kreuzberg 4.0.0.pre.rc.13 → 4.0.0.pre.rc.15

data/vendor/kreuzberg/src/chunking/processor.rs

@@ -1,220 +1,220 @@
-//! Text chunking post-processor.
-//!
-//! This module provides a PostProcessor plugin that chunks text content in
-//! extraction results.
-
-use crate::plugins::{Plugin, PostProcessor, ProcessingStage};
-use crate::{ExtractionConfig, ExtractionResult, KreuzbergError, Result};
-use async_trait::async_trait;
-
-/// Post-processor that chunks text in document content.
-///
-/// This processor:
-/// - Runs in the Middle processing stage
-/// - Only processes when `config.chunking` is configured
-/// - Stores chunks in `result.chunks`
-/// - Uses configurable chunk size and overlap
-///
-/// # Example
-///
-/// ```rust,no_run
-/// use kreuzberg::plugins::{Plugin, PostProcessor};
-/// use kreuzberg::chunking::processor::ChunkingProcessor;
-///
-/// let processor = ChunkingProcessor;
-/// assert_eq!(processor.name(), "text-chunking");
-/// ```
-#[derive(Debug, Clone, Copy)]
-pub struct ChunkingProcessor;
-
-impl Plugin for ChunkingProcessor {
-    fn name(&self) -> &str {
-        "text-chunking"
-    }
-
-    fn version(&self) -> String {
-        env!("CARGO_PKG_VERSION").to_string()
-    }
-
-    fn initialize(&self) -> Result<()> {
-        Ok(())
-    }
-
-    fn shutdown(&self) -> Result<()> {
-        Ok(())
-    }
-}
-
-#[cfg_attr(not(target_arch = "wasm32"), async_trait)]
-#[cfg_attr(target_arch = "wasm32", async_trait(?Send))]
-impl PostProcessor for ChunkingProcessor {
-    async fn process(&self, result: &mut ExtractionResult, config: &ExtractionConfig) -> Result<()> {
-        let chunking_config = match &config.chunking {
-            Some(cfg) => cfg,
-            None => return Ok(()),
-        };
-
-        let chunk_config = crate::chunking::ChunkingConfig {
-            max_characters: chunking_config.max_chars,
-            overlap: chunking_config.max_overlap,
-            trim: true,
-            chunker_type: crate::chunking::ChunkerType::Text,
-        };
-
-        let chunking_result = crate::chunking::chunk_text(&result.content, &chunk_config, None)
-            .map_err(|e| KreuzbergError::Other(format!("Chunking failed: {}", e)))?;
-        result.chunks = Some(chunking_result.chunks);
-
-        Ok(())
-    }
-
-    fn processing_stage(&self) -> ProcessingStage {
-        ProcessingStage::Middle
-    }
-
-    fn should_process(&self, _result: &ExtractionResult, config: &ExtractionConfig) -> bool {
-        config.chunking.is_some()
-    }
-
-    fn estimated_duration_ms(&self, result: &ExtractionResult) -> u64 {
-        let text_length = result.content.len();
-        // Chunking is fast: ~1ms per 10KB
-        (text_length / 10240).max(1) as u64
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-    use crate::core::config::ChunkingConfig;
-    use crate::types::Metadata;
-
-    #[tokio::test]
-    async fn test_chunking_processor() {
-        let processor = ChunkingProcessor;
-        let config = ExtractionConfig {
-            chunking: Some(ChunkingConfig {
-                max_chars: 100,
-                max_overlap: 10,
-                embedding: None,
-                preset: None,
-            }),
-            ..Default::default()
-        };
-
-        let mut result = ExtractionResult {
-	            content: "This is a longer text that should be split into multiple chunks to test the chunking processor functionality.".to_string(),
-	            mime_type: "text/plain".to_string(),
-	            metadata: Metadata::default(),
-	            tables: vec![],
-	            detected_languages: None,
-	            chunks: None,
-	            images: None,
-	            pages: None,
-	        };
-
-        processor.process(&mut result, &config).await.unwrap();
-
-        assert!(result.chunks.is_some());
-        let chunks = result.chunks.unwrap();
-        assert!(!chunks.is_empty());
-    }
-
-    #[tokio::test]
-    async fn test_chunking_processor_no_config() {
-        let processor = ChunkingProcessor;
-        let config = ExtractionConfig::default();
-
-        let mut result = ExtractionResult {
-            content: "Some text".to_string(),
-            mime_type: "text/plain".to_string(),
-            metadata: Metadata::default(),
-            tables: vec![],
-            detected_languages: None,
-            chunks: None,
-            images: None,
-            pages: None,
-        };
-
-        processor.process(&mut result, &config).await.unwrap();
-
-        assert!(result.chunks.is_none());
-    }
-
-    #[test]
-    fn test_chunking_processor_plugin_interface() {
-        let processor = ChunkingProcessor;
-        assert_eq!(processor.name(), "text-chunking");
-        assert!(!processor.version().is_empty());
-        assert!(processor.initialize().is_ok());
-        assert!(processor.shutdown().is_ok());
-    }
-
-    #[test]
-    fn test_chunking_processor_stage() {
-        let processor = ChunkingProcessor;
-        assert_eq!(processor.processing_stage(), ProcessingStage::Middle);
-    }
-
-    #[test]
-    fn test_chunking_processor_should_process() {
-        let processor = ChunkingProcessor;
-
-        let result = ExtractionResult {
-            content: "Sample text".to_string(),
-            mime_type: "text/plain".to_string(),
-            metadata: Metadata::default(),
-            tables: vec![],
-            detected_languages: None,
-            chunks: None,
-            images: None,
-            pages: None,
-        };
-
-        let config_with_chunking = ExtractionConfig {
-            chunking: Some(crate::core::config::ChunkingConfig {
-                max_chars: 100,
-                max_overlap: 10,
-                embedding: None,
-                preset: None,
-            }),
-            ..Default::default()
-        };
-        assert!(processor.should_process(&result, &config_with_chunking));
-
-        let config_without_chunking = ExtractionConfig::default();
-        assert!(!processor.should_process(&result, &config_without_chunking));
-    }
-
-    #[test]
-    fn test_chunking_processor_estimated_duration() {
-        let processor = ChunkingProcessor;
-
-        let short_result = ExtractionResult {
-            content: "Short".to_string(),
-            mime_type: "text/plain".to_string(),
-            metadata: Metadata::default(),
-            tables: vec![],
-            detected_languages: None,
-            chunks: None,
-            images: None,
-            pages: None,
-        };
-
-        let long_result = ExtractionResult {
-            content: "a".repeat(100000),
-            mime_type: "text/plain".to_string(),
-            metadata: Metadata::default(),
-            tables: vec![],
-            detected_languages: None,
-            chunks: None,
-            images: None,
-            pages: None,
-        };
-
-        let short_duration = processor.estimated_duration_ms(&short_result);
-        let long_duration = processor.estimated_duration_ms(&long_result);
-
-        assert!(long_duration > short_duration);
-    }
-}
+//! Text chunking post-processor.
+//!
+//! This module provides a PostProcessor plugin that chunks text content in
+//! extraction results.
+
+use crate::plugins::{Plugin, PostProcessor, ProcessingStage};
+use crate::{ExtractionConfig, ExtractionResult, KreuzbergError, Result};
+use async_trait::async_trait;
+
+/// Post-processor that chunks text in document content.
+///
+/// This processor:
+/// - Runs in the Middle processing stage
+/// - Only processes when `config.chunking` is configured
+/// - Stores chunks in `result.chunks`
+/// - Uses configurable chunk size and overlap
+///
+/// # Example
+///
+/// ```rust,no_run
+/// use kreuzberg::plugins::{Plugin, PostProcessor};
+/// use kreuzberg::chunking::processor::ChunkingProcessor;
+///
+/// let processor = ChunkingProcessor;
+/// assert_eq!(processor.name(), "text-chunking");
+/// ```
+#[derive(Debug, Clone, Copy)]
+pub struct ChunkingProcessor;
+
+impl Plugin for ChunkingProcessor {
+    fn name(&self) -> &str {
+        "text-chunking"
+    }
+
+    fn version(&self) -> String {
+        env!("CARGO_PKG_VERSION").to_string()
+    }
+
+    fn initialize(&self) -> Result<()> {
+        Ok(())
+    }
+
+    fn shutdown(&self) -> Result<()> {
+        Ok(())
+    }
+}
+
+#[cfg_attr(not(target_arch = "wasm32"), async_trait)]
+#[cfg_attr(target_arch = "wasm32", async_trait(?Send))]
+impl PostProcessor for ChunkingProcessor {
+    async fn process(&self, result: &mut ExtractionResult, config: &ExtractionConfig) -> Result<()> {
+        let chunking_config = match &config.chunking {
+            Some(cfg) => cfg,
+            None => return Ok(()),
+        };
+
+        let chunk_config = crate::chunking::ChunkingConfig {
+            max_characters: chunking_config.max_chars,
+            overlap: chunking_config.max_overlap,
+            trim: true,
+            chunker_type: crate::chunking::ChunkerType::Text,
+        };
+
+        let chunking_result = crate::chunking::chunk_text(&result.content, &chunk_config, None)
+            .map_err(|e| KreuzbergError::Other(format!("Chunking failed: {}", e)))?;
+        result.chunks = Some(chunking_result.chunks);
+
+        Ok(())
+    }
+
+    fn processing_stage(&self) -> ProcessingStage {
+        ProcessingStage::Middle
+    }
+
+    fn should_process(&self, _result: &ExtractionResult, config: &ExtractionConfig) -> bool {
+        config.chunking.is_some()
+    }
+
+    fn estimated_duration_ms(&self, result: &ExtractionResult) -> u64 {
+        let text_length = result.content.len();
+        // Chunking is fast: ~1ms per 10KB
+        (text_length / 10240).max(1) as u64
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::core::config::ChunkingConfig;
+    use crate::types::Metadata;
+
+    #[tokio::test]
+    async fn test_chunking_processor() {
+        let processor = ChunkingProcessor;
+        let config = ExtractionConfig {
+            chunking: Some(ChunkingConfig {
+                max_chars: 100,
+                max_overlap: 10,
+                embedding: None,
+                preset: None,
+            }),
+            ..Default::default()
+        };
+
+        let mut result = ExtractionResult {
+	            content: "This is a longer text that should be split into multiple chunks to test the chunking processor functionality.".to_string(),
+	            mime_type: "text/plain".to_string(),
+	            metadata: Metadata::default(),
+	            tables: vec![],
+	            detected_languages: None,
+	            chunks: None,
+	            images: None,
+	            pages: None,
+	        };
+
+        processor.process(&mut result, &config).await.unwrap();
+
+        assert!(result.chunks.is_some());
+        let chunks = result.chunks.unwrap();
+        assert!(!chunks.is_empty());
+    }
+
+    #[tokio::test]
+    async fn test_chunking_processor_no_config() {
+        let processor = ChunkingProcessor;
+        let config = ExtractionConfig::default();
+
+        let mut result = ExtractionResult {
+            content: "Some text".to_string(),
+            mime_type: "text/plain".to_string(),
+            metadata: Metadata::default(),
+            tables: vec![],
+            detected_languages: None,
+            chunks: None,
+            images: None,
+            pages: None,
+        };
+
+        processor.process(&mut result, &config).await.unwrap();
+
+        assert!(result.chunks.is_none());
+    }
+
+    #[test]
+    fn test_chunking_processor_plugin_interface() {
+        let processor = ChunkingProcessor;
+        assert_eq!(processor.name(), "text-chunking");
+        assert!(!processor.version().is_empty());
+        assert!(processor.initialize().is_ok());
+        assert!(processor.shutdown().is_ok());
+    }
+
+    #[test]
+    fn test_chunking_processor_stage() {
+        let processor = ChunkingProcessor;
+        assert_eq!(processor.processing_stage(), ProcessingStage::Middle);
+    }
+
+    #[test]
+    fn test_chunking_processor_should_process() {
+        let processor = ChunkingProcessor;
+
+        let result = ExtractionResult {
+            content: "Sample text".to_string(),
+            mime_type: "text/plain".to_string(),
+            metadata: Metadata::default(),
+            tables: vec![],
+            detected_languages: None,
+            chunks: None,
+            images: None,
+            pages: None,
+        };
+
+        let config_with_chunking = ExtractionConfig {
+            chunking: Some(crate::core::config::ChunkingConfig {
+                max_chars: 100,
+                max_overlap: 10,
+                embedding: None,
+                preset: None,
+            }),
+            ..Default::default()
+        };
+        assert!(processor.should_process(&result, &config_with_chunking));
+
+        let config_without_chunking = ExtractionConfig::default();
+        assert!(!processor.should_process(&result, &config_without_chunking));
+    }
+
+    #[test]
+    fn test_chunking_processor_estimated_duration() {
+        let processor = ChunkingProcessor;
+
+        let short_result = ExtractionResult {
+            content: "Short".to_string(),
+            mime_type: "text/plain".to_string(),
+            metadata: Metadata::default(),
+            tables: vec![],
+            detected_languages: None,
+            chunks: None,
+            images: None,
+            pages: None,
+        };
+
+        let long_result = ExtractionResult {
+            content: "a".repeat(100000),
+            mime_type: "text/plain".to_string(),
+            metadata: Metadata::default(),
+            tables: vec![],
+            detected_languages: None,
+            chunks: None,
+            images: None,
+            pages: None,
+        };
+
+        let short_duration = processor.estimated_duration_ms(&short_result);
+        let long_duration = processor.estimated_duration_ms(&long_result);
+
+        assert!(long_duration > short_duration);
+    }
+}

data/vendor/kreuzberg/src/core/batch_mode.rs

@@ -1,95 +1,95 @@
-//! Internal batch mode tracking using tokio task-local storage.
-//!
-//! This module provides a way to track whether we're in batch processing mode
-//! without exposing it in the public API. Extractors check this flag to decide
-//! whether to use `spawn_blocking` for CPU-intensive work.
-
-use std::cell::Cell;
-use tokio::task_local;
-
-task_local! {
-    /// Task-local flag indicating batch processing mode.
-    ///
-    /// When true, extractors use `spawn_blocking` for CPU-intensive work to enable
-    /// parallelism. When false (single-file mode), extractors run directly to avoid
-    /// spawn overhead.
-    static BATCH_MODE: Cell<bool>;
-}
-
-/// Check if we're currently in batch processing mode.
-///
-/// Returns `false` if the task-local is not set (single-file mode).
-#[allow(dead_code)]
-pub fn is_batch_mode() -> bool {
-    BATCH_MODE.try_with(|cell| cell.get()).unwrap_or(false)
-}
-
-/// Run a future with batch mode enabled.
-///
-/// This sets the task-local BATCH_MODE flag for the duration of the future.
-pub async fn with_batch_mode<F, T>(future: F) -> T
-where
-    F: std::future::Future<Output = T>,
-{
-    BATCH_MODE.scope(Cell::new(true), future).await
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-
-    #[tokio::test]
-    async fn test_batch_mode_not_set_by_default() {
-        let result = is_batch_mode();
-        assert!(!result, "batch mode should be false by default");
-    }
-
-    #[tokio::test]
-    async fn test_with_batch_mode_sets_flag() {
-        let result = with_batch_mode(async { is_batch_mode() }).await;
-
-        assert!(result, "batch mode should be true inside with_batch_mode");
-    }
-
-    #[tokio::test]
-    async fn test_batch_mode_scoped_to_future() {
-        assert!(!is_batch_mode(), "batch mode should be false before");
-
-        with_batch_mode(async {
-            assert!(is_batch_mode(), "batch mode should be true inside");
-        })
-        .await;
-
-        assert!(!is_batch_mode(), "batch mode should be false after future completes");
-    }
-
-    #[tokio::test]
-    async fn test_nested_batch_mode_calls() {
-        let result = with_batch_mode(async {
-            let outer = is_batch_mode();
-            let inner = with_batch_mode(async { is_batch_mode() }).await;
-            (outer, inner)
-        })
-        .await;
-
-        assert!(result.0, "outer batch mode should be true");
-        assert!(result.1, "inner batch mode should be true");
-    }
-
-    #[tokio::test]
-    async fn test_batch_mode_unaffected_after_with_batch_mode() {
-        with_batch_mode(async {
-            assert!(is_batch_mode(), "first call should set batch mode");
-        })
-        .await;
-
-        assert!(!is_batch_mode(), "batch mode should be false between calls");
-
-        with_batch_mode(async {
-            assert!(is_batch_mode(), "second call should set batch mode");
-        })
-        .await;
-
-        assert!(!is_batch_mode(), "batch mode should be false after all calls");
-    }
-}
+//! Internal batch mode tracking using tokio task-local storage.
+//!
+//! This module provides a way to track whether we're in batch processing mode
+//! without exposing it in the public API. Extractors check this flag to decide
+//! whether to use `spawn_blocking` for CPU-intensive work.
+
+use std::cell::Cell;
+use tokio::task_local;
+
+task_local! {
+    /// Task-local flag indicating batch processing mode.
+    ///
+    /// When true, extractors use `spawn_blocking` for CPU-intensive work to enable
+    /// parallelism. When false (single-file mode), extractors run directly to avoid
+    /// spawn overhead.
+    static BATCH_MODE: Cell<bool>;
+}
+
+/// Check if we're currently in batch processing mode.
+///
+/// Returns `false` if the task-local is not set (single-file mode).
+#[allow(dead_code)]
+pub fn is_batch_mode() -> bool {
+    BATCH_MODE.try_with(|cell| cell.get()).unwrap_or(false)
+}
+
+/// Run a future with batch mode enabled.
+///
+/// This sets the task-local BATCH_MODE flag for the duration of the future.
+pub async fn with_batch_mode<F, T>(future: F) -> T
+where
+    F: std::future::Future<Output = T>,
+{
+    BATCH_MODE.scope(Cell::new(true), future).await
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[tokio::test]
+    async fn test_batch_mode_not_set_by_default() {
+        let result = is_batch_mode();
+        assert!(!result, "batch mode should be false by default");
+    }
+
+    #[tokio::test]
+    async fn test_with_batch_mode_sets_flag() {
+        let result = with_batch_mode(async { is_batch_mode() }).await;
+
+        assert!(result, "batch mode should be true inside with_batch_mode");
+    }
+
+    #[tokio::test]
+    async fn test_batch_mode_scoped_to_future() {
+        assert!(!is_batch_mode(), "batch mode should be false before");
+
+        with_batch_mode(async {
+            assert!(is_batch_mode(), "batch mode should be true inside");
+        })
+        .await;
+
+        assert!(!is_batch_mode(), "batch mode should be false after future completes");
+    }
+
+    #[tokio::test]
+    async fn test_nested_batch_mode_calls() {
+        let result = with_batch_mode(async {
+            let outer = is_batch_mode();
+            let inner = with_batch_mode(async { is_batch_mode() }).await;
+            (outer, inner)
+        })
+        .await;
+
+        assert!(result.0, "outer batch mode should be true");
+        assert!(result.1, "inner batch mode should be true");
+    }
+
+    #[tokio::test]
+    async fn test_batch_mode_unaffected_after_with_batch_mode() {
+        with_batch_mode(async {
+            assert!(is_batch_mode(), "first call should set batch mode");
+        })
+        .await;
+
+        assert!(!is_batch_mode(), "batch mode should be false between calls");
+
+        with_batch_mode(async {
+            assert!(is_batch_mode(), "second call should set batch mode");
+        })
+        .await;
+
+        assert!(!is_batch_mode(), "batch mode should be false after all calls");
+    }
+}