fix(tests): update test document paths after directory refactoring

Update test file paths across all language bindings to reflect the
reorganized test_documents directory structure:

- pdfs_with_tables/ → pdf/
- pdfs/ → pdf/
- tables/ → images/
- legacy_office/ → doc/
- office/ → docx/
- orgmode/ → org/
- web/ → html/
- pandoc/ → markdown/
- misc/ → org/, epub/
- data_formats/ → json/, yaml/
- documents/ → docx/

Fix broken symlinks in test_documents/ directory.
Add validate_ground_truth.py script for benchmark validation.

Affected bindings: Rust, TypeScript, C#, PHP, Ruby, Python, Go, Java,
Elixir, WASM

Author: Goldziher

crates/kreuzberg-node/tests/binding/batch-operations.spec.ts

@@ -34,7 +34,7 @@ let sampleTxtBytes: Uint8Array;
 
 beforeAll(() => {
 	// Get test documents
-	samplePdfPath = getTestDocumentPath("pdfs/tiny.pdf");
+	samplePdfPath = getTestDocumentPath("pdf/tiny.pdf");
 	sampleDocxPath = getTestDocumentPath("documents/sample.docx");
 	sampleTxtPath = getTestDocumentPath("text/sample.txt");
 

crates/kreuzberg-node/tests/binding/extraction-functions.spec.ts

@@ -19,7 +19,7 @@ beforeAll(() => {
 	// Resolve symlinks to get the actual file path (important for Windows compatibility)
 	samplePdfBytes = new Uint8Array(readFileSync(realpathSync(samplePdfPath)));
 
-	sampleTextPath = getTestDocumentPath("pandoc/simple_metadata.md");
+	sampleTextPath = getTestDocumentPath("markdown/simple_metadata.md");
 	sampleTextBytes = new Uint8Array(readFileSync(realpathSync(sampleTextPath)));
 });
 

crates/kreuzberg-node/tests/binding/helpers.spec.ts

@@ -168,7 +168,7 @@ describe("Helper Functions and Edge Cases", () => {
 
 	describe("Edge cases", () => {
 		it("should handle very small files", () => {
-			const textPath = getTestDocumentPath("pandoc/simple_metadata.md");
+			const textPath = getTestDocumentPath("markdown/simple_metadata.md");
 			const result = extractFileSync(textPath, null, null);
 
 			expect(result.content).toBeTruthy();

crates/kreuzberg-node/tests/binding/images.spec.ts

@@ -26,7 +26,7 @@ let pptxPath: string;
 let docxPath: string;
 
 beforeAll(() => {
-	samplePdfPath = getTestDocumentPath("pdfs/embedded_images_tables.pdf");
+	samplePdfPath = getTestDocumentPath("pdf/embedded_images_tables.pdf");
 	// Resolve symlinks to get the actual file path (important for Windows compatibility)
 	samplePdfBytes = new Uint8Array(readFileSync(realpathSync(samplePdfPath)));
 

crates/kreuzberg-node/tests/binding/metadata-types.spec.ts

@@ -392,8 +392,8 @@ describe("Metadata Types - Type Compatibility Tests", () => {
 		});
 
 		it("test_extract_html_file_integration", () => {
-			const htmlPath = getTestDocumentPath("web/taylor_swift.html");
-			const buffer = loadTestDocument("web/taylor_swift.html");
+			const htmlPath = getTestDocumentPath("html/taylor_swift.html");
+			const buffer = loadTestDocument("html/taylor_swift.html");
 
 			const result = extractBytesSync(buffer, "text/html", null);
 

crates/kreuzberg-node/tests/binding/sync-async-coverage.spec.ts

@@ -37,7 +37,7 @@ beforeAll(() => {
 	const resolvedPdfPath = realpathSync(samplePdfPath);
 	samplePdfBytes = new Uint8Array(readFileSync(resolvedPdfPath));
 
-	sampleTextPath = getTestDocumentPath("pandoc/simple_metadata.md");
+	sampleTextPath = getTestDocumentPath("markdown/simple_metadata.md");
 	const resolvedTextPath = realpathSync(sampleTextPath);
 	sampleTextBytes = new Uint8Array(readFileSync(resolvedTextPath));
 });

crates/kreuzberg-node/tests/binding/tables.spec.ts

@@ -29,9 +29,9 @@ let tinyPdfBytes: Uint8Array;
 
 beforeAll(() => {
 	// Use PDFs with tables for testing
-	tinyPdfPath = getTestDocumentPath("pdfs_with_tables/tiny.pdf");
-	mediumPdfPath = getTestDocumentPath("pdfs_with_tables/medium.pdf");
-	largePdfPath = getTestDocumentPath("pdfs_with_tables/large.pdf");
+	tinyPdfPath = getTestDocumentPath("pdf/tiny.pdf");
+	mediumPdfPath = getTestDocumentPath("pdf/medium.pdf");
+	largePdfPath = getTestDocumentPath("pdf/large.pdf");
 
 	try {
 		// Resolve symlinks to get the actual file path (important for Windows compatibility)

crates/kreuzberg-tesseract/tests/integration_test.rs

@@ -108,7 +108,7 @@ fn test_ocr_on_table_image() {
     api.set_variable("tessedit_pageseg_mode", "1")
         .expect("Failed to set PSM");
 
-    let (image_data, width, height) = load_test_image("tables/simple_table.png").expect("Failed to load test image");
+    let (image_data, width, height) = load_test_image("images/simple_table.png").expect("Failed to load test image");
     api.set_image(&image_data, width as i32, height as i32, 3, 3 * width as i32)
         .expect("Failed to set image");
 

crates/kreuzberg-wasm/typescript/pages.spec.ts

@@ -22,7 +22,7 @@ beforeAll(async () => {
 	await initWasm();
 
 	// Load test PDF file (path relative to crates/kreuzberg-wasm)
-	const pdfPath = join(process.cwd(), "../../test_documents/pdfs/embedded_images_tables.pdf");
+	const pdfPath = join(process.cwd(), "../../test_documents/pdf/embedded_images_tables.pdf");
 	try {
 		samplePdfBytes = new Uint8Array(readFileSync(pdfPath));
 	} catch {

crates/kreuzberg-wasm/typescript/tables.spec.ts

@@ -15,7 +15,7 @@ let samplePdfBytes: Uint8Array;
 beforeAll(async () => {
 	await initWasm();
 	// Load test PDF file (path relative to crates/kreuzberg-wasm)
-	const pdfPath = join(process.cwd(), "../../test_documents/pdfs/embedded_images_tables.pdf");
+	const pdfPath = join(process.cwd(), "../../test_documents/pdf/embedded_images_tables.pdf");
 	try {
 		samplePdfBytes = new Uint8Array(readFileSync(pdfPath));
 	} catch {