kreuzberg 4.9.1 → 4.9.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 559b3104e6e21f2f14a92949d427703b51ea8b35b7a643d8964b6953785aa6e1
-  data.tar.gz: bfce92579c45ecba0da0d8e1f077ecca0bb9dd6a1e96c950e8beb8d0a39b5884
+  metadata.gz: 431f2a7c19d9a4404369b7c9fe80fd10233b34b3462cf3c0076aac9b7f3065cb
+  data.tar.gz: bf36fd9dd46ffb84374e8f64ffb1f61ac6774cc8dd56d43d404e0b8995813ffc
 SHA512:
-  metadata.gz: 1ea8af57d65eb5008126758041df2bfe07acca9d47ebfbf9c9de79f12b5d5ff2336d55b643268d1ea420db1825eaf9bfef6e5deb7335bd2449e9ccb62800492d
-  data.tar.gz: '019f2abaa7dcaf2b91925f7ed0b5332ce569a41a5ae1f152d698423cce1276ea68e177c2088b4377c369943e235c70584f61a76da5600d6dd8b3bd075bc266ab'
+  metadata.gz: 2cf3ad1e3edd0778537e542bcafc121b6cd14235e271d2b9b27747f7beab7a2899362426f450a06f9ff0d0d4365982acb68300e46568887a871ad1bda6767d00
+  data.tar.gz: e7bcbe73f50701fbd4a0cb7e75e588096c02202198b234ba6f830b846f4fabd95e051ed0c25ed4a1c1d72a933cf5ccf9d755129eb1dc54ea446a60d9d9946997

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=v4.9.1" alt="Go">
+    <img src="https://img.shields.io/github/v/tag/kreuzberg-dev/kreuzberg?label=Go&color=007ec6&filter=v4.9.2" 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/ext/kreuzberg_rb/native/Cargo.lock

@@ -2916,7 +2916,7 @@ dependencies = [
 
 [[package]]
 name = "kreuzberg-rb"
-version = "4.9.1"
+version = "4.9.2"
 dependencies = [
  "async-trait",
  "html-to-markdown-rs",
@@ -3040,9 +3040,9 @@ checksum = "92daf443525c4cce67b150400bc2316076100ce0b3686209eb8cf3c31612e6f0"
 
 [[package]]
 name = "liter-llm"
-version = "1.2.1"
+version = "1.2.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1884be380e45da823105c85ef0fa188af81d57be7de9b65016576e1774fdd5f8"
+checksum = "4e4ce5d2d0b09f2e63537ba40b15b0a95c2d6818ed0454eb04d9593ba4a0cad3"
 dependencies = [
  "base64 0.22.1",
  "bytes",

data/ext/kreuzberg_rb/native/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "kreuzberg-rb"
-version = "4.9.1"
+version = "4.9.2"
 edition = "2024"
 rust-version = "1.91"
 authors = ["Na'aman Hirschfeld <naaman@kreuzberg.dev>"]

data/lib/kreuzberg/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Kreuzberg
-  VERSION = '4.9.1'
+  VERSION = '4.9.2'
 end

data/vendor/Cargo.toml

@@ -2,7 +2,7 @@
 members = ["kreuzberg", "kreuzberg-ffi", "kreuzberg-tesseract", "kreuzberg-paddle-ocr", "kreuzberg-pdfium-render"]
 
 [workspace.package]
-version = "4.9.1"
+version = "4.9.2"
 edition = "2024"
 rust-version = "1.91"
 authors = ["Na'aman Hirschfeld <naaman@kreuzberg.dev>"]
@@ -32,8 +32,8 @@ html-to-markdown-rs = { version = "3.2.5", default-features = false }
 image = { version = "0.25.10", default-features = false }
 itertools = "0.14"
 js-sys = "0.3"
-kreuzberg = { path = "./crates/kreuzberg", version = "4.9.1", default-features = false }
-kreuzberg-ffi = { path = "./crates/kreuzberg-ffi", version = "4.9.1" }
+kreuzberg = { path = "./crates/kreuzberg", version = "4.9.2", default-features = false }
+kreuzberg-ffi = { path = "./crates/kreuzberg-ffi", version = "4.9.2" }
 lazy_static = "1.5.0"
 libc = "0.2.185"
 liter-llm = { version = "1.2", features = ["native-http", "tracing"], default-features = false }

data/vendor/kreuzberg/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "kreuzberg"
-version = "4.9.1"
+version = "4.9.2"
 edition = "2024"
 rust-version = "1.91"
 authors = ["Na'aman Hirschfeld <naaman@kreuzberg.dev>"]

data/vendor/kreuzberg/README.md

@@ -18,7 +18,7 @@ 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 4.9.1 Release**
+> **🚀 Version 4.9.2 Release**
 > 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.
 >
 > **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/src/core/config/ocr.rs

@@ -323,9 +323,12 @@ impl OcrConfig {
     /// Returns the effective pipeline config.
     ///
     /// - If `pipeline` is explicitly set, returns it.
-    /// - If `paddle-ocr` feature is compiled in and no explicit pipeline is set,
-    ///   auto-constructs a default pipeline: primary backend (priority 100) + paddleocr (priority 50).
-    /// - Otherwise returns `None` (single-backend mode, same as today).
+    /// - If `paddle-ocr` is compiled in and the backend is the default
+    ///   (tesseract), auto-constructs `[tesseract @ 100, paddleocr @ 50]`.
+    /// - Otherwise returns `None` (single-backend mode).
+    ///
+    /// Explicit non-default backend selections are honored as-is — a silent
+    /// paddleocr fallback would mask errors from the chosen backend.
     pub fn effective_pipeline(&self) -> Option<OcrPipelineConfig> {
         if self.pipeline.is_some() {
             return self.pipeline.clone();
@@ -333,25 +336,28 @@ impl OcrConfig {
 
         #[cfg(feature = "paddle-ocr")]
         {
-            let mut stages = vec![OcrPipelineStage {
-                backend: self.backend.clone(),
-                priority: 100,
-                language: None,
-                tesseract_config: self.tesseract_config.clone(),
-                paddle_ocr_config: None,
-                vlm_config: self.vlm_config.clone(),
-            }];
-            // Only add paddleocr fallback if primary backend isn't already paddleocr
-            if self.backend != "paddleocr" {
-                stages.push(OcrPipelineStage {
+            if self.backend != default_tesseract_backend() {
+                return None;
+            }
+
+            let stages = vec![
+                OcrPipelineStage {
+                    backend: self.backend.clone(),
+                    priority: 100,
+                    language: None,
+                    tesseract_config: self.tesseract_config.clone(),
+                    paddle_ocr_config: None,
+                    vlm_config: self.vlm_config.clone(),
+                },
+                OcrPipelineStage {
                     backend: "paddleocr".to_string(),
                     priority: 50,
                     language: None,
                     tesseract_config: None,
                     paddle_ocr_config: self.paddle_ocr_config.clone(),
                     vlm_config: None,
-                });
-            }
+                },
+            ];
             Some(OcrPipelineConfig {
                 stages,
                 quality_thresholds: self.effective_thresholds(),
@@ -485,29 +491,21 @@ mod tests {
     }
 
     #[test]
-    fn test_effective_pipeline_paddleocr_backend_no_duplicate() {
-        // When primary backend is "paddleocr", effective_pipeline should NOT add
-        // a second paddleocr stage (issue #6 fix).
+    fn test_effective_pipeline_explicit_paddleocr_no_autofallback() {
         let config = OcrConfig {
             backend: "paddleocr".to_string(),
             ..Default::default()
         };
-        let result = config.effective_pipeline();
-        // With paddle-ocr feature: should have exactly 1 stage (no duplicate)
-        // Without paddle-ocr feature: should be None
-        #[cfg(feature = "paddle-ocr")]
-        {
-            let pipeline = result.unwrap();
-            let paddle_count = pipeline.stages.iter().filter(|s| s.backend == "paddleocr").count();
-            assert_eq!(
-                paddle_count, 1,
-                "Should not have duplicate paddleocr stages, found {paddle_count}"
-            );
-        }
-        #[cfg(not(feature = "paddle-ocr"))]
-        {
-            assert!(result.is_none());
-        }
+        assert!(config.effective_pipeline().is_none());
+    }
+
+    #[test]
+    fn test_effective_pipeline_explicit_easyocr_no_autofallback() {
+        let config = OcrConfig {
+            backend: "easyocr".to_string(),
+            ..Default::default()
+        };
+        assert!(config.effective_pipeline().is_none());
     }
 
     #[test]

data/vendor/kreuzberg/src/extraction/pptx/mod.rs

@@ -282,9 +282,7 @@ fn extract_pptx_from_container<R: std::io::Read + std::io::Seek>(
         }),
     });
 
-    let document = doc_builder
-        .map(|b| b.build())
-        .filter(|d| !d.is_empty());
+    let document = doc_builder.map(|b| b.build()).filter(|d| !d.is_empty());
 
     Ok(PptxExtractionResult {
         content,

data/vendor/kreuzberg/src/extractors/doc.rs

@@ -85,7 +85,12 @@ impl DocumentExtractor for DocExtractor {
             }
 
             #[cfg(not(feature = "tokio-runtime"))]
-            extract_doc_text(content)
+            {
+                if config.cancel_token.as_ref().map(|t| t.is_cancelled()).unwrap_or(false) {
+                    return Err(crate::error::KreuzbergError::Cancelled);
+                }
+                extract_doc_text(content)
+            }
         }?;
 
         let mut doc = InternalDocument::new("doc");

data/vendor/kreuzberg/src/extractors/excel.rs

@@ -215,6 +215,9 @@ impl DocumentExtractor for ExcelExtractor {
             }
             #[cfg(not(feature = "tokio-runtime"))]
             {
+                if config.cancel_token.as_ref().map(|t| t.is_cancelled()).unwrap_or(false) {
+                    return Err(crate::error::KreuzbergError::Cancelled);
+                }
                 crate::extraction::excel::read_excel_bytes(content, extension)?
             }
         };

data/vendor/kreuzberg/src/extractors/iwork/keynote.rs

@@ -172,7 +172,12 @@ impl DocumentExtractor for KeynoteExtractor {
             }
 
             #[cfg(not(feature = "tokio-runtime"))]
-            parse_keynote(content)?
+            {
+                if config.cancel_token.as_ref().map(|t| t.is_cancelled()).unwrap_or(false) {
+                    return Err(crate::error::KreuzbergError::Cancelled);
+                }
+                parse_keynote(content)?
+            }
         };
 
         let mut doc = build_keynote_internal_document(&data);

data/vendor/kreuzberg/src/extractors/iwork/numbers.rs

@@ -181,7 +181,12 @@ impl DocumentExtractor for NumbersExtractor {
             }
 
             #[cfg(not(feature = "tokio-runtime"))]
-            parse_numbers(content)?
+            {
+                if config.cancel_token.as_ref().map(|t| t.is_cancelled()).unwrap_or(false) {
+                    return Err(crate::error::KreuzbergError::Cancelled);
+                }
+                parse_numbers(content)?
+            }
         };
 
         let mut doc = build_numbers_internal_document(&data);

data/vendor/kreuzberg/src/extractors/iwork/pages.rs

@@ -163,7 +163,12 @@ impl DocumentExtractor for PagesExtractor {
             }
 
             #[cfg(not(feature = "tokio-runtime"))]
-            parse_pages(content)?
+            {
+                if config.cancel_token.as_ref().map(|t| t.is_cancelled()).unwrap_or(false) {
+                    return Err(crate::error::KreuzbergError::Cancelled);
+                }
+                parse_pages(content)?
+            }
         };
 
         let mut doc = build_pages_internal_document(&data);

data/vendor/kreuzberg/src/extractors/pdf/extraction.rs

@@ -420,6 +420,10 @@ pub(crate) fn extract_all_from_oxide_document(
     content: &[u8],
     config: &ExtractionConfig,
     layout_hints: Option<&[Vec<crate::pdf::structure::types::LayoutHint>]>,
+    #[cfg(feature = "layout-detection")] layout_images: Option<&[image::DynamicImage]>,
+    #[cfg(not(feature = "layout-detection"))] _layout_images: Option<()>,
+    #[cfg(feature = "layout-detection")] layout_results: Option<&[crate::pdf::layout_runner::PageLayoutResult]>,
+    #[cfg(not(feature = "layout-detection"))] _layout_results: Option<()>,
 ) -> Result<PdfExtractionPhaseResult> {
     let _span = tracing::debug_span!("extract_pdf_oxide").entered();
 
@@ -460,6 +464,11 @@ pub(crate) fn extract_all_from_oxide_document(
         OutputFormat::Markdown | OutputFormat::Djot | OutputFormat::Html
     );
 
+    let allow_single_column = config
+        .pdf_options
+        .as_ref()
+        .is_some_and(|o| o.allow_single_column_tables);
+
     let pre_rendered_doc =
         if needs_structured && !config.force_ocr {
             let k = config
@@ -503,6 +512,15 @@ pub(crate) fn extract_all_from_oxide_document(
                     used_structure_tree,
                     image_positions: &image_positions,
                     layout_hints,
+                    allow_single_column,
+                    #[cfg(feature = "layout-detection")]
+                    layout_images,
+                    #[cfg(feature = "layout-detection")]
+                    layout_results,
+                    #[cfg(feature = "layout-detection")]
+                    table_model: config.layout.as_ref().map(|l| l.table_model).unwrap_or_default(),
+                    #[cfg(feature = "layout-detection")]
+                    acceleration: config.acceleration.as_ref(),
                 },
             ) {
                 Ok(structured_doc) if !structured_doc.elements.is_empty() => {

data/vendor/kreuzberg/src/extractors/pdf/mod.rs

@@ -1192,7 +1192,14 @@ impl PdfExtractor {
         #[cfg(feature = "layout-detection")]
         let layout_bundle = run_layout_detection(content, config);
         #[cfg(feature = "layout-detection")]
-        let layout_hints = layout_bundle.as_ref().map(|b| b.hints.as_slice());
+        let (layout_hints, layout_images, layout_results) = match layout_bundle {
+            Some(ref bundle) => (
+                Some(bundle.hints.as_slice()),
+                Some(bundle.images.as_slice()),
+                Some(bundle.results.as_slice()),
+            ),
+            None => (None, None, None),
+        };
         #[cfg(not(feature = "layout-detection"))]
         let layout_hints: Option<&[Vec<crate::pdf::structure::types::LayoutHint>]> = None;
 
@@ -1206,7 +1213,19 @@ impl PdfExtractor {
             pre_rendered_doc,
             _has_font_encoding_issues,
             pdf_annotations,
-        ) = extract_all_from_oxide_document(content, config, layout_hints)?;
+        ) = extract_all_from_oxide_document(
+            content,
+            config,
+            layout_hints,
+            #[cfg(feature = "layout-detection")]
+            layout_images,
+            #[cfg(not(feature = "layout-detection"))]
+            None,
+            #[cfg(feature = "layout-detection")]
+            layout_results,
+            #[cfg(not(feature = "layout-detection"))]
+            None,
+        )?;
 
         // --- OCR evaluation (reuses the same logic as the pdfium path) ---
         #[cfg(feature = "ocr")]

data/vendor/kreuzberg/src/extractors/ppt.rs

@@ -136,7 +136,12 @@ impl DocumentExtractor for PptExtractor {
             }
 
             #[cfg(not(feature = "tokio-runtime"))]
-            crate::extraction::ppt::extract_ppt_text_with_options(content, include_master_slides)
+            {
+                if config.cancel_token.as_ref().map(|t| t.is_cancelled()).unwrap_or(false) {
+                    return Err(crate::error::KreuzbergError::Cancelled);
+                }
+                crate::extraction::ppt::extract_ppt_text_with_options(content, include_master_slides)
+            }
         }?;
 
         let mut metadata_map = AHashMap::new();

data/vendor/kreuzberg/src/layout/model_manager.rs

@@ -9,8 +9,18 @@ use std::path::{Path, PathBuf};
 use crate::layout::error::LayoutError;
 use crate::model_download;
 
+#[cfg(feature = "paddle-ocr")]
 use crate::paddle_ocr::ModelManifestEntry;
 
+#[cfg(not(feature = "paddle-ocr"))]
+#[derive(Debug, Clone, serde::Serialize)]
+pub struct ModelManifestEntry {
+    pub relative_path: String,
+    pub sha256: String,
+    pub size_bytes: u64,
+    pub source_url: String,
+}
+
 /// Model definition for a layout model.
 #[derive(Debug, Clone)]
 struct ModelDefinition {