kreuzberg 4.0.0.pre.rc.6 → 4.0.0.rc1

data/ext/kreuzberg_rb/native/Cargo.toml

@@ -1,24 +1,17 @@
-# This crate is excluded from the workspace to use a vendored kreuzberg crate for gem packaging
-[workspace]
-
-[patch.crates-io]
-# Patch rb-sys to fix Windows i32/i64 type mismatch in tracking_allocator.rs
-rb-sys = { path = "../../../vendor/rb-sys" }
-
 [package]
 name = "kreuzberg-rb"
-version = "4.0.0-rc.6"
+version = "4.0.0-rc.1"
 edition = "2024"
-rust-version = "1.91"
+rust-version = "1.85"
 authors = ["Na'aman Hirschfeld <nhirschfeld@gmail.com>"]
 license = "MIT"
-repository = "https://github.com/kreuzberg-dev/kreuzberg"
+repository = "https://github.com/Goldziher/kreuzberg"
 homepage = "https://kreuzberg.dev"
 documentation = "https://docs.rs/kreuzberg"
 readme = "README.md"
 description = "Ruby bindings (Magnus) for Kreuzberg - high-performance document intelligence framework"
-keywords = ["ruby", "magnus", "document", "extraction", "bindings"]
-categories = ["api-bindings", "text-processing"]
+keywords = ["document", "extraction", "ocr", "pdf", "ruby"]
+categories = ["api-bindings"]
 
 [lib]
 name = "kreuzberg_rb"
@@ -28,17 +21,16 @@ crate-type = ["cdylib", "rlib"]
 default = []
 
 [dependencies]
-async-trait = "0.1.89"
-kreuzberg = { path = "../../../vendor/kreuzberg", features = ["full", "embeddings"] }
-magnus = { git = "https://github.com/matsadler/magnus", rev = "f6db11769efb517427bf7f121f9c32e18b059b38", features = [
-    "rb-sys",
-] }
-rb-sys = { version = "0.9.119", default-features = false, features = [
-    "stable-api-compiled-fallback",
-] }
-serde_json = "1.0.145"
-tokio = { version = "1.48.0", features = ["rt", "rt-multi-thread", "macros", "sync", "process", "fs", "time", "io-util"] }
-html-to-markdown-rs = { version = "2.11.1", default-features = false }
+async-trait = "0.1"
+kreuzberg = { version = "4.0.0-rc.1", features = ["full", "embeddings"] }
+magnus = { git = "https://github.com/matsadler/magnus", rev = "f6db11769efb517427bf7f121f9c32e18b059b38", features = ["rb-sys"] }
+rb-sys = { version = "0.9.117", default-features = false, features = ["stable-api-compiled-fallback"] }
+serde_json = "1.0"
+tokio = { version = "1.48", features = ["rt", "macros"] }
+html-to-markdown-rs = { version = "2.9.1", default-features = false }
 
 [dev-dependencies]
 pretty_assertions = "1.4"
+
+[patch.crates-io]
+kreuzberg = { path = "../../../../vendor/kreuzberg" }

data/ext/kreuzberg_rb/native/build.rs

@@ -1,12 +1,14 @@
 #[cfg(target_os = "macos")]
 fn main() {
     println!("cargo:rustc-link-arg=-Wl,-undefined,dynamic_lookup");
+    // Set rpath to look for libpdfium.dylib in the same directory as the Ruby extension
     println!("cargo:rustc-link-arg=-Wl,-rpath,@loader_path");
     println!("cargo:rustc-link-arg=-Wl,-rpath,@loader_path/.");
 }
 
 #[cfg(target_os = "linux")]
 fn main() {
+    // Set rpath to look for libpdfium.so in the same directory as the Ruby extension
     println!("cargo:rustc-link-arg=-Wl,-rpath,$ORIGIN");
     println!("cargo:rustc-link-arg=-Wl,-rpath,$ORIGIN/.");
 }

data/ext/kreuzberg_rb/native/include/ieeefp.h

@@ -8,4 +8,4 @@
 #include_next <ieeefp.h>
 #endif
 
-#endif
+#endif  // KREUZBERG_RUBY_IEEFP_H

data/ext/kreuzberg_rb/native/include/msvc_compat/strings.h

@@ -11,4 +11,4 @@
 #define strncasecmp _strnicmp
 #endif
 
-#endif
+#endif /* KREUZBERG_RB_MSVC_COMPAT_STRINGS_H */

data/ext/kreuzberg_rb/native/include/strings.h

@@ -15,6 +15,6 @@
 #ifndef bzero
 #define bzero(ptr, size) memset((ptr), 0, (size))
 #endif
-#endif
+#endif  // _MSC_VER
 
-#endif
+#endif  // KREUZBERG_RUBY_STRINGS_H

data/ext/kreuzberg_rb/native/include/unistd.h

@@ -44,4 +44,4 @@ typedef long ssize_t;
 #include_next <unistd.h>
 #endif
 
-#endif
+#endif  // KREUZBERG_RUBY_UNISTD_H

data/ext/kreuzberg_rb/native/src/lib.rs

@@ -49,33 +49,6 @@ impl Drop for GcGuardedValue {
     }
 }
 
-unsafe extern "C" {
-    fn kreuzberg_last_error_code() -> i32;
-    fn kreuzberg_last_panic_context() -> *const std::ffi::c_char;
-    fn kreuzberg_free_string(s: *mut std::ffi::c_char);
-}
-
-/// Retrieve panic context from FFI if available
-fn get_panic_context() -> Option<String> {
-    unsafe {
-        let ctx_ptr = kreuzberg_last_panic_context();
-        if ctx_ptr.is_null() {
-            return None;
-        }
-
-        let c_str = std::ffi::CStr::from_ptr(ctx_ptr);
-        let context = c_str.to_string_lossy().to_string();
-        kreuzberg_free_string(ctx_ptr as *mut std::ffi::c_char);
-
-        if context.is_empty() { None } else { Some(context) }
-    }
-}
-
-/// Retrieve error code from FFI
-fn get_error_code() -> i32 {
-    unsafe { kreuzberg_last_error_code() }
-}
-
 /// Convert Kreuzberg errors to Ruby exceptions
 fn kreuzberg_error(err: KreuzbergError) -> Error {
     let ruby = Ruby::get().expect("Ruby not initialized");
@@ -282,10 +255,10 @@ fn ruby_value_to_json(value: Value) -> Result<serde_json::Value, Error> {
         return Ok(serde_json::Value::Number(serde_json::Number::from(unsigned)));
     }
 
-    if let Ok(float) = f64::try_convert(value)
-        && let Some(num) = serde_json::Number::from_f64(float)
-    {
-        return Ok(serde_json::Value::Number(num));
+    if let Ok(float) = f64::try_convert(value) {
+        if let Some(num) = serde_json::Number::from_f64(float) {
+            return Ok(serde_json::Value::Number(num));
+        }
     }
 
     if let Ok(sym) = Symbol::try_convert(value) {
@@ -660,10 +633,10 @@ fn parse_keyword_config(ruby: &Ruby, hash: RHash) -> Result<RustKeywordConfig, E
         }
     }
 
-    if let Some(val) = get_kw(ruby, hash, "language")
-        && !val.is_nil()
-    {
-        config.language = Some(symbol_to_string(val)?);
+    if let Some(val) = get_kw(ruby, hash, "language") {
+        if !val.is_nil() {
+            config.language = Some(symbol_to_string(val)?);
+        }
     }
 
     if let Some(val) = get_kw(ruby, hash, "yake_params")
@@ -2024,6 +1997,9 @@ fn register_post_processor(args: &[Value]) -> Result<(), Error> {
             let processor = self.processor.value();
             let result_clone = result.clone();
 
+            // Use block_in_place to avoid GVL deadlocks (same pattern as Python PostProcessor)
+            // See crates/kreuzberg-py/README.md:151-158 for explanation
+            // CRITICAL: spawn_blocking causes GVL deadlocks, must use block_in_place
             let updated_result = tokio::task::block_in_place(|| {
                 let ruby = Ruby::get().expect("Ruby not initialized");
                 let result_hash = extraction_result_to_ruby(&ruby, result_clone.clone()).map_err(|e| {
@@ -2230,6 +2206,9 @@ fn register_validator(args: &[Value]) -> Result<(), Error> {
             let validator = self.validator.value();
             let result_clone = result.clone();
 
+            // Use block_in_place to avoid GVL deadlocks (same pattern as Python Validator)
+            // See crates/kreuzberg-py/README.md:151-158 for explanation
+            // CRITICAL: spawn_blocking causes GVL deadlocks, must use block_in_place
             tokio::task::block_in_place(|| {
                 let ruby = Ruby::get().expect("Ruby not initialized");
                 let result_hash =
@@ -2701,6 +2680,8 @@ fn get_embedding_preset(ruby: &Ruby, name: String) -> Result<Value, Error> {
             set_hash_entry(ruby, &hash, "chunk_size", preset.chunk_size.into_value_with(ruby))?;
             set_hash_entry(ruby, &hash, "overlap", preset.overlap.into_value_with(ruby))?;
 
+            // Note: When embeddings feature is enabled in kreuzberg, the model field is EmbeddingModel
+            // Since Ruby bindings typically build with all features, we use the model field and format it.
             let model_name = format!("{:?}", preset.model);
 
             set_hash_entry(ruby, &hash, "model_name", ruby.str_new(&model_name).as_value())?;
@@ -2713,43 +2694,6 @@ fn get_embedding_preset(ruby: &Ruby, name: String) -> Result<Value, Error> {
     }
 }
 
-/// Get the last error code from FFI
-///
-/// Returns an i32 error code indicating the type of error that occurred:
-/// - 0: Success (no error)
-/// - 1: GenericError
-/// - 2: Panic
-/// - 3: InvalidArgument
-/// - 4: IoError
-/// - 5: ParsingError
-/// - 6: OcrError
-/// - 7: MissingDependency
-///
-/// @return [Integer] The error code
-fn last_error_code() -> i32 {
-    get_error_code()
-}
-
-/// Get the last panic context from FFI as a JSON string
-///
-/// Returns a JSON string containing panic context if the last error was a panic,
-/// or nil if no panic context is available.
-///
-/// The JSON structure contains:
-/// - file: Source file where panic occurred
-/// - line: Line number
-/// - function: Function name
-/// - message: Panic message
-/// - timestamp_secs: Unix timestamp
-///
-/// @return [String, nil] JSON string with panic context or nil
-fn last_panic_context_json(ruby: &Ruby) -> Value {
-    match get_panic_context() {
-        Some(json) => ruby.str_new(&json).as_value(),
-        None => ruby.qnil().as_value(),
-    }
-}
-
 /// Initialize the Kreuzberg Ruby module
 #[magnus::init]
 fn init(ruby: &Ruby) -> Result<(), Error> {
@@ -2801,9 +2745,6 @@ fn init(ruby: &Ruby) -> Result<(), Error> {
     module.define_module_function("list_embedding_presets", function!(list_embedding_presets, 0))?;
     module.define_module_function("get_embedding_preset", function!(get_embedding_preset, 1))?;
 
-    module.define_module_function("_last_error_code_native", function!(last_error_code, 0))?;
-    module.define_module_function("_last_panic_context_json_native", function!(last_panic_context_json, 0))?;
-
     Ok(())
 }
 

data/kreuzberg.gemspec

@@ -9,17 +9,17 @@ ruby_prefix = 'packages/ruby/'
 ruby_cmd = %(git -C "#{repo_root}" ls-files -z #{ruby_prefix})
 ruby_files =
   `#{ruby_cmd}`.split("\x0")
-               .select { |path| path.start_with?(ruby_prefix) }
-               .map { |path| path.delete_prefix(ruby_prefix) }
+              .select { |path| path.start_with?(ruby_prefix) }
+              .map { |path| path.delete_prefix(ruby_prefix) }
 
 # Include the kreuzberg core crate (needed for path patch in Cargo.toml)
 core_prefix = 'crates/kreuzberg/'
 core_cmd = %(git -C "#{repo_root}" ls-files -z #{core_prefix})
 core_files =
   `#{core_cmd}`.split("\x0")
-               .select { |path| path.start_with?(core_prefix) }
-               .map { |path| path.delete_prefix('crates/') }
-               .map { |path| "vendor/#{path}" }
+              .select { |path| path.start_with?(core_prefix) }
+              .map { |path| path.delete_prefix('crates/') }
+              .map { |path| "vendor/#{path}" }
 
 fallback_files = Dir.chdir(__dir__) do
   ruby_fallback = Dir.glob(
@@ -44,57 +44,14 @@ fallback_files = Dir.chdir(__dir__) do
   # Fallback for core crate - copy from repo root
   core_fallback = Dir.chdir(repo_root) do
     Dir.glob('crates/kreuzberg/**/*', File::FNM_DOTMATCH)
-       .reject { |f| File.directory?(f) }
-       .map { |path| "vendor/#{path.delete_prefix('crates/')}" }
+      .reject { |f| File.directory?(f) }
+      .map { |path| "vendor/#{path.delete_prefix('crates/')}" }
   end
 
   ruby_fallback + core_fallback
 end
 
-# Check for vendored crates (copied during CI/packaging)
-vendor_files = Dir.chdir(__dir__) do
-  kreuzberg_files = if Dir.exist?('vendor/kreuzberg')
-                      Dir.glob('vendor/kreuzberg/**/*', File::FNM_DOTMATCH)
-                         .reject { |f| File.directory?(f) }
-                         .reject { |f| f.include?('/.fastembed_cache/') }
-                         .reject { |f| f.include?('/target/') }
-                         .grep_v(/\.(swp|bak|tmp)$/)
-                         .grep_v(/~$/)
-                    else
-                      []
-                    end
-
-  rb_sys_files = if Dir.exist?('vendor/rb-sys')
-                   Dir.glob('vendor/rb-sys/**/*', File::FNM_DOTMATCH)
-                      .reject { |f| File.directory?(f) }
-                      .reject { |f| f.include?('/target/') }
-                      .grep_v(/\.(swp|bak|tmp)$/)
-                      .grep_v(/~$/)
-                 else
-                   []
-                 end
-
-  workspace_toml = if File.exist?('vendor/Cargo.toml')
-                     ['vendor/Cargo.toml']
-                   else
-                     []
-                   end
-
-  kreuzberg_files + rb_sys_files + workspace_toml
-end
-
-# Use git-tracked files if available, otherwise fallback to glob
-# Always include vendored files if they exist on disk (for CI packaging)
-files = if (ruby_files + core_files).empty?
-          fallback_files
-        elsif vendor_files.any?
-          ruby_files + vendor_files
-        else
-          ruby_files + core_files
-        end
-
-# Filter to only include files that actually exist
-files = files.select { |f| File.exist?(f) }
+files = (ruby_files + core_files).empty? ? fallback_files : (ruby_files + core_files)
 
 Gem::Specification.new do |spec|
   spec.name = 'kreuzberg'
@@ -108,16 +65,16 @@ Gem::Specification.new do |spec|
     Rust core. Supports extraction, OCR, chunking, and language detection for 30+ file formats
     including PDF, DOCX, PPTX, XLSX, images, and more.
   DESC
-  spec.homepage = 'https://github.com/kreuzberg-dev/kreuzberg'
+  spec.homepage = 'https://github.com/Goldziher/kreuzberg'
   spec.license = 'MIT'
   spec.required_ruby_version = '>= 3.2.0'
 
   spec.metadata = {
     'homepage_uri' => spec.homepage,
-    'source_code_uri' => 'https://github.com/kreuzberg-dev/kreuzberg',
-    'changelog_uri' => 'https://github.com/kreuzberg-dev/kreuzberg/blob/main/CHANGELOG.md',
+    'source_code_uri' => 'https://github.com/Goldziher/kreuzberg',
+    'changelog_uri' => 'https://github.com/Goldziher/kreuzberg/blob/main/CHANGELOG.md',
     'documentation_uri' => 'https://docs.kreuzberg.dev',
-    'bug_tracker_uri' => 'https://github.com/kreuzberg-dev/kreuzberg/issues',
+    'bug_tracker_uri' => 'https://github.com/Goldziher/kreuzberg/issues',
     'rubygems_mfa_required' => 'true',
     'keywords' => 'document-intelligence,document-extraction,ocr,rust,bindings'
   }
@@ -132,10 +89,10 @@ Gem::Specification.new do |spec|
   # None - the gem is self-contained with the Rust extension
 
   # Development dependencies
-  spec.add_development_dependency 'bundler', '~> 4.0'
+  spec.add_development_dependency 'bundler', '~> 2.0'
   spec.add_development_dependency 'rake', '~> 13.0'
   spec.add_development_dependency 'rake-compiler', '~> 1.2'
-  spec.add_development_dependency 'rb_sys', '~> 0.9.119'
+  spec.add_development_dependency 'rb_sys', '~> 0.9'
   spec.add_development_dependency 'rspec', '~> 3.12'
   unless Gem.win_platform?
     spec.add_development_dependency 'rbs', '~> 3.0'

data/lib/kreuzberg/cache_api.rb

@@ -1,7 +1,6 @@
 # frozen_string_literal: true
 
 module Kreuzberg
-  # Provides caching capabilities for extraction results.
   module CacheAPI
     def clear_cache
       native_clear_cache

data/lib/kreuzberg/cli.rb

@@ -22,8 +22,8 @@ module Kreuzberg
     # @return [String] Extracted content
     #
     def extract(path, output: 'text', ocr: false)
-      args = ['extract', path, '--format', output]
-      args.push('--ocr', ocr ? 'true' : 'false')
+      args = ['extract', path, '--output', output]
+      args << '--ocr' if ocr
       CLIProxy.call(args)
     end
 

data/lib/kreuzberg/config.rb

@@ -39,7 +39,6 @@ module Kreuzberg
       end
     end
 
-    # Tesseract OCR engine configuration
     class Tesseract
       attr_reader :options
 
@@ -121,7 +120,6 @@ module Kreuzberg
       end
     end
 
-    # Embedding model configuration for document chunking
     class Embedding
       attr_reader :model, :normalize, :batch_size, :show_download_progress, :cache_dir
 
@@ -359,7 +357,6 @@ module Kreuzberg
       end
     end
 
-    # HTML preprocessing configuration for content extraction
     class HtmlPreprocessing
       attr_reader :enabled, :preset, :remove_navigation, :remove_forms
 
@@ -388,7 +385,6 @@ module Kreuzberg
       end
     end
 
-    # HTML rendering options for document conversion
     class HtmlOptions
       attr_reader :options
 
@@ -416,7 +412,6 @@ module Kreuzberg
       end
     end
 
-    # YAKE keyword extraction parameters
     class KeywordYakeParams
       attr_reader :window_size
 
@@ -429,7 +424,6 @@ module Kreuzberg
       end
     end
 
-    # RAKE keyword extraction parameters
     class KeywordRakeParams
       attr_reader :min_word_length, :max_words_per_phrase
 
@@ -446,7 +440,6 @@ module Kreuzberg
       end
     end
 
-    # Keyword extraction configuration for document analysis
     class Keywords
       attr_reader :algorithm, :max_keywords, :min_score, :ngram_range,
                   :language, :yake_params, :rake_params
@@ -652,7 +645,7 @@ module Kreuzberg
         @max_concurrent_extractions = max_concurrent_extractions&.to_i
       end
 
-      # rubocop:disable Metrics/CyclomaticComplexity
+      # rubocop:disable Metrics/PerceivedComplexity
       def to_h
         {
           use_cache: @use_cache,
@@ -671,7 +664,7 @@ module Kreuzberg
           max_concurrent_extractions: @max_concurrent_extractions
         }.compact
       end
-      # rubocop:enable Metrics/CyclomaticComplexity
+      # rubocop:enable Metrics/PerceivedComplexity
 
       private
 

data/lib/kreuzberg/errors.rb

@@ -1,77 +1,9 @@
 # frozen_string_literal: true
 
-require 'json'
-
 module Kreuzberg
-  # Error code constants matching kreuzberg-ffi error codes
-  ERROR_CODE_SUCCESS = 0
-  ERROR_CODE_GENERIC = 1
-  ERROR_CODE_PANIC = 2
-  ERROR_CODE_INVALID_ARGUMENT = 3
-  ERROR_CODE_IO = 4
-  ERROR_CODE_PARSING = 5
-  ERROR_CODE_OCR = 6
-  ERROR_CODE_MISSING_DEPENDENCY = 7
-
   module Errors
-    # Panic context information from FFI error introspection
-    class PanicContext
-      attr_reader :file, :line, :function, :message, :timestamp_secs
-
-      def initialize(file:, line:, function:, message:, timestamp_secs:)
-        @file = file
-        @line = line
-        @function = function
-        @message = message
-        @timestamp_secs = timestamp_secs
-      end
-
-      def to_s
-        "#{file}:#{line}:#{function}: #{message}"
-      end
-
-      def to_h
-        {
-          file:,
-          line:,
-          function:,
-          message:,
-          timestamp_secs:
-        }
-      end
-
-      def self.from_json(json_string)
-        return nil if json_string.nil? || json_string.empty?
-
-        data = JSON.parse(json_string, symbolize_names: true)
-        sliced = data.slice(:file, :line, :function, :message, :timestamp_secs)
-        new(**with_defaults(sliced))
-      rescue JSON::ParserError
-        nil
-      end
-
-      def self.with_defaults(sliced)
-        {
-          file: sliced[:file] || '',
-          line: sliced[:line] || 0,
-          function: sliced[:function] || '',
-          message: sliced[:message] || '',
-          timestamp_secs: sliced[:timestamp_secs] || 0
-        }
-      end
-      private_class_method :with_defaults
-    end
-
     # Base error class for all Kreuzberg errors
-    class Error < StandardError
-      attr_reader :panic_context, :error_code
-
-      def initialize(message, panic_context: nil, error_code: nil)
-        super(message)
-        @panic_context = panic_context
-        @error_code = error_code
-      end
-    end
+    class Error < StandardError; end
 
     # Raised when validation fails
     class ValidationError < Error; end
@@ -80,8 +12,8 @@ module Kreuzberg
     class ParsingError < Error
       attr_reader :context
 
-      def initialize(message, context: nil, panic_context: nil, error_code: nil)
-        super(message, panic_context:, error_code:)
+      def initialize(message, context: nil)
+        super(message)
         @context = context
       end
     end
@@ -90,8 +22,8 @@ module Kreuzberg
     class OCRError < Error
       attr_reader :context
 
-      def initialize(message, context: nil, panic_context: nil, error_code: nil)
-        super(message, panic_context:, error_code:)
+      def initialize(message, context: nil)
+        super(message)
         @context = context
       end
     end
@@ -100,8 +32,8 @@ module Kreuzberg
     class MissingDependencyError < Error
       attr_reader :dependency
 
-      def initialize(message, dependency: nil, panic_context: nil, error_code: nil)
-        super(message, panic_context:, error_code:)
+      def initialize(message, dependency: nil)
+        super(message)
         @dependency = dependency
       end
     end

data/lib/kreuzberg/extraction_api.rb

@@ -1,7 +1,6 @@
 # frozen_string_literal: true
 
 module Kreuzberg
-  # Provides extraction methods for documents and text.
   module ExtractionAPI
     def extract_file_sync(path, mime_type: nil, config: nil)
       opts = normalize_config(config)

data/lib/kreuzberg/setup_lib_path.rb

@@ -4,7 +4,6 @@ require 'rbconfig'
 require 'open3'
 
 module Kreuzberg
-  # Configures library paths for dynamic linking on different platforms.
   module SetupLibPath
     module_function
 

data/lib/kreuzberg/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Kreuzberg
-  VERSION = '4.0.0-rc.6'
+  VERSION = '4.0.0.rc1'
 end

data/lib/kreuzberg.rb

@@ -6,8 +6,6 @@ Kreuzberg::SetupLibPath.configure
 require_relative 'kreuzberg/version'
 require 'kreuzberg_rb'
 
-# Kreuzberg is a Ruby binding for the Rust core library providing document extraction,
-# text extraction, and OCR capabilities.
 module Kreuzberg
   autoload :Config, 'kreuzberg/config'
   autoload :Result, 'kreuzberg/result'
@@ -16,7 +14,6 @@ module Kreuzberg
   autoload :APIProxy, 'kreuzberg/api_proxy'
   autoload :MCPProxy, 'kreuzberg/mcp_proxy'
   autoload :Errors, 'kreuzberg/errors'
-  autoload :ErrorContext, 'kreuzberg/error_context'
   autoload :PostProcessorProtocol, 'kreuzberg/post_processor_protocol'
   autoload :ValidatorProtocol, 'kreuzberg/validator_protocol'
   autoload :OcrBackendProtocol, 'kreuzberg/ocr_backend_protocol'
@@ -76,24 +73,6 @@ module Kreuzberg
 
   # List all registered OCR backends.
   module_function :list_ocr_backends
-
-  # Detect MIME type from file bytes.
-  module_function :detect_mime_type
-
-  # Detect MIME type from a file path.
-  module_function :detect_mime_type_from_path
-
-  # Validate a MIME type string.
-  module_function :validate_mime_type
-
-  # Get file extensions for a given MIME type.
-  module_function :get_extensions_for_mime
-
-  # List all available embedding presets.
-  module_function :list_embedding_presets
-
-  # Get a specific embedding preset by name.
-  module_function :get_embedding_preset
 end
 
 require_relative 'kreuzberg/cache_api'