binding/cli_spec.rb

@@ -1,35 +1,34 @@
 # frozen_string_literal: true
 
 RSpec.describe Kreuzberg::CLI do
-  describe '.extract' do
+  describe '.extract', :skip do
     it 'extracts content from a file' do
-      path = test_document_path('documents/simple.odt')
+      # Skip in environments without CLI binary
+      path = create_test_file('CLI test content')
       output = described_class.extract(path)
 
       expect(output).to be_a(String)
-      expect(output).not_to be_empty
+      expect(output).to include('CLI test content')
     end
 
     it 'accepts output format option' do
-      path = test_document_path('documents/simple.odt')
+      path = create_test_file('JSON output test')
       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)
+      path = create_test_file('OCR test')
+      output = described_class.extract(path, ocr: true)
 
       expect(output).to be_a(String)
-      expect(output).not_to be_empty
     end
   end
 
-  describe '.detect' do
+  describe '.detect', :skip do
     it 'detects MIME type' do
-      path = test_document_path('documents/simple.odt')
+      path = create_test_file('MIME detection test')
       mime_type = described_class.detect(path)
 
       expect(mime_type).to be_a(String)
@@ -37,7 +36,7 @@ RSpec.describe Kreuzberg::CLI do
     end
   end
 
-  describe '.version' do
+  describe '.version', :skip do
     it 'returns version string' do
       version = described_class.version
       expect(version).to be_a(String)
@@ -45,7 +44,7 @@ RSpec.describe Kreuzberg::CLI do
     end
   end
 
-  describe '.help' do
+  describe '.help', :skip do
     it 'returns help text' do
       help_text = described_class.help
       expect(help_text).to be_a(String)

data/spec/examples.txt

@@ -0,0 +1,104 @@
+example_id                                                                         | status | run_time        |
+---------------------------------------------------------------------------------- | ------ | --------------- |
+./spec/binding/plugins/ocr_backend_spec.rb[1:1:1]                                  | failed | 0.00173 seconds |
+./spec/binding/plugins/ocr_backend_spec.rb[1:1:2]                                  | failed | 0.0018 seconds  |
+./spec/binding/plugins/ocr_backend_spec.rb[1:2:1]                                  | failed | 0.00192 seconds |
+./spec/binding/plugins/ocr_backend_spec.rb[1:2:2]                                  | failed | 0.00581 seconds |
+./spec/binding/plugins/ocr_backend_spec.rb[1:3:1]                                  | failed | 0.00184 seconds |
+./spec/binding/plugins/ocr_backend_spec.rb[1:4:1]                                  | passed | 0.00088 seconds |
+./spec/binding/plugins/ocr_backend_spec.rb[1:4:2]                                  | passed | 0.00045 seconds |
+./spec/binding/plugins/ocr_backend_spec.rb[1:5:1]                                  | passed | 0.00007 seconds |
+./spec/binding/plugins/ocr_backend_spec.rb[1:5:2]                                  | passed | 0.00052 seconds |
+./spec/binding/plugins/ocr_backend_spec.rb[1:6:1:1]                                | passed | 0.00012 seconds |
+./spec/binding/plugins/ocr_backend_spec.rb[1:6:1:2]                                | passed | 0.00079 seconds |
+./spec/binding/plugins/ocr_backend_spec.rb[1:6:2:1]                                | passed | 0.00004 seconds |
+./spec/binding/plugins/ocr_backend_spec.rb[1:6:2:2]                                | passed | 0.00029 seconds |
+./spec/binding/plugins/postprocessor_spec.rb[1:1:1]                                | failed | 0.00139 seconds |
+./spec/binding/plugins/postprocessor_spec.rb[1:1:2]                                | failed | 0.00153 seconds |
+./spec/binding/plugins/postprocessor_spec.rb[1:1:3]                                | failed | 0.0014 seconds  |
+./spec/binding/plugins/postprocessor_spec.rb[1:2:1]                                | failed | 0.00182 seconds |
+./spec/binding/plugins/postprocessor_spec.rb[1:2:2]                                | failed | 0.00209 seconds |
+./spec/binding/plugins/postprocessor_spec.rb[1:3:1]                                | failed | 0.00165 seconds |
+./spec/binding/plugins/postprocessor_spec.rb[1:4:1]                                | failed | 0.00142 seconds |
+./spec/binding/plugins/postprocessor_spec.rb[1:4:2]                                | failed | 0.00148 seconds |
+./spec/binding/plugins/postprocessor_spec.rb[1:5:1]                                | failed | 0.00148 seconds |
+./spec/binding/plugins/postprocessor_spec.rb[1:6:1]                                | passed | 0.0001 seconds  |
+./spec/binding/plugins/postprocessor_spec.rb[1:6:2]                                | passed | 0.00011 seconds |
+./spec/binding/plugins/postprocessor_spec.rb[1:7:1]                                | passed | 0.00003 seconds |
+./spec/binding/plugins/postprocessor_spec.rb[1:7:2]                                | passed | 0.00002 seconds |
+./spec/binding/plugins/postprocessor_spec.rb[1:7:3]                                | passed | 0.00003 seconds |
+./spec/binding/plugins/postprocessor_spec.rb[1:7:4]                                | passed | 0.00006 seconds |
+./spec/binding/plugins/validator_spec.rb[1:1:1]                                    | failed | 0.00154 seconds |
+./spec/binding/plugins/validator_spec.rb[1:1:2]                                    | failed | 0.00171 seconds |
+./spec/binding/plugins/validator_spec.rb[1:1:3]                                    | passed | 0.00099 seconds |
+./spec/binding/plugins/validator_spec.rb[1:2:1]                                    | failed | 0.00186 seconds |
+./spec/binding/plugins/validator_spec.rb[1:2:2]                                    | failed | 0.0016 seconds  |
+./spec/binding/plugins/validator_spec.rb[1:3:1]                                    | failed | 0.00182 seconds |
+./spec/binding/plugins/validator_spec.rb[1:3:2]                                    | failed | 0.0128 seconds  |
+./spec/binding/plugins/validator_spec.rb[1:4:1]                                    | failed | 0.00156 seconds |
+./spec/binding/plugins/validator_spec.rb[1:4:2]                                    | passed | 0.0001 seconds  |
+./spec/binding/plugins/validator_spec.rb[1:5:1]                                    | failed | 0.00445 seconds |
+./spec/binding/plugins/validator_spec.rb[1:5:2]                                    | failed | 0.00198 seconds |
+./spec/binding/plugins/validator_spec.rb[1:6:1]                                    | failed | 0.00179 seconds |
+./spec/binding/plugins/validator_spec.rb[1:7:1]                                    | passed | 0.00068 seconds |
+./spec/binding/plugins/validator_spec.rb[1:7:2]                                    | passed | 0.00088 seconds |
+./spec/binding/plugins/validator_spec.rb[1:7:3]                                    | passed | 0.00045 seconds |
+./spec/binding/plugins/validator_spec.rb[1:7:4]                                    | passed | 0.00004 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/email_spec.rb[1:1]       | passed | 0.01048 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/html_spec.rb[1:1]        | passed | 1.95 seconds    |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/html_spec.rb[1:2]        | passed | 0.00031 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/image_spec.rb[1:1]       | passed | 0.0027 seconds  |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/ocr_spec.rb[1:1]         | passed | 0.04721 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/ocr_spec.rb[1:2]         | passed | 0.04402 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/ocr_spec.rb[1:3]         | passed | 3.41 seconds    |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/ocr_spec.rb[1:4]         | passed | 0.34493 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/ocr_spec.rb[1:5]         | passed | 0.33223 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:1]      | passed | 2.74 seconds    |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:2]      | passed | 0.00021 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:3]      | passed | 0.00035 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:4]      | passed | 0.00021 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:5]      | passed | 0.0003 seconds  |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:6]      | passed | 0.00027 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:7]      | passed | 0.00023 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:8]      | passed | 0.00016 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:9]      | passed | 2.65 seconds    |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:10]     | passed | 0.0003 seconds  |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:11]     | passed | 0.0002 seconds  |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:12]     | passed | 0.00984 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:13]     | passed | 0.00096 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:14]     | passed | 0.00115 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:15]     | passed | 0.00038 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/office_spec.rb[1:16]     | passed | 0.00448 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/pdf_spec.rb[1:1]         | passed | 0.99668 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/pdf_spec.rb[1:2]         | passed | 4.11 seconds    |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/pdf_spec.rb[1:3]         | passed | 0.00451 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/pdf_spec.rb[1:4]         | passed | 0.07588 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/pdf_spec.rb[1:5]         | passed | 0.00339 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/pdf_spec.rb[1:6]         | passed | 0.00238 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/pdf_spec.rb[1:7]         | passed | 0.24683 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/pdf_spec.rb[1:8]         | passed | 0.07999 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/pdf_spec.rb[1:9]         | passed | 0.01214 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/pdf_spec.rb[1:10]        | passed | 0.00095 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/pdf_spec.rb[1:11]        | passed | 0.03728 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/pdf_spec.rb[1:12]        | passed | 0.01741 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/pdf_spec.rb[1:13]        | passed | 0.0036 seconds  |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/pdf_spec.rb[1:14]        | passed | 0.89424 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/plugin_apis_spec.rb[1:1] | passed | 0.00228 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/plugin_apis_spec.rb[1:2] | passed | 0.0012 seconds  |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/plugin_apis_spec.rb[2:1] | passed | 0.0008 seconds  |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/plugin_apis_spec.rb[2:2] | passed | 0.00119 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/plugin_apis_spec.rb[2:3] | passed | 0.0013 seconds  |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/plugin_apis_spec.rb[3:1] | passed | 0.00184 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/plugin_apis_spec.rb[3:2] | passed | 0.00053 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/plugin_apis_spec.rb[3:3] | passed | 0.00004 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/plugin_apis_spec.rb[4:1] | passed | 0.00049 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/plugin_apis_spec.rb[4:2] | passed | 0.00006 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/plugin_apis_spec.rb[4:3] | passed | 0.00005 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/plugin_apis_spec.rb[5:1] | passed | 0.00007 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/plugin_apis_spec.rb[5:2] | passed | 0.00011 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/plugin_apis_spec.rb[6:1] | passed | 0.00003 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/plugin_apis_spec.rb[6:2] | passed | 0.00002 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/structured_spec.rb[1:1]  | passed | 0.00101 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/structured_spec.rb[1:2]  | passed | 0.00041 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/structured_spec.rb[1:3]  | passed | 0.00035 seconds |
+/Users/naamanhirschfeld/workspace/kreuzberg/e2e/ruby/spec/xml_spec.rb[1:1]         | passed | 0.00078 seconds |

data/spec/fixtures/config.yaml

@@ -1,3 +1,4 @@
+# Test configuration file for Kreuzberg Ruby bindings
 
 use_cache: false
 enable_quality_processing: true

data/spec/spec_helper.rb

@@ -30,7 +30,7 @@ RSpec.configure do |config|
 
     def test_document_path(relative_path)
       # Go up from packages/ruby/spec to project root, then into test_documents
-      File.expand_path(File.join(__dir__, '..', '..', '..', 'test_documents', relative_path))
+      File.join(__dir__, '..', '..', '..', 'test_documents', relative_path)
     end
 
     def create_test_file(content, filename: 'test.txt')

data/vendor/kreuzberg/Cargo.toml

@@ -1,62 +1,41 @@
 [package]
 name = "kreuzberg"
-version.workspace = true
+version = "4.0.0-rc.1"
 edition.workspace = true
 rust-version.workspace = true
-authors.workspace = true
+authors = ["Na'aman Hirschfeld <nhirschfeld@gmail.com>"]
 description = "High-performance document intelligence library for Rust. Extract text, metadata, and structured data from PDFs, Office documents, images, and 50+ formats with async/sync APIs."
-license.workspace = true
-repository = "https://github.com/kreuzberg-dev/kreuzberg"
+license = "MIT"
+repository = "https://github.com/Goldziher/kreuzberg"
 homepage = "https://kreuzberg.dev"
 documentation = "https://docs.rs/kreuzberg"
-keywords = ["document", "extraction", "pdf", "ocr", "parser"]
-categories = ["parser-implementations", "text-processing"]
+keywords = ["document", "extraction", "pdf", "ocr", "parsing"]
+categories = ["parsing", "text-processing", "asynchronous", "data-structures"]
 readme = "README.md"
 
 [lib]
 crate-type = ["rlib"]
 
 [features]
-default = ["tokio-runtime"]
+default = []
 
-tokio-runtime = []
 profiling = ["dep:pprof"]
 
 # Format extractors
 pdf = ["dep:pdfium-render", "dep:lopdf", "dep:image"]
 excel = ["dep:calamine", "dep:polars"]
-office = [
-    "dep:roxmltree",
-    "dep:zip",
-    "dep:docx-lite",
-    "dep:quick-xml",
-    "dep:pulldown-cmark",
-    "dep:biblatex",
-    "dep:org",
-    "dep:rtf-parser",
-    "dep:rst_parser",
-    "dep:fb2",
-    "dep:typst-syntax",
-    "html",  # EPUB needs HTML parsing (zip + roxmltree + html-to-markdown-rs)
-]
+office = ["dep:roxmltree", "dep:zip", "dep:docx-lite"]
 email = ["dep:mail-parser", "dep:msg_parser"]
-html = ["dep:html-to-markdown-rs"]
+html = ["dep:html-to-markdown-rs", "dep:html-escape", "dep:scraper"]
 xml = ["dep:quick-xml", "dep:roxmltree"]
 archives = ["dep:zip", "dep:tar", "dep:sevenz-rust"]
 
 # Processing features
-ocr = [
-    "dep:kreuzberg-tesseract",
-    "dep:image",
-    "dep:fast_image_resize",
-    "dep:ndarray",
-    "dep:kamadak-exif",
-    "html",
-]
+ocr = ["dep:kreuzberg-tesseract", "dep:image", "dep:fast_image_resize", "dep:ndarray", "dep:kamadak-exif", "html"]
 language-detection = ["dep:whatlang"]
 chunking = ["dep:text-splitter"]
 embeddings = ["dep:fastembed", "chunking"]
-stopwords = [] # Stopwords for keyword extraction and token reduction
+stopwords = []  # Stopwords for keyword extraction and token reduction
 quality = ["dep:unicode-normalization", "dep:chardetng", "dep:encoding_rs", "stopwords"]
 
 # Keyword extraction (requires stopwords)
@@ -68,59 +47,48 @@ keywords = ["keywords-yake", "keywords-rake"]
 api = ["dep:axum", "dep:tower", "dep:tower-http"]
 mcp = ["dep:rmcp"]
 
-# Observability features
-otel = ["dep:opentelemetry", "dep:opentelemetry_sdk", "dep:tracing-opentelemetry"]
-
 # Convenience bundles
 full = [
-    "pdf",
-    "excel",
-    "office",
-    "email",
-    "html",
-    "xml",
-    "archives",
-    "ocr",
-    "language-detection",
-    "chunking",
-    "quality",
-    "keywords",
+    "pdf", "excel", "office", "email", "html", "xml", "archives",
+    "ocr", "language-detection", "chunking", "quality", "keywords"
 ]
 server = ["pdf", "excel", "html", "ocr", "api", "mcp"]
 cli = ["pdf", "excel", "office", "html", "ocr", "language-detection", "chunking", "quality"]
 
 [build-dependencies]
-tracing = { workspace = true }
+tracing = "0.1"
 
 [dependencies]
 # Core dependencies (always included)
-ahash = { workspace = true }
-async-trait = { workspace = true }
-base64 = { workspace = true }
-base64-simd = "0.8"
-hex = { workspace = true }
+ahash = "0.8.12"
+async-trait = "0.1.89"
+base64 = "0.22.1"
+hex = "0.4.3"
 lazy_static = "1.5.0"
-libc = { workspace = true }
+libc = "0.2"
 memchr = "2.7.6"
-num_cpus = { workspace = true }
+num_cpus = "1.17.0"
 once_cell = "1.21.3"
 paste = "1.0"
 rayon = "1.11.0"
 regex = "1.12.2"
-serde = { workspace = true }
-serde_json = { workspace = true }
+serde = { version = "1.0.228", features = ["derive"] }
+serde_json = "1.0.145"
 serde_yaml_ng = "0.10.0"
 toml = "0.9.8"
 mime_guess = "2.0"
 rmp-serde = "1.3"
-thiserror = { workspace = true }
-tokio = { workspace = true }
-uuid = { version = "1.19.0", features = ["v4"] }
+thiserror = "2.0.17"
+tokio = { version = "1.48.0", features = ["process", "fs", "rt", "rt-multi-thread", "macros", "time", "sync", "io-util"] }
+uuid = { version = "1.18.1", features = ["v4"] }
 indexmap = "2.12.1"
-tracing = { workspace = true }
-reqwest = { workspace = true, default-features = false, features = ["json", "rustls-tls"] }
+tracing = "0.1"
+pprof = { version = "0.15.0", features = ["flamegraph"], optional = true }
+reqwest = { version = "0.12", default-features = false, features = ["json", "rustls-tls"] }
+base64-simd = "0.8"
+
 # Format extractors (optional)
-pdfium-render = { version = "0.8.37", features = ["thread_safe", "image"], optional = true }
+pdfium-render = { version = "0.8.36", features = ["thread_safe", "image"], optional = true }
 lopdf = { version = "0.38.0", optional = true }
 calamine = { version = "0.32.0", features = ["dates"], optional = true }
 polars = { version = "0.52.0", default-features = false, features = ["ipc"], optional = true }
@@ -128,77 +96,39 @@ roxmltree = { version = "0.21.1", optional = true }
 zip = { version = "6.0.0", optional = true }
 mail-parser = { version = "0.11.1", optional = true }
 msg_parser = { version = "0.1.1", optional = true }
-html-to-markdown-rs = { version = "2.12.0", features = ["inline-images"], optional = true }
-quick-xml = { version = "0.38.4", features = ["serialize"], optional = true }
+html-to-markdown-rs = { version = "2.9.1", features = ["inline-images"], optional = true }
+html-escape = { version = "0.2.13", optional = true }
+scraper = { version = "0.24.0", optional = true }
+quick-xml = { version = "0.38.4", optional = true }
 tar = { version = "0.4.44", optional = true }
 sevenz-rust = { version = "0.6.1", optional = true }
 docx-lite = { version = "0.2.0", optional = true }
 
-pulldown-cmark = { version = "0.13", optional = true }
-biblatex = { version = "0.11", optional = true }
-org = { version = "0.3", optional = true }
-rtf-parser = { version = "0.4", optional = true }
-rst_parser = { version = "0.4", optional = true }
-fb2 = { version = "0.4", optional = true }
-typst-syntax = { version = "0.14", optional = true }
-
 # Processing features (optional)
-kreuzberg-tesseract = { version = "4.0.0-rc.6", optional = true }
-image = { workspace = true, default-features = false, features = [
-    "png",
-    "jpeg",
-    "webp",
-    "bmp",
-    "tiff",
-    "gif",
-    "rayon",
-], optional = true }
-fast_image_resize = { version = "5.4.0", optional = true }
+kreuzberg-tesseract = { version = "1.0.0-rc.1", optional = true }
+image = { version = "0.25.9", default-features = false, features = ["png", "jpeg", "webp", "bmp", "tiff", "gif", "rayon"], optional = true }
+fast_image_resize = { version = "5.1.4", optional = true }
 ndarray = { version = "0.17.1", optional = true }
 kamadak-exif = { version = "0.6.1", optional = true }
 whatlang = { version = "0.18.0", optional = true }
 text-splitter = { version = "0.28.0", features = ["markdown"], optional = true }
-fastembed = { version = "5.4", default-features = false, features = [
-    "hf-hub-rustls-tls",
-    "ort-download-binaries",
-], optional = true }
+fastembed = { version = "*", git = "https://github.com/kreuzberg-dev/fastembed-rs", default-features = false, features = ["hf-hub", "hf-hub-rustls-tls", "ort-download-binaries"], optional = true }
 unicode-normalization = { version = "0.1.25", optional = true }
 chardetng = { version = "0.1.17", optional = true }
 encoding_rs = { version = "0.8.35", optional = true }
 yake-rust = { version = "1.0.3", optional = true }
 rake = { version = "0.3.6", optional = true }
+
 # Server features (optional)
 axum = { version = "0.8", features = ["macros", "json", "multipart"], optional = true }
 tower = { version = "0.5", optional = true }
 tower-http = { version = "0.6", features = ["cors", "trace", "limit"], optional = true }
-rmcp = { version = "0.11.0", features = [
-    "server",
-    "macros",
-    "base64",
-    "transport-io",
-], optional = true }
-# Observability features (optional)
-opentelemetry = { version = "0.31", features = ["trace"], optional = true }
-opentelemetry_sdk = { version = "0.31", features = ["rt-tokio"], optional = true }
-tracing-opentelemetry = { version = "0.32", optional = true }
+rmcp = { version = "0.9.0", features = ["server", "macros", "base64", "transport-io"], optional = true }
 infer = "0.19.0"
 
 [dev-dependencies]
-tempfile = { workspace = true }
+tempfile = "3.23.0"
 filetime = "0.2"
 tar = "0.4.44"
 zip = "6.0.0"
 serial_test = "3.2.0"
-anyhow = { workspace = true }
-tokio-test = "0.4"
-tracing-subscriber = { version = "0.3", features = ["env-filter"] }
-criterion = { workspace = true }
-image = { workspace = true, default-features = false, features = ["png"] }
-
-[[bench]]
-name = "otel_overhead"
-harness = false
-
-# Only build profiling tooling on non-Windows targets (pprof depends on Unix APIs)
-[target.'cfg(not(target_os = "windows"))'.dependencies]
-pprof = { version = "0.15.0", features = ["flamegraph"], optional = true }

data/vendor/kreuzberg/README.md

@@ -2,7 +2,7 @@
 
 [![Crates.io](https://img.shields.io/crates/v/kreuzberg)](https://crates.io/crates/kreuzberg)
 [![PyPI](https://img.shields.io/pypi/v/kreuzberg)](https://pypi.org/project/kreuzberg/)
-[![npm](https://img.shields.io/npm/v/kreuzberg)](https://www.npmjs.com/package/kreuzberg)
+[![npm](https://img.shields.io/npm/v/@goldziher/kreuzberg)](https://www.npmjs.com/package/@goldziher/kreuzberg)
 [![RubyGems](https://img.shields.io/gem/v/kreuzberg)](https://rubygems.org/gems/kreuzberg)
 [![docs.rs](https://docs.rs/kreuzberg/badge.svg)](https://docs.rs/kreuzberg)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
@@ -13,7 +13,7 @@ High-performance document intelligence library for Rust. Extract text, metadata,
 This is the core Rust library that powers the Python, TypeScript, and Ruby bindings.
 
 > **🚀 Version 4.0.0 Release Candidate**
-> This is a pre-release version. We invite you to test the library and [report any issues](https://github.com/kreuzberg-dev/kreuzberg/issues) you encounter.
+> This is a pre-release version. We invite you to test the library and [report any issues](https://github.com/Goldziher/kreuzberg/issues) you encounter.
 >
 > **Note**: The Rust crate is not currently published to crates.io for this RC. Use git dependencies or language bindings (Python, TypeScript, Ruby) instead.
 

data/vendor/kreuzberg/build.rs

@@ -290,6 +290,7 @@ fn copy_lib_to_package(pdfium_dir: &Path, target: &str) {
         return;
     }
 
+    // Fix install_name on macOS to use @rpath
     if target.contains("darwin") {
         fix_macos_install_name(&src_lib, &runtime_lib_name);
         codesign_if_needed(target, &src_lib);
@@ -298,13 +299,9 @@ fn copy_lib_to_package(pdfium_dir: &Path, target: &str) {
     let crate_dir = PathBuf::from(env::var("CARGO_MANIFEST_DIR").unwrap());
     let workspace_root = crate_dir.parent().unwrap().parent().unwrap();
 
+    // Copy to target directory for CLI binary
     if let Ok(profile) = env::var("PROFILE") {
-        let target_dir = if let Ok(cargo_target) = env::var("TARGET") {
-            workspace_root.join("target").join(cargo_target).join(&profile)
-        } else {
-            workspace_root.join("target").join(&profile)
-        };
-
+        let target_dir = workspace_root.join("target").join(profile);
         if target_dir.exists() {
             copy_lib_if_needed(
                 &src_lib,
@@ -313,18 +310,6 @@ fn copy_lib_to_package(pdfium_dir: &Path, target: &str) {
                 target,
             );
         }
-
-        // Also copy to target/{profile} for Java FFI (Maven expects it here)
-        let simple_target_dir = workspace_root.join("target").join(&profile);
-        if simple_target_dir != target_dir {
-            fs::create_dir_all(&simple_target_dir).ok();
-            copy_lib_if_needed(
-                &src_lib,
-                &simple_target_dir.join(&runtime_lib_name),
-                "Java FFI target directory",
-                target,
-            );
-        }
     }
 
     let python_dest_dir = workspace_root.join("packages").join("python").join("kreuzberg");
@@ -450,6 +435,7 @@ fn copy_dir_all(src: &Path, dst: &Path) -> io::Result<()> {
 fn fix_macos_install_name(lib_path: &Path, lib_name: &str) {
     use std::process::Command;
 
+    // Change install_name from ./libpdfium.dylib to @rpath/libpdfium.dylib
     let new_install_name = format!("@rpath/{}", lib_name);
 
     tracing::debug!("Fixing install_name for {} to {}", lib_path.display(), new_install_name);