fix: propagate ocr_elements, paddle_ocr_config, element_config in Ruby binding

Add missing ocr_elements field conversion in extraction_result_to_ruby
and complete OCR config parsing (paddle_ocr_config, element_config,
output_format) in the Ruby native extension.

Author: Goldziher

packages/ruby/ext/kreuzberg_rb/native/src/config/types.rs

@@ -58,6 +58,35 @@ pub fn parse_ocr_config(ruby: &Ruby, hash: RHash) -> Result<OcrConfig, Error> {
         config.tesseract_config = Some(parsed);
     }
 
+    if let Some(val) = get_kw(ruby, hash, "paddle_ocr_config")
+        && !val.is_nil()
+    {
+        config.paddle_ocr_config = Some(ruby_value_to_json(val)?);
+    }
+
+    if let Some(val) = get_kw(ruby, hash, "element_config")
+        && !val.is_nil()
+    {
+        let ec_json = ruby_value_to_json(val)?;
+        let parsed: kreuzberg::types::OcrElementConfig =
+            serde_json::from_value(ec_json).map_err(|e| runtime_error(format!("Invalid element_config: {}", e)))?;
+        config.element_config = Some(parsed);
+    }
+
+    if let Some(val) = get_kw(ruby, hash, "output_format")
+        && !val.is_nil()
+    {
+        let format_str = symbol_to_string(val)?;
+        let format: OutputFormat = match format_str.as_str() {
+            "plain" | "Plain" => OutputFormat::Plain,
+            "markdown" | "Markdown" => OutputFormat::Markdown,
+            "djot" | "Djot" => OutputFormat::Djot,
+            "html" | "Html" => OutputFormat::Html,
+            other => return Err(runtime_error(format!("Invalid ocr output_format: '{}'", other))),
+        };
+        config.output_format = Some(format);
+    }
+
     Ok(config)
 }
 

packages/ruby/ext/kreuzberg_rb/native/src/result.rs

@@ -324,6 +324,20 @@ pub fn extraction_result_to_ruby(ruby: &Ruby, result: RustExtractionResult) -> R
         set_hash_entry(ruby, &hash, "elements", ruby.qnil().as_value())?;
     }
 
+    // Convert ocr_elements
+    if let Some(ocr_elements) = result.ocr_elements {
+        let elements_array = ruby.ary_new();
+        for elem in ocr_elements {
+            let elem_json = serde_json::to_value(&elem)
+                .map_err(|e| runtime_error(format!("Failed to serialize ocr_element: {}", e)))?;
+            let elem_ruby = json_value_to_ruby(ruby, &elem_json)?;
+            elements_array.push(elem_ruby)?;
+        }
+        set_hash_entry(ruby, &hash, "ocr_elements", elements_array.into_value_with(ruby))?;
+    } else {
+        set_hash_entry(ruby, &hash, "ocr_elements", ruby.qnil().as_value())?;
+    }
+
     // Convert document structure
     if let Some(doc_structure) = result.document {
         let document_hash = ruby.hash_new();