kreuzberg 3.11.3__tar.gz → 3.11.4__tar.gz

{kreuzberg-3.11.3 → kreuzberg-3.11.4}/.docker/Dockerfile

@@ -46,9 +46,13 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
 COPY --from=builder /app/.venv /app/.venv
 COPY --from=builder /app/kreuzberg /app/kreuzberg
 
-# Create non-root user
+# Create non-root user and cache directory
 RUN groupadd -r appuser && useradd -r -g appuser -d /app -s /sbin/nologin appuser && \
+    mkdir -p /app/.kreuzberg && \
     chown -R appuser:appuser /app
 
+# Set default cache directory to prevent permission issues
+ENV KREUZBERG_CACHE_DIR=/app/.kreuzberg
+
 USER appuser
 CMD ["litestar", "--app", "kreuzberg._api.main:app", "run", "--host", "0.0.0.0"]

{kreuzberg-3.11.3 → kreuzberg-3.11.4}/.github/workflows/docker-e2e-tests.yml

@@ -15,6 +15,7 @@ jobs:
           - { name: "easyocr", extras: "easyocr" }
           - { name: "paddle", extras: "paddleocr" }
           - { name: "gmft", extras: "gmft" }
+          - { name: "all", extras: "all" }
       fail-fast: false
 
     steps:

{kreuzberg-3.11.3 → kreuzberg-3.11.4}/.gitignore

@@ -39,3 +39,6 @@ requirements.txt
 site/
 .cache/
 dist/
+.task/
+tests/e2e/test_report.json
+tests/e2e/logs/

{kreuzberg-3.11.3 → kreuzberg-3.11.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: kreuzberg
-Version: 3.11.3
+Version: 3.11.4
 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

kreuzberg-3.11.4/docker-compose.example.yml

@@ -0,0 +1,37 @@
+services:
+  kreuzberg:
+    # Choose your image variant:
+    # - goldziher/kreuzberg:latest (core, 270MB)
+    # - goldziher/kreuzberg:latest-paddle (+ PaddleOCR, 878MB)
+    # - goldziher/kreuzberg:latest-easyocr (+ EasyOCR, 8.7GB)
+    # - goldziher/kreuzberg:latest-gmft (+ table extraction, 8.6GB)
+    # - goldziher/kreuzberg:latest-all (all features, 9.6GB - testing only!)
+    image: goldziher/kreuzberg:latest
+
+    ports:
+      - "8000:8000"
+
+    volumes:
+      # Mount your configuration file (optional)
+      - "./kreuzberg.toml:/app/kreuzberg.toml:ro"
+      # Persist cache across restarts (recommended)
+      - "kreuzberg-cache:/app/.kreuzberg"
+
+    environment:
+      - PYTHONUNBUFFERED=1
+      - KREUZBERG_CACHE_DIR=/app/.kreuzberg
+      # Optional: Cache size limits
+      # - KREUZBERG_OCR_CACHE_SIZE_MB=500
+      # - KREUZBERG_DOCUMENT_CACHE_SIZE_MB=1000
+
+    restart: unless-stopped
+
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
+      interval: 30s
+      timeout: 10s
+      retries: 3
+      start_period: 40s
+
+volumes:
+  kreuzberg-cache:

kreuzberg-3.11.4/docs/user-guide/docker.md

@@ -0,0 +1,417 @@
+# Docker
+
+Kreuzberg provides official Docker images for easy deployment and containerized usage.
+
+## Image Variants
+
+Docker images are available on [Docker Hub](https://hub.docker.com/r/goldziher/kreuzberg):
+
+### Core Image
+
+- **Image**: `goldziher/kreuzberg:latest`
+- **Size**: ~270MB
+- **Includes**: Base library + API server + Tesseract OCR
+- **Use Case**: Basic text extraction from documents
+- **Limitations**: No chunking, language detection, entity extraction, or alternative OCR backends
+
+### OCR Backend Variants
+
+- **EasyOCR**: `goldziher/kreuzberg:latest-easyocr` (~8.7GB)
+
+    - Deep learning-based OCR with support for 80+ languages
+    - Better accuracy for complex layouts and handwriting
+
+- **PaddleOCR**: `goldziher/kreuzberg:latest-paddle` (~878MB)
+
+    - Lightweight deep learning OCR
+    - Good balance between size and accuracy
+
+### Table Extraction
+
+- **GMFT**: `goldziher/kreuzberg:latest-gmft` (~8.6GB)
+    - Advanced table detection and extraction from PDFs
+    - Uses Microsoft's Table Transformer models
+
+### All-in-One (Testing Only)
+
+- **Image**: `goldziher/kreuzberg:latest-all`
+- **Size**: ~9.6GB
+- **⚠️ WARNING**: For testing only, NOT for production use
+- **Includes**: All OCR backends and features
+- **Why not production?**: Unnecessarily large, includes conflicting dependencies, slower startup
+
+## Quick Start
+
+### Basic Usage
+
+```bash
+# Pull and run the core image
+docker pull goldziher/kreuzberg:latest
+docker run -p 8000:8000 goldziher/kreuzberg:latest
+
+# Extract text from a document
+curl -X POST http://localhost:8000/extract \
+  -F "data=@document.pdf"
+```
+
+### With Cache Volume
+
+```bash
+# Create cache directory
+mkdir -p kreuzberg-cache
+
+# Run with persistent cache
+docker run -p 8000:8000 \
+  -v "$(pwd)/kreuzberg-cache:/app/.kreuzberg" \
+  goldziher/kreuzberg:latest
+```
+
+## Customizing Docker Images
+
+For production, create a custom image with only the features you need:
+
+### Example 1: Core + Chunking Support
+
+```dockerfile
+FROM goldziher/kreuzberg:latest
+
+USER root
+
+# Install only chunking dependency
+RUN uv pip install --python /app/.venv/bin/python semantic-text-splitter
+
+USER appuser
+```
+
+Build and run:
+
+```bash
+# Build the image
+docker build -t kreuzberg-chunking .
+
+# Run with external configuration
+docker run -p 8000:8000 \
+  -v "$(pwd)/kreuzberg.toml:/app/kreuzberg.toml:ro" \
+  -v "$(pwd)/cache:/app/.kreuzberg" \
+  kreuzberg-chunking
+```
+
+### Example 2: Core + Language Detection + Chunking
+
+```dockerfile
+FROM goldziher/kreuzberg:latest
+
+USER root
+
+# Install specific features
+RUN uv pip install --python /app/.venv/bin/python \
+    semantic-text-splitter \
+    fast-langdetect
+
+USER appuser
+```
+
+Create configuration file `kreuzberg.toml`:
+
+```toml
+chunk_content = true
+auto_detect_language = true
+max_chars = 2000
+max_overlap = 100
+```
+
+Run with:
+
+```bash
+docker run -p 8000:8000 \
+  -v "$(pwd)/kreuzberg.toml:/app/kreuzberg.toml:ro" \
+  -v "$(pwd)/cache:/app/.kreuzberg" \
+  kreuzberg-multilang
+```
+
+### Example 3: Core + PaddleOCR (Custom Build)
+
+```dockerfile
+FROM goldziher/kreuzberg:latest
+
+USER root
+
+# Install PaddleOCR dependencies
+RUN uv pip install --python /app/.venv/bin/python \
+    paddleocr \
+    paddlepaddle
+
+USER appuser
+```
+
+Run with PaddleOCR backend:
+
+```bash
+docker run -p 8000:8000 \
+  -e KREUZBERG_OCR_BACKEND=paddleocr \
+  -v "$(pwd)/cache:/app/.kreuzberg" \
+  kreuzberg-paddle
+```
+
+### Example 4: Optimized Production Build
+
+```dockerfile
+FROM goldziher/kreuzberg:latest
+
+USER root
+
+# Install only the features you need
+RUN uv pip install --python /app/.venv/bin/python \
+    semantic-text-splitter \
+    fast-langdetect && \
+    # Clean up cache to reduce image size
+    rm -rf /root/.cache/uv
+
+USER appuser
+
+# Set production environment variables
+ENV PYTHONUNBUFFERED=1
+ENV PYTHONDONTWRITEBYTECODE=1
+```
+
+Deploy with Docker Compose:
+
+```yaml
+services:
+  kreuzberg:
+    build: .
+    ports:
+      - "8000:8000"
+    volumes:
+      - "./config/kreuzberg.toml:/app/kreuzberg.toml:ro"
+      - "kreuzberg-cache:/app/.kreuzberg"
+    environment:
+      - PYTHONUNBUFFERED=1
+      - KREUZBERG_CACHE_DIR=/app/.kreuzberg
+    restart: unless-stopped
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
+      interval: 30s
+      timeout: 10s
+      retries: 3
+
+volumes:
+  kreuzberg-cache:
+```
+
+## Docker Compose
+
+### Production Setup
+
+```yaml
+services:
+  kreuzberg:
+    image: goldziher/kreuzberg:latest  # Or your custom image
+    ports:
+      - "8000:8000"
+    volumes:
+      - "./kreuzberg-cache:/app/.kreuzberg"  # Persistent cache
+      - "./kreuzberg.toml:/app/kreuzberg.toml:ro"  # Configuration
+    environment:
+      - PYTHONUNBUFFERED=1
+      - KREUZBERG_CACHE_DIR=/app/.kreuzberg
+      # Cache configuration
+      - KREUZBERG_OCR_CACHE_SIZE_MB=500
+      - KREUZBERG_DOCUMENT_CACHE_SIZE_MB=1000
+    restart: unless-stopped
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
+      interval: 30s
+      timeout: 10s
+      retries: 3
+```
+
+## Configuration
+
+### Using Configuration Files
+
+Create `kreuzberg.toml`:
+
+```toml
+force_ocr = false
+chunk_content = true  # Requires semantic-text-splitter
+extract_tables = false  # Requires gmft
+ocr_backend = "tesseract"
+
+[tesseract]
+language = "eng"
+psm = 6
+```
+
+Mount the configuration:
+
+```bash
+docker run -p 8000:8000 \
+  -v "$(pwd)/kreuzberg.toml:/app/kreuzberg.toml" \
+  goldziher/kreuzberg:latest
+```
+
+### Environment Variables
+
+**Cache Configuration:**
+
+- `KREUZBERG_CACHE_DIR`: Cache directory (default: `/app/.kreuzberg`)
+- `KREUZBERG_OCR_CACHE_SIZE_MB`: OCR cache size limit (default: `500`)
+- `KREUZBERG_DOCUMENT_CACHE_SIZE_MB`: Document cache size limit (default: `1000`)
+
+**Runtime Configuration:**
+
+- `PYTHONUNBUFFERED=1`: Ensures proper logging output
+- `PYTHONDONTWRITEBYTECODE=1`: Prevents .pyc file creation
+
+## Production Deployment
+
+### Kubernetes
+
+```yaml
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: kreuzberg-api
+spec:
+  replicas: 3
+  selector:
+    matchLabels:
+      app: kreuzberg-api
+  template:
+    metadata:
+      labels:
+        app: kreuzberg-api
+    spec:
+      containers:
+      - name: kreuzberg
+        image: your-registry/kreuzberg-custom:latest
+        ports:
+        - containerPort: 8000
+        volumeMounts:
+        - name: cache
+          mountPath: /app/.kreuzberg
+        - name: config
+          mountPath: /app/kreuzberg.toml
+          subPath: kreuzberg.toml
+        livenessProbe:
+          httpGet:
+            path: /health
+            port: 8000
+          initialDelaySeconds: 30
+        readinessProbe:
+          httpGet:
+            path: /health
+            port: 8000
+          initialDelaySeconds: 5
+        resources:
+          requests:
+            memory: "512Mi"
+            cpu: "500m"
+          limits:
+            memory: "2Gi"
+            cpu: "2000m"
+      volumes:
+      - name: cache
+        emptyDir: {}
+      - name: config
+        configMap:
+          name: kreuzberg-config
+---
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: kreuzberg-config
+data:
+  kreuzberg.toml: |
+    chunk_content = false
+    ocr_backend = "tesseract"
+    [tesseract]
+    language = "eng"
+```
+
+### With nginx Reverse Proxy
+
+```nginx
+server {
+    listen 80;
+    server_name api.example.com;
+
+    location / {
+        proxy_pass http://localhost:8000;
+        proxy_set_header Host $host;
+        proxy_set_header X-Real-IP $remote_addr;
+
+        # File upload settings
+        client_max_body_size 100M;
+        proxy_read_timeout 300s;
+    }
+
+    location /health {
+        proxy_pass http://localhost:8000/health;
+        access_log off;
+    }
+}
+```
+
+## Resource Requirements
+
+| Variant     | CPU      | Memory | Storage |
+| ----------- | -------- | ------ | ------- |
+| Core        | 1+ cores | 512MB+ | 1GB     |
+| + Chunking  | 1+ cores | 1GB+   | 1GB     |
+| + PaddleOCR | 2+ cores | 2GB+   | 2GB     |
+| + EasyOCR   | 2+ cores | 4GB+   | 10GB    |
+| + GMFT      | 2+ cores | 4GB+   | 10GB    |
+
+## Troubleshooting
+
+### Common Issues
+
+#### Permission Denied on Cache Directory
+
+```bash
+# Fix: Ensure proper ownership
+docker run --rm -v "$(pwd)/cache:/app/.kreuzberg" --user root \
+  goldziher/kreuzberg:latest \
+  chown -R 999:999 /app/.kreuzberg
+```
+
+#### Missing Dependencies Error
+
+```bash
+# Solution: Use appropriate image variant or build custom image
+# For chunking: Install semantic-text-splitter
+# For language detection: Install fast-langdetect
+```
+
+#### Out of Memory
+
+- Increase Docker memory allocation
+- Use a smaller OCR engine (Tesseract instead of EasyOCR)
+- Disable unnecessary features
+
+### Debugging
+
+```bash
+# Check logs
+docker logs <container-id>
+
+# Shell access
+docker exec -it <container-id> /bin/bash
+
+# Test extraction
+docker exec <container-id> python3 -c "
+from kreuzberg import extract_file_sync
+result = extract_file_sync('/path/to/file.pdf')
+print(result.content[:100])
+"
+```
+
+## Security Considerations
+
+- Runs as non-root user (`appuser`) by default
+- No external API calls or cloud dependencies
+- Process files locally within the container
+- Use read-only mounts where possible (`:ro`)
+- Consider adding authentication for production use

{kreuzberg-3.11.3 → kreuzberg-3.11.4}/kreuzberg/cli.py

@@ -19,11 +19,9 @@ except ImportError as e:  # pragma: no cover
 
 from kreuzberg import __version__, extract_bytes_sync, extract_file_sync
 from kreuzberg._config import build_extraction_config, find_config_file, load_config_from_file
+from kreuzberg._constants import DEFAULT_MAX_CHARACTERS, DEFAULT_MAX_OVERLAP
 from kreuzberg.exceptions import KreuzbergError, MissingDependencyError
 
-DEFAULT_MAX_CHARACTERS = 4000
-DEFAULT_MAX_OVERLAP = 200
-
 if TYPE_CHECKING:
     from kreuzberg._types import ExtractionConfig, ExtractionResult
 

{kreuzberg-3.11.3 → kreuzberg-3.11.4}/pyproject.toml

@@ -5,7 +5,7 @@ requires = [ "hatchling" ]
 
 [project]
 name = "kreuzberg"
-version = "3.11.3"
+version = "3.11.4"
 description = "Document intelligence framework for Python - Extract text, metadata, and structured data from diverse file formats"
 readme = "README.md"
 keywords = [
@@ -202,9 +202,7 @@ lint.per-file-ignores."tests/**/*.*" = [
   "SLF001",
 ]
 lint.per-file-ignores."tests/e2e/*_test.py" = [
-  "ANN",    # Disable type annotations in E2E tests
-  "PTH123", # Allow open() in E2E tests
-  "T201",   # print statements are needed for test output
+  "T201", # print statements are needed for test output
 ]
 lint.isort.known-first-party = [ "kreuzberg", "tests" ]
 lint.mccabe.max-complexity = 15

{kreuzberg-3.11.3 → kreuzberg-3.11.4}/tests/e2e/docker_images_test.py

@@ -31,10 +31,11 @@ DOCKER_IMAGES = {
     "easyocr": "kreuzberg:easyocr",
     "paddle": "kreuzberg:paddle",
     "gmft": "kreuzberg:gmft",
+    "all": "kreuzberg:all",
 }
 
 # Images that are optional (may not be built due to space constraints)
-OPTIONAL_IMAGES = {"paddle", "gmft"}
+OPTIONAL_IMAGES = {"paddle", "gmft", "all"}
 
 # Security configuration
 SECURITY_CONFIG = {
@@ -111,7 +112,7 @@ def test_cli_help(image_name: str) -> bool:
         "--help",
     ]
     exit_code, stdout, stderr = run_command(cmd)
-    success = exit_code == 0 and "Extract text from documents" in stdout
+    success = exit_code == 0 and "Text extraction from documents" in stdout
     if not success:
         pass
     return success
@@ -536,13 +537,11 @@ def print_summary(all_results: dict[str, dict[str, bool]]) -> bool:
     success_rate = (total_passed / total_tests * 100) if total_tests > 0 else 0
     print(f"Success rate: {success_rate:.1f}%")
 
-    if success_rate >= 90:
+    # ALL tests must pass - no partial success allowed
+    if success_rate == 100:
         print("✅ Test suite PASSED")
         return True
-    if success_rate >= 70:
-        print("⚠️ Test suite PASSED with warnings")
-        return True
-    print("❌ Test suite FAILED")
+    print("❌ Test suite FAILED - all tests must pass")
     return False
 
 

kreuzberg-3.11.4/tests/test_source_files/contract.txt

@@ -0,0 +1 @@
+Test content for contract.txt file

{kreuzberg-3.11.3 → kreuzberg-3.11.4}/uv.lock

@@ -956,7 +956,7 @@ name = "exceptiongroup"
 version = "1.3.0"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
-    { name = "typing-extensions", marker = "python_full_version < '3.13'" },
+    { name = "typing-extensions", marker = "python_full_version < '3.11'" },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/0b/9f/a65090624ecf468cdca03533906e7c69ed7588582240cfe7cc9e770b50eb/exceptiongroup-1.3.0.tar.gz", hash = "sha256:b241f5885f560bc56a59ee63ca4c6a8bfa46ae4ad651af316d4e81817bb9fd88", size = 29749, upload-time = "2025-05-10T17:42:51.123Z" }
 wheels = [
@@ -1632,7 +1632,7 @@ wheels = [
 
 [[package]]
 name = "kreuzberg"
-version = "3.11.2"
+version = "3.11.4"
 source = { editable = "." }
 dependencies = [
     { name = "anyio" },

kreuzberg-3.11.3/.task/checksum/docker-build-core

@@ -1 +0,0 @@
-f2683e2fcbf67fa54bee4baa3ead293a

kreuzberg-3.11.3/.task/checksum/docker-build-easyocr

@@ -1 +0,0 @@
-f2683e2fcbf67fa54bee4baa3ead293a

kreuzberg-3.11.3/.task/checksum/docker-build-gmft

@@ -1 +0,0 @@
-f2683e2fcbf67fa54bee4baa3ead293a

kreuzberg-3.11.3/.task/checksum/docker-build-paddle

@@ -1 +0,0 @@
-f2683e2fcbf67fa54bee4baa3ead293a