kreuzberg 3.10.1__tar.gz → 3.11.0__tar.gz

{kreuzberg-3.10.1 → kreuzberg-3.11.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: kreuzberg
-Version: 3.10.1
+Version: 3.11.0
 Summary: Document intelligence framework for Python - Extract text, metadata, and structured data from diverse file formats
 Project-URL: documentation, https://kreuzberg.dev
 Project-URL: homepage, https://github.com/Goldziher/kreuzberg
@@ -32,7 +32,7 @@ Requires-Dist: anyio>=4.9.0
 Requires-Dist: chardetng-py>=0.3.5
 Requires-Dist: exceptiongroup>=1.2.2; python_version < '3.11'
 Requires-Dist: html-to-markdown[lxml]>=1.9.0
-Requires-Dist: mcp>=1.12.2
+Requires-Dist: mcp>=1.12.3
 Requires-Dist: msgspec>=0.18.0
 Requires-Dist: playa-pdf>=0.6.4
 Requires-Dist: psutil>=7.0.0
@@ -45,6 +45,7 @@ Requires-Dist: mailparse>=1.0.15; extra == 'additional-extensions'
 Requires-Dist: tomli>=2.0.0; (python_version < '3.11') and extra == 'additional-extensions'
 Provides-Extra: all
 Requires-Dist: click>=8.2.1; extra == 'all'
+Requires-Dist: deep-translator>=1.11.4; extra == 'all'
 Requires-Dist: easyocr>=1.7.2; extra == 'all'
 Requires-Dist: fast-langdetect>=0.3.2; extra == 'all'
 Requires-Dist: gmft>=0.4.2; extra == 'all'
@@ -53,6 +54,7 @@ Requires-Dist: litestar[opentelemetry,standard,structlog]>=2.16.0; extra == 'all
 Requires-Dist: mailparse>=1.0.15; extra == 'all'
 Requires-Dist: paddleocr>=3.1.0; extra == 'all'
 Requires-Dist: paddlepaddle>=3.1.0; extra == 'all'
+Requires-Dist: pandas>=2.3.1; extra == 'all'
 Requires-Dist: playa-pdf[crypto]>=0.6.4; extra == 'all'
 Requires-Dist: rich>=14.1.0; extra == 'all'
 Requires-Dist: semantic-text-splitter>=0.27.0; extra == 'all'
@@ -61,9 +63,6 @@ Requires-Dist: spacy>=3.8.7; extra == 'all'
 Requires-Dist: tomli>=2.0.0; (python_version < '3.11') and extra == 'all'
 Provides-Extra: api
 Requires-Dist: litestar[opentelemetry,standard,structlog]>=2.16.0; extra == 'api'
-Provides-Extra: auto-classify-document-type
-Requires-Dist: deep-translator>=1.11.4; extra == 'auto-classify-document-type'
-Requires-Dist: pandas>=2.3.1; extra == 'auto-classify-document-type'
 Provides-Extra: chunking
 Requires-Dist: semantic-text-splitter>=0.27.0; extra == 'chunking'
 Provides-Extra: cli
@@ -72,6 +71,9 @@ Requires-Dist: rich>=14.1.0; extra == 'cli'
 Requires-Dist: tomli>=2.0.0; (python_version < '3.11') and extra == 'cli'
 Provides-Extra: crypto
 Requires-Dist: playa-pdf[crypto]>=0.6.4; extra == 'crypto'
+Provides-Extra: document-classification
+Requires-Dist: deep-translator>=1.11.4; extra == 'document-classification'
+Requires-Dist: pandas>=2.3.1; extra == 'document-classification'
 Provides-Extra: easyocr
 Requires-Dist: easyocr>=1.7.2; extra == 'easyocr'
 Provides-Extra: entity-extraction

{kreuzberg-3.10.1 → kreuzberg-3.11.0}/docs/contributing.md

@@ -34,7 +34,7 @@ All commands run through `uv run`:
 # Testing
 uv run pytest                      # Run all tests
 uv run pytest tests/foo_test.py    # Run specific test
-uv run pytest --cov                # With coverage (must be ≥95%)
+uv run pytest --cov                # With coverage (must be ≥85%)
 
 # Code quality
 uv run ruff format                 # Format code

{kreuzberg-3.10.1 → kreuzberg-3.11.0}/docs/getting-started/installation.md

@@ -134,6 +134,16 @@ python -m spacy download es_core_news_sm  # Spanish
 
     spaCy language models are large (50-500MB each) and are downloaded separately. Only download the models for languages you actually need to process. See the [spaCy models documentation](https://spacy.io/models) for a complete list of available models.
 
+### Document Classification
+
+For automatic document type detection (invoice, contract, receipt, etc.), install the document classification extra:
+
+```shell
+pip install "kreuzberg[document-classification]"
+```
+
+This feature uses Google Translate for multi-language support and requires explicit opt-in by setting `auto_detect_document_type=True` in your configuration.
+
 ### All Optional Dependencies
 
 To install Kreuzberg with all optional dependencies, you can use the `all` extra group:
@@ -145,5 +155,5 @@ pip install "kreuzberg[all]"
 This is equivalent to:
 
 ```shell
-pip install "kreuzberg[chunking,easyocr,entity-extraction,gmft,langdetect,paddleocr]"
+pip install "kreuzberg[chunking,document-classification,easyocr,entity-extraction,gmft,langdetect,paddleocr]"
 ```

{kreuzberg-3.10.1 → kreuzberg-3.11.0}/docs/index.md

@@ -22,7 +22,7 @@ Kreuzberg addresses the complete document intelligence pipeline through a modula
 
 ### Engineering Principles
 
-- **Test Coverage**: 95%+ coverage with comprehensive test suites
+- **Test Coverage**: Comprehensive test suites ensuring code reliability
 - **API Design**: True async/await implementation alongside synchronous APIs
 - **Error Handling**: Consistent exception hierarchy with detailed context
 - **Type Safety**: Full type annotations for enhanced developer experience

{kreuzberg-3.10.1 → kreuzberg-3.11.0}/docs/user-guide/document-classification.md

@@ -2,9 +2,17 @@
 
 Kreuzberg can automatically classify documents into common types like invoices, contracts, and receipts. This allows you to build custom processing pipelines tailored to each document type.
 
+## Installation
+
+Document classification requires the `document-classification` extra to be installed:
+
+```bash
+pip install "kreuzberg[document-classification]"
+```
+
 ## Enabling Document Classification
 
-To enable this feature, set `auto_detect_document_type=True` in your `ExtractionConfig`:
+Document classification is disabled by default. To enable this feature, set `auto_detect_document_type=True` in your `ExtractionConfig`:
 
 ```python
 from kreuzberg import ExtractionConfig, extract_file

{kreuzberg-3.10.1 → kreuzberg-3.11.0}/kreuzberg/_config.py

@@ -97,19 +97,21 @@ def parse_ocr_backend_config(
     if not isinstance(backend_config, dict):
         return None
 
-    if backend == "tesseract":
-        # Convert psm integer to PSMMode enum if needed
-        processed_config = backend_config.copy()
-        if "psm" in processed_config and isinstance(processed_config["psm"], int):
-            from kreuzberg._ocr._tesseract import PSMMode  # noqa: PLC0415
-
-            processed_config["psm"] = PSMMode(processed_config["psm"])
-        return TesseractConfig(**processed_config)
-    if backend == "easyocr":
-        return EasyOCRConfig(**backend_config)
-    if backend == "paddleocr":
-        return PaddleOCRConfig(**backend_config)
-    return None
+    match backend:
+        case "tesseract":
+            # Convert psm integer to PSMMode enum if needed
+            processed_config = backend_config.copy()
+            if "psm" in processed_config and isinstance(processed_config["psm"], int):
+                from kreuzberg._ocr._tesseract import PSMMode  # noqa: PLC0415
+
+                processed_config["psm"] = PSMMode(processed_config["psm"])
+            return TesseractConfig(**processed_config)
+        case "easyocr":
+            return EasyOCRConfig(**backend_config)
+        case "paddleocr":
+            return PaddleOCRConfig(**backend_config)
+        case _:
+            return None
 
 
 def build_extraction_config_from_dict(config_dict: dict[str, Any]) -> ExtractionConfig:
@@ -140,7 +142,9 @@ def build_extraction_config_from_dict(config_dict: dict[str, Any]) -> Extraction
         "document_classification_mode",
         "keyword_count",
     }
-    extraction_config.update({field: config_dict[field] for field in basic_fields if field in config_dict})
+    extraction_config = extraction_config | {
+        field: config_dict[field] for field in basic_fields if field in config_dict
+    }
 
     # Handle OCR backend configuration
     ocr_backend = extraction_config.get("ocr_backend")

{kreuzberg-3.10.1 → kreuzberg-3.11.0}/kreuzberg/_document_classification.py

@@ -62,7 +62,7 @@ def _get_translated_text(result: ExtractionResult) -> str:
         from deep_translator import GoogleTranslator  # noqa: PLC0415
     except ImportError as e:  # pragma: no cover
         raise MissingDependencyError(
-            "The 'deep-translator' library is not installed. Please install it with: pip install 'kreuzberg[auto-classify-document-type]'"
+            "The 'deep-translator' library is not installed. Please install it with: pip install 'kreuzberg[document-classification]'"
         ) from e
 
     try:

{kreuzberg-3.10.1 → kreuzberg-3.11.0}/kreuzberg/_extractors/_base.py

@@ -116,8 +116,7 @@ class Extractor(ABC):
         quality_score = calculate_quality_score(cleaned_content, dict(result.metadata) if result.metadata else None)
 
         # Add quality metadata
-        enhanced_metadata = dict(result.metadata) if result.metadata else {}
-        enhanced_metadata["quality_score"] = quality_score
+        enhanced_metadata = (dict(result.metadata) if result.metadata else {}) | {"quality_score": quality_score}
 
         # Return enhanced result
         return ExtractionResult(

{kreuzberg-3.10.1 → kreuzberg-3.11.0}/kreuzberg/_extractors/_image.py

@@ -85,23 +85,24 @@ class ImageExtractor(Extractor):
 
         backend = get_ocr_backend(self.config.ocr_backend)
 
-        if self.config.ocr_backend == "tesseract":
-            config = (
-                self.config.ocr_config if isinstance(self.config.ocr_config, TesseractConfig) else TesseractConfig()
-            )
-            result = backend.process_file_sync(path, **asdict(config))
-        elif self.config.ocr_backend == "paddleocr":
-            paddle_config = (
-                self.config.ocr_config if isinstance(self.config.ocr_config, PaddleOCRConfig) else PaddleOCRConfig()
-            )
-            result = backend.process_file_sync(path, **asdict(paddle_config))
-        elif self.config.ocr_backend == "easyocr":
-            easy_config = (
-                self.config.ocr_config if isinstance(self.config.ocr_config, EasyOCRConfig) else EasyOCRConfig()
-            )
-            result = backend.process_file_sync(path, **asdict(easy_config))
-        else:
-            raise NotImplementedError(f"Sync OCR not implemented for {self.config.ocr_backend}")
+        match self.config.ocr_backend:
+            case "tesseract":
+                config = (
+                    self.config.ocr_config if isinstance(self.config.ocr_config, TesseractConfig) else TesseractConfig()
+                )
+                result = backend.process_file_sync(path, **asdict(config))
+            case "paddleocr":
+                paddle_config = (
+                    self.config.ocr_config if isinstance(self.config.ocr_config, PaddleOCRConfig) else PaddleOCRConfig()
+                )
+                result = backend.process_file_sync(path, **asdict(paddle_config))
+            case "easyocr":
+                easy_config = (
+                    self.config.ocr_config if isinstance(self.config.ocr_config, EasyOCRConfig) else EasyOCRConfig()
+                )
+                result = backend.process_file_sync(path, **asdict(easy_config))
+            case _:
+                raise NotImplementedError(f"Sync OCR not implemented for {self.config.ocr_backend}")
         return self._apply_quality_processing(result)
 
     def _get_extension_from_mime_type(self, mime_type: str) -> str:

{kreuzberg-3.10.1 → kreuzberg-3.11.0}/kreuzberg/_extractors/_pdf.py

@@ -88,14 +88,12 @@ class PDFExtractor(Extractor):
             # Enhance metadata with table information
             if result.tables:
                 table_summary = generate_table_summary(result.tables)
-                result.metadata.update(
-                    {
-                        "table_count": table_summary["table_count"],
-                        "tables_summary": f"Document contains {table_summary['table_count']} tables "
-                        f"across {table_summary['pages_with_tables']} pages with "
-                        f"{table_summary['total_rows']} total rows",
-                    }
-                )
+                result.metadata = result.metadata | {
+                    "table_count": table_summary["table_count"],
+                    "tables_summary": f"Document contains {table_summary['table_count']} tables "
+                    f"across {table_summary['pages_with_tables']} pages with "
+                    f"{table_summary['total_rows']} total rows",
+                }
 
         return self._apply_quality_processing(result)
 
@@ -153,14 +151,12 @@ class PDFExtractor(Extractor):
         # Enhance metadata with table information
         if tables:
             table_summary = generate_table_summary(tables)
-            result.metadata.update(
-                {
-                    "table_count": table_summary["table_count"],
-                    "tables_summary": f"Document contains {table_summary['table_count']} tables "
-                    f"across {table_summary['pages_with_tables']} pages with "
-                    f"{table_summary['total_rows']} total rows",
-                }
-            )
+            result.metadata = result.metadata | {
+                "table_count": table_summary["table_count"],
+                "tables_summary": f"Document contains {table_summary['table_count']} tables "
+                f"across {table_summary['pages_with_tables']} pages with "
+                f"{table_summary['total_rows']} total rows",
+            }
 
         # Apply quality processing
         return self._apply_quality_processing(result)
@@ -386,23 +382,24 @@ class PDFExtractor(Extractor):
         backend = get_ocr_backend(self.config.ocr_backend)
         paths = [Path(p) for p in image_paths]
 
-        if self.config.ocr_backend == "tesseract":
-            config = (
-                self.config.ocr_config if isinstance(self.config.ocr_config, TesseractConfig) else TesseractConfig()
-            )
-            results = backend.process_batch_sync(paths, **asdict(config))
-        elif self.config.ocr_backend == "paddleocr":
-            paddle_config = (
-                self.config.ocr_config if isinstance(self.config.ocr_config, PaddleOCRConfig) else PaddleOCRConfig()
-            )
-            results = backend.process_batch_sync(paths, **asdict(paddle_config))
-        elif self.config.ocr_backend == "easyocr":
-            easy_config = (
-                self.config.ocr_config if isinstance(self.config.ocr_config, EasyOCRConfig) else EasyOCRConfig()
-            )
-            results = backend.process_batch_sync(paths, **asdict(easy_config))
-        else:
-            raise NotImplementedError(f"Sync OCR not implemented for {self.config.ocr_backend}")
+        match self.config.ocr_backend:
+            case "tesseract":
+                config = (
+                    self.config.ocr_config if isinstance(self.config.ocr_config, TesseractConfig) else TesseractConfig()
+                )
+                results = backend.process_batch_sync(paths, **asdict(config))
+            case "paddleocr":
+                paddle_config = (
+                    self.config.ocr_config if isinstance(self.config.ocr_config, PaddleOCRConfig) else PaddleOCRConfig()
+                )
+                results = backend.process_batch_sync(paths, **asdict(paddle_config))
+            case "easyocr":
+                easy_config = (
+                    self.config.ocr_config if isinstance(self.config.ocr_config, EasyOCRConfig) else EasyOCRConfig()
+                )
+                results = backend.process_batch_sync(paths, **asdict(easy_config))
+            case _:
+                raise NotImplementedError(f"Sync OCR not implemented for {self.config.ocr_backend}")
 
         # Use list comprehension and join for efficient string building
         return "\n\n".join(result.content for result in results)

{kreuzberg-3.10.1 → kreuzberg-3.11.0}/kreuzberg/_mcp/server.py

@@ -51,7 +51,7 @@ def _create_config_with_overrides(**kwargs: Any) -> ExtractionConfig:
     }
 
     # Override with provided parameters
-    config_dict.update(kwargs)
+    config_dict = config_dict | kwargs
 
     return ExtractionConfig(**config_dict)
 

{kreuzberg-3.10.1 → kreuzberg-3.11.0}/kreuzberg/_types.py

@@ -349,7 +349,7 @@ class ExtractionConfig:
     """Configuration for language detection. If None, uses default settings."""
     spacy_entity_extraction_config: SpacyEntityExtractionConfig | None = None
     """Configuration for spaCy entity extraction. If None, uses default settings."""
-    auto_detect_document_type: bool = True
+    auto_detect_document_type: bool = False
     """Whether to automatically detect the document type."""
     document_type_confidence_threshold: float = 0.5
     """Confidence threshold for document type detection."""
@@ -398,15 +398,16 @@ class ExtractionConfig:
             return asdict(self.ocr_config)
 
         # Lazy load and cache default configs instead of creating new instances
-        if self.ocr_backend == "tesseract":
-            from kreuzberg._ocr._tesseract import TesseractConfig  # noqa: PLC0415
+        match self.ocr_backend:
+            case "tesseract":
+                from kreuzberg._ocr._tesseract import TesseractConfig  # noqa: PLC0415
 
-            return asdict(TesseractConfig())
-        if self.ocr_backend == "easyocr":
-            from kreuzberg._ocr._easyocr import EasyOCRConfig  # noqa: PLC0415
+                return asdict(TesseractConfig())
+            case "easyocr":
+                from kreuzberg._ocr._easyocr import EasyOCRConfig  # noqa: PLC0415
 
-            return asdict(EasyOCRConfig())
-        # paddleocr
-        from kreuzberg._ocr._paddleocr import PaddleOCRConfig  # noqa: PLC0415
+                return asdict(EasyOCRConfig())
+            case _:  # paddleocr or any other backend
+                from kreuzberg._ocr._paddleocr import PaddleOCRConfig  # noqa: PLC0415
 
-        return asdict(PaddleOCRConfig())
+                return asdict(PaddleOCRConfig())

{kreuzberg-3.10.1 → kreuzberg-3.11.0}/mkdocs.yaml

@@ -158,4 +158,3 @@ nav:
       - Custom Hooks: advanced/custom-hooks.md
       - Custom Extractors: advanced/custom-extractors.md
   - Contributing: contributing.md
-  - Changelog: changelog.md

{kreuzberg-3.10.1 → kreuzberg-3.11.0}/pyproject.toml

@@ -5,7 +5,7 @@ requires = [ "hatchling" ]
 
 [project]
 name = "kreuzberg"
-version = "3.10.1"
+version = "3.11.0"
 description = "Document intelligence framework for Python - Extract text, metadata, and structured data from diverse file formats"
 readme = "README.md"
 keywords = [
@@ -61,7 +61,7 @@ dependencies = [
   "chardetng-py>=0.3.5",
   "exceptiongroup>=1.2.2; python_version<'3.11'",
   "html-to-markdown[lxml]>=1.9.0",
-  "mcp>=1.12.2",
+  "mcp>=1.12.3",
   "msgspec>=0.18.0",
   "playa-pdf>=0.6.4",                                 # pinned due to breaking changes in 0.5.0
   "psutil>=7.0.0",
@@ -76,15 +76,11 @@ optional-dependencies.additional-extensions = [
   "tomli>=2.0.0; python_version<'3.11'",
 ]
 optional-dependencies.all = [
-  "kreuzberg[additional-extensions,api,chunking,cli,crypto,easyocr,entity-extraction,gmft,langdetect,paddleocr]",
+  "kreuzberg[additional-extensions,api,chunking,cli,crypto,document-classification,easyocr,entity-extraction,gmft,langdetect,paddleocr]",
 ]
 optional-dependencies.api = [
   "litestar[standard,structlog,opentelemetry]>=2.16.0",
 ]
-optional-dependencies.auto-classify-document-type = [
-  "deep-translator>=1.11.4",
-  "pandas>=2.3.1",
-]
 optional-dependencies.chunking = [ "semantic-text-splitter>=0.27.0" ]
 optional-dependencies.cli = [
   "click>=8.2.1",
@@ -92,6 +88,10 @@ optional-dependencies.cli = [
   "tomli>=2.0.0; python_version<'3.11'",
 ]
 optional-dependencies.crypto = [ "playa-pdf[crypto]>=0.6.4" ]
+optional-dependencies.document-classification = [
+  "deep-translator>=1.11.4",
+  "pandas>=2.3.1",
+]
 optional-dependencies.easyocr = [ "easyocr>=1.7.2" ]
 optional-dependencies.entity-extraction = [ "keybert>=0.9.0", "spacy>=3.8.7" ]
 optional-dependencies.gmft = [ "gmft>=0.4.2" ]
@@ -256,7 +256,7 @@ exclude_lines = [
   "class .*\\bProtocol\\):",
   "@(abc\\.)?abstractmethod",
 ]
-fail_under = 95
+fail_under = 85
 
 [tool.mypy]
 packages = [ "kreuzberg", "tests", "benchmarks.src.kreuzberg_benchmarks" ]

{kreuzberg-3.10.1 → kreuzberg-3.11.0}/tests/document_classification_test.py

@@ -2,8 +2,10 @@
 
 from __future__ import annotations
 
+import builtins
+import sys
 from pathlib import Path
-from typing import TYPE_CHECKING
+from typing import TYPE_CHECKING, Any
 
 import pandas as pd
 import pytest
@@ -15,6 +17,7 @@ from kreuzberg._document_classification import (
     classify_document_from_layout,
 )
 from kreuzberg._types import ExtractionConfig, ExtractionResult
+from kreuzberg.exceptions import MissingDependencyError
 
 if TYPE_CHECKING:
     from pytest_mock import MockerFixture
@@ -112,7 +115,7 @@ def test_classify_document_with_metadata() -> None:
         mime_type="text/plain",
         metadata={"title": "Invoice #12345", "subject": "Payment Due"},
     )
-    config = ExtractionConfig()
+    config = ExtractionConfig(auto_detect_document_type=True)
 
     doc_type, confidence = classify_document(result, config)
 
@@ -142,7 +145,7 @@ def test_classify_document_empty_content() -> None:
         mime_type="text/plain",
         metadata={},
     )
-    config = ExtractionConfig()
+    config = ExtractionConfig(auto_detect_document_type=True)
 
     doc_type, confidence = classify_document(result, config)
 
@@ -158,7 +161,7 @@ def test_classify_document_with_exclusions() -> None:
         mime_type="text/plain",
         metadata={},
     )
-    config = ExtractionConfig()
+    config = ExtractionConfig(auto_detect_document_type=True)
 
     doc_type, confidence = classify_document(result, config)
 
@@ -184,7 +187,7 @@ def test_classify_document_from_layout_basic() -> None:
         metadata={},
         layout=layout_df,
     )
-    config = ExtractionConfig()
+    config = ExtractionConfig(auto_detect_document_type=True)
 
     doc_type, confidence = classify_document_from_layout(result, config)
 
@@ -200,7 +203,7 @@ def test_classify_document_from_layout_no_layout() -> None:
         mime_type="text/plain",
         metadata={},
     )
-    config = ExtractionConfig()
+    config = ExtractionConfig(auto_detect_document_type=True)
 
     doc_type, confidence = classify_document_from_layout(result, config)
 
@@ -218,7 +221,7 @@ def test_classify_document_from_layout_empty_layout() -> None:
         metadata={},
         layout=layout_df,
     )
-    config = ExtractionConfig()
+    config = ExtractionConfig(auto_detect_document_type=True)
 
     doc_type, confidence = classify_document_from_layout(result, config)
 
@@ -236,7 +239,7 @@ def test_classify_document_from_layout_missing_columns() -> None:
         metadata={},
         layout=layout_df,
     )
-    config = ExtractionConfig()
+    config = ExtractionConfig(auto_detect_document_type=True)
 
     doc_type, confidence = classify_document_from_layout(result, config)
 
@@ -260,7 +263,7 @@ def test_classify_document_from_layout_no_pattern_matches() -> None:
         metadata={},
         layout=layout_df,
     )
-    config = ExtractionConfig()
+    config = ExtractionConfig(auto_detect_document_type=True)
 
     doc_type, confidence = classify_document_from_layout(result, config)
 
@@ -285,7 +288,7 @@ def test_classify_document_from_layout_header_patterns() -> None:
         metadata={},
         layout=layout_df,
     )
-    config = ExtractionConfig()
+    config = ExtractionConfig(auto_detect_document_type=True)
 
     doc_type, confidence = classify_document_from_layout(result, config)
 
@@ -312,7 +315,7 @@ def test_classify_document_from_layout_position_scoring() -> None:
         metadata={},
         layout=layout_df,
     )
-    config = ExtractionConfig()
+    config = ExtractionConfig(auto_detect_document_type=True)
 
     doc_type, confidence = classify_document_from_layout(result, config)
 
@@ -327,7 +330,7 @@ def test_auto_detect_document_type_from_content() -> None:
         mime_type="text/plain",
         metadata={},
     )
-    config = ExtractionConfig()
+    config = ExtractionConfig(auto_detect_document_type=True)
 
     detection_result = auto_detect_document_type(result, config)
 
@@ -352,7 +355,7 @@ def test_auto_detect_document_type_from_layout() -> None:
         metadata={},
         layout=layout_df,
     )
-    config = ExtractionConfig()
+    config = ExtractionConfig(auto_detect_document_type=True)
 
     detection_result = auto_detect_document_type(result, config)
 
@@ -382,7 +385,7 @@ def test_auto_detect_document_type_no_matches() -> None:
         mime_type="text/plain",
         metadata={},
     )
-    config = ExtractionConfig()
+    config = ExtractionConfig(auto_detect_document_type=True)
 
     detection_result = auto_detect_document_type(result, config)
 
@@ -884,3 +887,35 @@ def test_classify_document_confidence_calculation(mocker: MockerFixture) -> None
 
     assert doc_type == "invoice"
     assert confidence == 1.0  # All 3 matches are for invoice, so 3/3 = 1.0
+
+
+def test_missing_deep_translator_import_error(mocker: MockerFixture) -> None:
+    """Test that MissingDependencyError is raised when deep-translator is not installed."""
+    # Temporarily remove deep_translator from sys.modules if it exists
+    original_module = sys.modules.pop("deep_translator", None)
+
+    try:
+        # Mock the import to raise ImportError when importing deep_translator
+        def mock_import(name: str, *args: Any, **kwargs: Any) -> Any:
+            if name == "deep_translator":
+                raise ImportError("No module named 'deep_translator'")
+            return original_import(name, *args, **kwargs)
+
+        original_import = builtins.__import__
+        mocker.patch("builtins.__import__", side_effect=mock_import)
+
+        # Import _get_translated_text after setting up the mock
+        from kreuzberg._document_classification import _get_translated_text
+
+        result = ExtractionResult(content="Test content", mime_type="text/plain", metadata={})
+
+        # Should raise MissingDependencyError when trying to import deep_translator
+        with pytest.raises(MissingDependencyError) as exc_info:
+            _get_translated_text(result)
+
+        assert "deep-translator" in str(exc_info.value)
+        assert "pip install 'kreuzberg[document-classification]'" in str(exc_info.value)
+    finally:
+        # Restore original module if it existed
+        if original_module is not None:
+            sys.modules["deep_translator"] = original_module