kreuzberg 4.0.8 → 4.1.0

data/vendor/kreuzberg/src/core/extractor/mod.rs

@@ -0,0 +1,490 @@
+//! Main extraction entry points.
+//!
+//! This module provides the primary API for extracting content from files and byte arrays.
+//! It orchestrates the entire extraction pipeline: cache checking, MIME detection,
+//! extractor selection, extraction, post-processing, and cache storage.
+//!
+//! # Functions
+//!
+//! - [`extract_file`] - Extract content from a file path
+//! - [`extract_bytes`] - Extract content from a byte array
+//! - [`batch_extract_file`] - Extract content from multiple files concurrently
+//! - [`batch_extract_bytes`] - Extract content from multiple byte arrays concurrently
+
+mod bytes;
+mod file;
+mod helpers;
+mod legacy;
+mod sync;
+
+#[cfg(feature = "tokio-runtime")]
+mod batch;
+
+// Re-export public API
+pub use bytes::extract_bytes;
+pub use file::extract_file;
+pub use helpers::get_pool_sizing_hint;
+pub use sync::{batch_extract_bytes_sync, extract_bytes_sync};
+
+#[cfg(feature = "tokio-runtime")]
+pub use sync::extract_file_sync;
+
+#[cfg(feature = "tokio-runtime")]
+pub use batch::{batch_extract_bytes, batch_extract_file};
+#[cfg(feature = "tokio-runtime")]
+pub use sync::batch_extract_file_sync;
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::core::config::ExtractionConfig;
+    use serial_test::serial;
+    use std::fs::File;
+    use std::io::Write;
+    use std::sync::Arc;
+    use tempfile::tempdir;
+
+    fn assert_text_content(actual: &str, expected: &str) {
+        assert_eq!(actual.trim_end_matches('\n'), expected);
+    }
+
+    #[tokio::test]
+    async fn test_extract_file_basic() {
+        let dir = tempdir().unwrap();
+        let file_path = dir.path().join("test.txt");
+        let mut file = File::create(&file_path).unwrap();
+        file.write_all(b"Hello, world!").unwrap();
+
+        let config = ExtractionConfig::default();
+        let result = extract_file(&file_path, None, &config).await;
+
+        assert!(result.is_ok());
+        let result = result.unwrap();
+        assert_text_content(&result.content, "Hello, world!");
+        assert_eq!(result.mime_type, "text/plain");
+    }
+
+    #[tokio::test]
+    async fn test_extract_file_with_mime_override() {
+        let dir = tempdir().unwrap();
+        let file_path = dir.path().join("test.dat");
+        let mut file = File::create(&file_path).unwrap();
+        file.write_all(b"test content").unwrap();
+
+        let config = ExtractionConfig::default();
+        let result = extract_file(&file_path, Some("text/plain"), &config).await;
+
+        assert!(result.is_ok());
+        let result = result.unwrap();
+        assert_eq!(result.mime_type, "text/plain");
+    }
+
+    #[tokio::test]
+    async fn test_extract_file_nonexistent() {
+        let config = ExtractionConfig::default();
+        let result = extract_file("/nonexistent/file.txt", None, &config).await;
+        assert!(result.is_err());
+    }
+
+    #[tokio::test]
+    async fn test_extract_bytes_basic() {
+        let config = ExtractionConfig::default();
+        let result = extract_bytes(b"test content", "text/plain", &config).await;
+
+        assert!(result.is_ok());
+        let result = result.unwrap();
+        assert_text_content(&result.content, "test content");
+        assert_eq!(result.mime_type, "text/plain");
+    }
+
+    #[tokio::test]
+    async fn test_extract_bytes_invalid_mime() {
+        let config = ExtractionConfig::default();
+        let result = extract_bytes(b"test", "invalid/mime", &config).await;
+        assert!(result.is_err());
+    }
+
+    #[tokio::test]
+    async fn test_batch_extract_file() {
+        let dir = tempdir().unwrap();
+
+        let file1 = dir.path().join("test1.txt");
+        let file2 = dir.path().join("test2.txt");
+
+        File::create(&file1).unwrap().write_all(b"content 1").unwrap();
+        File::create(&file2).unwrap().write_all(b"content 2").unwrap();
+
+        let config = ExtractionConfig::default();
+        let paths = vec![file1, file2];
+        let results = batch_extract_file(paths, &config).await;
+
+        assert!(results.is_ok());
+        let results = results.unwrap();
+        assert_eq!(results.len(), 2);
+        assert_text_content(&results[0].content, "content 1");
+        assert_text_content(&results[1].content, "content 2");
+    }
+
+    #[tokio::test]
+    async fn test_batch_extract_file_empty() {
+        let config = ExtractionConfig::default();
+        let paths: Vec<std::path::PathBuf> = vec![];
+        let results = batch_extract_file(paths, &config).await;
+
+        assert!(results.is_ok());
+        assert_eq!(results.unwrap().len(), 0);
+    }
+
+    #[tokio::test]
+    async fn test_batch_extract_bytes() {
+        let config = ExtractionConfig::default();
+        let contents = vec![
+            (b"content 1".as_slice(), "text/plain"),
+            (b"content 2".as_slice(), "text/plain"),
+        ];
+        let owned_contents: Vec<(Vec<u8>, String)> = contents
+            .into_iter()
+            .map(|(bytes, mime)| (bytes.to_vec(), mime.to_string()))
+            .collect();
+        let results = batch_extract_bytes(owned_contents, &config).await;
+
+        assert!(results.is_ok());
+        let results = results.unwrap();
+        assert_eq!(results.len(), 2);
+        assert_text_content(&results[0].content, "content 1");
+        assert_text_content(&results[1].content, "content 2");
+    }
+
+    #[test]
+    fn test_sync_wrappers() {
+        let dir = tempdir().unwrap();
+        let file_path = dir.path().join("test.txt");
+        File::create(&file_path).unwrap().write_all(b"sync test").unwrap();
+
+        let config = ExtractionConfig::default();
+
+        let result = extract_file_sync(&file_path, None, &config);
+        assert!(result.is_ok());
+        let result = result.unwrap();
+        assert_text_content(&result.content, "sync test");
+
+        let result = extract_bytes_sync(b"test", "text/plain", &config);
+        assert!(result.is_ok());
+    }
+
+    #[tokio::test]
+    async fn test_extractor_cache() {
+        let config = ExtractionConfig::default();
+
+        let result1 = extract_bytes(b"test 1", "text/plain", &config).await;
+        assert!(result1.is_ok());
+        let result1 = result1.unwrap();
+
+        let result2 = extract_bytes(b"test 2", "text/plain", &config).await;
+        assert!(result2.is_ok());
+        let result2 = result2.unwrap();
+
+        assert_text_content(&result1.content, "test 1");
+        assert_text_content(&result2.content, "test 2");
+
+        let result3 = extract_bytes(b"# test 3", "text/markdown", &config).await;
+        assert!(result3.is_ok());
+    }
+
+    #[tokio::test]
+    async fn test_extract_file_empty() {
+        let dir = tempdir().unwrap();
+        let file_path = dir.path().join("empty.txt");
+        File::create(&file_path).unwrap();
+
+        let config = ExtractionConfig::default();
+        let result = extract_file(&file_path, None, &config).await;
+
+        assert!(result.is_ok());
+        let result = result.unwrap();
+        assert_eq!(result.content, "");
+    }
+
+    #[tokio::test]
+    async fn test_extract_bytes_empty() {
+        let config = ExtractionConfig::default();
+        let result = extract_bytes(b"", "text/plain", &config).await;
+
+        assert!(result.is_ok());
+        let result = result.unwrap();
+        assert_eq!(result.content, "");
+    }
+
+    #[tokio::test]
+    async fn test_extract_file_whitespace_only() {
+        let dir = tempdir().unwrap();
+        let file_path = dir.path().join("whitespace.txt");
+        File::create(&file_path).unwrap().write_all(b"   \n\t  \n  ").unwrap();
+
+        let config = ExtractionConfig::default();
+        let result = extract_file(&file_path, None, &config).await;
+
+        assert!(result.is_ok());
+    }
+
+    #[tokio::test]
+    async fn test_extract_file_very_long_path() {
+        let dir = tempdir().unwrap();
+        let long_name = "a".repeat(200);
+        let file_path = dir.path().join(format!("{}.txt", long_name));
+
+        if let Ok(mut f) = File::create(&file_path) {
+            f.write_all(b"content").unwrap();
+            let config = ExtractionConfig::default();
+            let result = extract_file(&file_path, None, &config).await;
+            assert!(result.is_ok() || result.is_err());
+        }
+    }
+
+    #[tokio::test]
+    async fn test_extract_file_special_characters_in_path() {
+        let dir = tempdir().unwrap();
+        let file_path = dir.path().join("test with spaces & symbols!.txt");
+        File::create(&file_path).unwrap().write_all(b"content").unwrap();
+
+        let config = ExtractionConfig::default();
+        let result = extract_file(&file_path, None, &config).await;
+
+        assert!(result.is_ok());
+        let result = result.unwrap();
+        assert_text_content(&result.content, "content");
+    }
+
+    #[tokio::test]
+    async fn test_extract_file_unicode_filename() {
+        let dir = tempdir().unwrap();
+        let file_path = dir.path().join("测试文件名.txt");
+        File::create(&file_path).unwrap().write_all(b"content").unwrap();
+
+        let config = ExtractionConfig::default();
+        let result = extract_file(&file_path, None, &config).await;
+
+        assert!(result.is_ok());
+    }
+
+    #[tokio::test]
+    async fn test_extract_bytes_unsupported_mime() {
+        let config = ExtractionConfig::default();
+        let result = extract_bytes(b"test", "application/x-unknown-format", &config).await;
+
+        assert!(result.is_err());
+        use crate::KreuzbergError;
+        assert!(matches!(result.unwrap_err(), KreuzbergError::UnsupportedFormat(_)));
+    }
+
+    #[tokio::test]
+    async fn test_batch_extract_file_with_errors() {
+        let dir = tempdir().unwrap();
+
+        let valid_file = dir.path().join("valid.txt");
+        File::create(&valid_file).unwrap().write_all(b"valid content").unwrap();
+
+        let invalid_file = dir.path().join("nonexistent.txt");
+
+        let config = ExtractionConfig::default();
+        let paths = vec![valid_file, invalid_file];
+        let results = batch_extract_file(paths, &config).await;
+
+        assert!(results.is_ok());
+        let results = results.unwrap();
+        assert_eq!(results.len(), 2);
+        assert_text_content(&results[0].content, "valid content");
+        assert!(results[1].metadata.error.is_some());
+    }
+
+    #[tokio::test]
+    async fn test_batch_extract_bytes_mixed_valid_invalid() {
+        let config = ExtractionConfig::default();
+        let contents = vec![
+            (b"valid 1".as_slice(), "text/plain"),
+            (b"invalid".as_slice(), "invalid/mime"),
+            (b"valid 2".as_slice(), "text/plain"),
+        ];
+        let owned_contents: Vec<(Vec<u8>, String)> = contents
+            .into_iter()
+            .map(|(bytes, mime)| (bytes.to_vec(), mime.to_string()))
+            .collect();
+        let results = batch_extract_bytes(owned_contents, &config).await;
+
+        assert!(results.is_ok());
+        let results = results.unwrap();
+        assert_eq!(results.len(), 3);
+        assert_text_content(&results[0].content, "valid 1");
+        assert!(results[1].metadata.error.is_some());
+        assert_text_content(&results[2].content, "valid 2");
+    }
+
+    #[tokio::test]
+    async fn test_batch_extract_bytes_all_invalid() {
+        let config = ExtractionConfig::default();
+        let contents = vec![
+            (b"test 1".as_slice(), "invalid/mime1"),
+            (b"test 2".as_slice(), "invalid/mime2"),
+        ];
+        let owned_contents: Vec<(Vec<u8>, String)> = contents
+            .into_iter()
+            .map(|(bytes, mime)| (bytes.to_vec(), mime.to_string()))
+            .collect();
+        let results = batch_extract_bytes(owned_contents, &config).await;
+
+        assert!(results.is_ok());
+        let results = results.unwrap();
+        assert_eq!(results.len(), 2);
+        assert!(results[0].metadata.error.is_some());
+        assert!(results[1].metadata.error.is_some());
+    }
+
+    #[tokio::test]
+    async fn test_extract_bytes_very_large() {
+        let large_content = vec![b'a'; 10_000_000];
+        let config = ExtractionConfig::default();
+        let result = extract_bytes(&large_content, "text/plain", &config).await;
+
+        assert!(result.is_ok());
+        let result = result.unwrap();
+        let trimmed_len = result.content.trim_end_matches('\n').len();
+        assert_eq!(trimmed_len, 10_000_000);
+    }
+
+    #[tokio::test]
+    async fn test_batch_extract_large_count() {
+        let dir = tempdir().unwrap();
+        let mut paths = Vec::new();
+
+        for i in 0..100 {
+            let file_path = dir.path().join(format!("file{}.txt", i));
+            File::create(&file_path)
+                .unwrap()
+                .write_all(format!("content {}", i).as_bytes())
+                .unwrap();
+            paths.push(file_path);
+        }
+
+        let config = ExtractionConfig::default();
+        let results = batch_extract_file(paths, &config).await;
+
+        assert!(results.is_ok());
+        let results = results.unwrap();
+        assert_eq!(results.len(), 100);
+
+        for (i, result) in results.iter().enumerate() {
+            assert_text_content(&result.content, &format!("content {}", i));
+        }
+    }
+
+    #[tokio::test]
+    async fn test_extract_file_mime_detection_fallback() {
+        let dir = tempdir().unwrap();
+        let file_path = dir.path().join("testfile");
+        File::create(&file_path)
+            .unwrap()
+            .write_all(b"plain text content")
+            .unwrap();
+
+        let config = ExtractionConfig::default();
+        let result = extract_file(&file_path, None, &config).await;
+
+        assert!(result.is_ok() || result.is_err());
+    }
+
+    #[tokio::test]
+    async fn test_extract_file_wrong_mime_override() {
+        let dir = tempdir().unwrap();
+        let file_path = dir.path().join("test.txt");
+        File::create(&file_path).unwrap().write_all(b"plain text").unwrap();
+
+        let config = ExtractionConfig::default();
+        let result = extract_file(&file_path, Some("application/pdf"), &config).await;
+
+        assert!(result.is_err() || result.is_ok());
+    }
+
+    #[test]
+    fn test_sync_wrapper_nonexistent_file() {
+        let config = ExtractionConfig::default();
+        let result = extract_file_sync("/nonexistent/path.txt", None, &config);
+
+        assert!(result.is_err());
+        use crate::KreuzbergError;
+        assert!(matches!(result.unwrap_err(), KreuzbergError::Validation { .. }));
+    }
+
+    #[test]
+    fn test_sync_wrapper_batch_empty() {
+        let config = ExtractionConfig::default();
+        let paths: Vec<std::path::PathBuf> = vec![];
+        let results = batch_extract_file_sync(paths, &config);
+
+        assert!(results.is_ok());
+        assert_eq!(results.unwrap().len(), 0);
+    }
+
+    #[test]
+    fn test_sync_wrapper_batch_bytes_empty() {
+        let config = ExtractionConfig::default();
+        let contents: Vec<(Vec<u8>, String)> = vec![];
+        let results = batch_extract_bytes_sync(contents, &config);
+
+        assert!(results.is_ok());
+        assert_eq!(results.unwrap().len(), 0);
+    }
+
+    #[tokio::test]
+    async fn test_concurrent_extractions_same_mime() {
+        use tokio::task::JoinSet;
+
+        let config = Arc::new(ExtractionConfig::default());
+        let mut tasks = JoinSet::new();
+
+        for i in 0..50 {
+            let config_clone = Arc::clone(&config);
+            tasks.spawn(async move {
+                let content = format!("test content {}", i);
+                extract_bytes(content.as_bytes(), "text/plain", &config_clone).await
+            });
+        }
+
+        let mut success_count = 0;
+        while let Some(task_result) = tasks.join_next().await {
+            if let Ok(Ok(_)) = task_result {
+                success_count += 1;
+            }
+        }
+
+        assert_eq!(success_count, 50);
+    }
+
+    #[serial]
+    #[tokio::test]
+    async fn test_concurrent_extractions_different_mimes() {
+        use tokio::task::JoinSet;
+
+        let config = Arc::new(ExtractionConfig::default());
+        let mut tasks = JoinSet::new();
+
+        let mime_types = ["text/plain", "text/markdown"];
+
+        for i in 0..30 {
+            let config_clone = Arc::clone(&config);
+            let mime = mime_types[i % mime_types.len()];
+            tasks.spawn(async move {
+                let content = format!("test {}", i);
+                extract_bytes(content.as_bytes(), mime, &config_clone).await
+            });
+        }
+
+        let mut success_count = 0;
+        while let Some(task_result) = tasks.join_next().await {
+            if let Ok(Ok(_)) = task_result {
+                success_count += 1;
+            }
+        }
+
+        assert_eq!(success_count, 30);
+    }
+}

data/vendor/kreuzberg/src/core/extractor/sync.rs

@@ -0,0 +1,208 @@
+//! Synchronous wrappers for extraction operations.
+//!
+//! This module provides blocking synchronous wrappers around async extraction functions
+//! for use in non-async contexts. Uses a global Tokio runtime for optimal performance.
+
+use crate::Result;
+use crate::core::config::ExtractionConfig;
+use crate::types::ExtractionResult;
+
+#[cfg(feature = "tokio-runtime")]
+use std::path::Path;
+
+#[cfg(feature = "tokio-runtime")]
+use once_cell::sync::Lazy;
+
+#[cfg(feature = "tokio-runtime")]
+use super::batch::{batch_extract_bytes, batch_extract_file};
+#[cfg(feature = "tokio-runtime")]
+use super::bytes::extract_bytes;
+#[cfg(feature = "tokio-runtime")]
+use super::file::extract_file;
+
+/// Global Tokio runtime for synchronous operations.
+///
+/// This runtime is lazily initialized on first use and shared across all sync wrappers.
+/// Using a global runtime instead of creating one per call provides 100x+ performance improvement.
+///
+/// # Safety
+///
+/// The `.expect()` here is justified because:
+/// 1. Runtime creation can only fail due to system resource exhaustion (OOM, thread limit)
+/// 2. If runtime creation fails, the process is already in a critical state
+/// 3. This is a one-time initialization - if it fails, nothing will work
+/// 4. Better to fail fast than return errors from every sync operation
+///
+/// # Availability
+///
+/// This static is only available when the `tokio-runtime` feature is enabled.
+/// For WASM targets, use the truly synchronous extraction functions instead.
+#[cfg(feature = "tokio-runtime")]
+static GLOBAL_RUNTIME: Lazy<tokio::runtime::Runtime> = Lazy::new(|| {
+    tokio::runtime::Builder::new_multi_thread()
+        .enable_all()
+        .build()
+        .expect("Failed to create global Tokio runtime - system may be out of resources")
+});
+
+/// Synchronous wrapper for `extract_file`.
+///
+/// This is a convenience function that blocks the current thread until extraction completes.
+/// For async code, use `extract_file` directly.
+///
+/// Uses the global Tokio runtime for 100x+ performance improvement over creating
+/// a new runtime per call. Always uses the global runtime to avoid nested runtime issues.
+///
+/// This function is only available with the `tokio-runtime` feature. For WASM targets,
+/// use a truly synchronous extraction approach instead.
+///
+/// # Example
+///
+/// ```rust,no_run
+/// use kreuzberg::core::extractor::extract_file_sync;
+/// use kreuzberg::core::config::ExtractionConfig;
+///
+/// let config = ExtractionConfig::default();
+/// let result = extract_file_sync("document.pdf", None, &config)?;
+/// println!("Content: {}", result.content);
+/// # Ok::<(), kreuzberg::KreuzbergError>(())
+/// ```
+#[cfg(feature = "tokio-runtime")]
+pub fn extract_file_sync(
+    path: impl AsRef<Path>,
+    mime_type: Option<&str>,
+    config: &ExtractionConfig,
+) -> Result<ExtractionResult> {
+    GLOBAL_RUNTIME.block_on(extract_file(path, mime_type, config))
+}
+
+/// Synchronous wrapper for `extract_bytes`.
+///
+/// Uses the global Tokio runtime for 100x+ performance improvement over creating
+/// a new runtime per call.
+///
+/// With the `tokio-runtime` feature, this blocks the current thread using the global
+/// Tokio runtime. Without it (WASM), this calls a truly synchronous implementation.
+///
+/// # Example
+///
+/// ```rust,no_run
+/// use kreuzberg::core::extractor::extract_bytes_sync;
+/// use kreuzberg::core::config::ExtractionConfig;
+///
+/// let config = ExtractionConfig::default();
+/// let bytes = b"Hello, world!";
+/// let result = extract_bytes_sync(bytes, "text/plain", &config)?;
+/// println!("Content: {}", result.content);
+/// # Ok::<(), kreuzberg::KreuzbergError>(())
+/// ```
+#[cfg(feature = "tokio-runtime")]
+pub fn extract_bytes_sync(content: &[u8], mime_type: &str, config: &ExtractionConfig) -> Result<ExtractionResult> {
+    GLOBAL_RUNTIME.block_on(extract_bytes(content, mime_type, config))
+}
+
+/// Synchronous wrapper for `extract_bytes` (WASM-compatible version).
+///
+/// This is a truly synchronous implementation without tokio runtime dependency.
+/// It calls `extract_bytes_sync_impl()` to perform the extraction.
+#[cfg(not(feature = "tokio-runtime"))]
+pub fn extract_bytes_sync(content: &[u8], mime_type: &str, config: &ExtractionConfig) -> Result<ExtractionResult> {
+    super::legacy::extract_bytes_sync_impl(content.to_vec(), Some(mime_type.to_string()), Some(config.clone()))
+}
+
+/// Synchronous wrapper for `batch_extract_file`.
+///
+/// Uses the global Tokio runtime for 100x+ performance improvement over creating
+/// a new runtime per call.
+///
+/// This function is only available with the `tokio-runtime` feature. For WASM targets,
+/// use a truly synchronous extraction approach instead.
+///
+/// # Example
+///
+/// ```rust,no_run
+/// use kreuzberg::core::extractor::batch_extract_file_sync;
+/// use kreuzberg::core::config::ExtractionConfig;
+///
+/// let config = ExtractionConfig::default();
+/// let paths = vec!["doc1.pdf", "doc2.pdf"];
+/// let results = batch_extract_file_sync(paths, &config)?;
+/// println!("Processed {} files", results.len());
+/// # Ok::<(), kreuzberg::KreuzbergError>(())
+/// ```
+#[cfg(feature = "tokio-runtime")]
+pub fn batch_extract_file_sync(
+    paths: Vec<impl AsRef<Path>>,
+    config: &ExtractionConfig,
+) -> Result<Vec<ExtractionResult>> {
+    GLOBAL_RUNTIME.block_on(batch_extract_file(paths, config))
+}
+
+/// Synchronous wrapper for `batch_extract_bytes`.
+///
+/// Uses the global Tokio runtime for 100x+ performance improvement over creating
+/// a new runtime per call.
+///
+/// With the `tokio-runtime` feature, this blocks the current thread using the global
+/// Tokio runtime. Without it (WASM), this calls a truly synchronous implementation
+/// that iterates through items and calls `extract_bytes_sync()`.
+///
+/// # Example
+///
+/// ```rust,no_run
+/// use kreuzberg::core::extractor::batch_extract_bytes_sync;
+/// use kreuzberg::core::config::ExtractionConfig;
+///
+/// let config = ExtractionConfig::default();
+/// let contents = vec![
+///     (b"content 1".to_vec(), "text/plain".to_string()),
+///     (b"content 2".to_vec(), "text/plain".to_string()),
+/// ];
+/// let results = batch_extract_bytes_sync(contents, &config)?;
+/// println!("Processed {} items", results.len());
+/// # Ok::<(), kreuzberg::KreuzbergError>(())
+/// ```
+#[cfg(feature = "tokio-runtime")]
+pub fn batch_extract_bytes_sync(
+    contents: Vec<(Vec<u8>, String)>,
+    config: &ExtractionConfig,
+) -> Result<Vec<ExtractionResult>> {
+    GLOBAL_RUNTIME.block_on(batch_extract_bytes(contents, config))
+}
+
+/// Synchronous wrapper for `batch_extract_bytes` (WASM-compatible version).
+///
+/// This is a truly synchronous implementation that iterates through items
+/// and calls `extract_bytes_sync()` for each.
+#[cfg(not(feature = "tokio-runtime"))]
+pub fn batch_extract_bytes_sync(
+    contents: Vec<(Vec<u8>, String)>,
+    config: &ExtractionConfig,
+) -> Result<Vec<ExtractionResult>> {
+    use crate::types::{ErrorMetadata, Metadata};
+    use crate::utils::intern_mime_type;
+
+    let mut results = Vec::with_capacity(contents.len());
+    for (content, mime_type) in contents {
+        let result = extract_bytes_sync(&content, &mime_type, config);
+        results.push(result.unwrap_or_else(|e| ExtractionResult {
+            content: format!("Error: {}", e),
+            mime_type: intern_mime_type("text/plain").to_string(),
+            metadata: Metadata {
+                error: Some(ErrorMetadata {
+                    error_type: format!("{:?}", e),
+                    message: e.to_string(),
+                }),
+                ..Default::default()
+            },
+            tables: vec![],
+            detected_languages: None,
+            chunks: None,
+            images: None,
+            djot_content: None,
+            pages: None,
+            elements: None,
+        }));
+    }
+    Ok(results)
+}