kreuzberg 3.14.0__py3-none-any.whl → 3.14.1__py3-none-any.whl

kreuzberg/_api/main.py

@@ -1,26 +1,34 @@
 from __future__ import annotations
 
+import base64
+import io
 import traceback
 from functools import lru_cache
 from json import dumps, loads
 from typing import TYPE_CHECKING, Annotated, Any, Literal
 
 import msgspec
+import polars as pl
+from PIL import Image
 from typing_extensions import TypedDict
 
 from kreuzberg import (
     EasyOCRConfig,
     ExtractionConfig,
     ExtractionResult,
+    GMFTConfig,
     KreuzbergError,
+    LanguageDetectionConfig,
     MissingDependencyError,
     PaddleOCRConfig,
     ParsingError,
+    SpacyEntityExtractionConfig,
     TesseractConfig,
     ValidationError,
     batch_extract_bytes,
 )
 from kreuzberg._config import discover_config
+from kreuzberg._types import HTMLToMarkdownConfig
 
 if TYPE_CHECKING:
     from litestar.datastructures import UploadFile
@@ -162,6 +170,22 @@ def _merge_configs_cached(
         ocr_backend = config_dict.get("ocr_backend")
         config_dict["ocr_config"] = _create_ocr_config(ocr_backend, config_dict["ocr_config"])
 
+    if "gmft_config" in config_dict and isinstance(config_dict["gmft_config"], dict):
+        config_dict["gmft_config"] = GMFTConfig(**config_dict["gmft_config"])
+
+    if "language_detection_config" in config_dict and isinstance(config_dict["language_detection_config"], dict):
+        config_dict["language_detection_config"] = LanguageDetectionConfig(**config_dict["language_detection_config"])
+
+    if "spacy_entity_extraction_config" in config_dict and isinstance(
+        config_dict["spacy_entity_extraction_config"], dict
+    ):
+        config_dict["spacy_entity_extraction_config"] = SpacyEntityExtractionConfig(
+            **config_dict["spacy_entity_extraction_config"]
+        )
+
+    if "html_to_markdown_config" in config_dict and isinstance(config_dict["html_to_markdown_config"], dict):
+        config_dict["html_to_markdown_config"] = HTMLToMarkdownConfig(**config_dict["html_to_markdown_config"])
+
     return ExtractionConfig(**config_dict)
 
 
@@ -291,6 +315,19 @@ async def get_configuration() -> ConfigurationResponse:
     }
 
 
+def _polars_dataframe_encoder(obj: Any) -> Any:
+    """Convert polars DataFrame to dict for JSON serialization."""
+    return obj.to_dicts()
+
+
+def _pil_image_encoder(obj: Any) -> str:
+    """Convert PIL Image to base64 string for JSON serialization."""
+    buffer = io.BytesIO()
+    obj.save(buffer, format="PNG")
+    img_str = base64.b64encode(buffer.getvalue()).decode()
+    return f"data:image/png;base64,{img_str}"
+
+
 openapi_config = OpenAPIConfig(
     title="Kreuzberg API",
     version="3.14.0",
@@ -307,6 +344,12 @@ openapi_config = OpenAPIConfig(
     create_examples=True,
 )
 
+# Type encoders for custom serialization
+type_encoders = {
+    pl.DataFrame: _polars_dataframe_encoder,
+    Image.Image: _pil_image_encoder,
+}
+
 app = Litestar(
     route_handlers=[handle_files_upload, health_check, get_configuration],
     plugins=[OpenTelemetryPlugin(OpenTelemetryConfig())],
@@ -316,5 +359,6 @@ app = Litestar(
         KreuzbergError: exception_handler,
         Exception: general_exception_handler,
     },
+    type_encoders=type_encoders,
     request_max_body_size=1024 * 1024 * 1024,  # 1GB limit for large file uploads
 )

kreuzberg/_types.py

@@ -349,7 +349,7 @@ class GMFTConfig(ConfigDict):
     """
 
 
-@dataclass(frozen=True, slots=True)
+@dataclass(unsafe_hash=True, frozen=True, slots=True)
 class LanguageDetectionConfig(ConfigDict):
     low_memory: bool = True
     """If True, uses a smaller model (~200MB). If False, uses a larger, more accurate model.
@@ -751,7 +751,7 @@ PostProcessingHook = Callable[[ExtractionResult], ExtractionResult | Awaitable[E
 ValidationHook = Callable[[ExtractionResult], None | Awaitable[None]]
 
 
-@dataclass(unsafe_hash=True, slots=True)
+@dataclass(unsafe_hash=True, frozen=True, slots=True)
 class ExtractionConfig(ConfigDict):
     force_ocr: bool = False
     """Whether to force OCR."""

{kreuzberg-3.14.0.dist-info → kreuzberg-3.14.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: kreuzberg
-Version: 3.14.0
+Version: 3.14.1
 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-3.14.0.dist-info → kreuzberg-3.14.1.dist-info}/RECORD

@@ -10,13 +10,13 @@ kreuzberg/_language_detection.py,sha256=T9p6aimB7QFXAQiEntIMZeH_Z62E52E8fBQ43hWu
 kreuzberg/_mime_types.py,sha256=kGBDSMO4XPgzUKC7iaBeChCtRQXZ9_zXq6eJydejX_k,7739
 kreuzberg/_playa.py,sha256=p4G5ymSSCbQoDeXJjH-yuVzdd4y-wKcolqDthjPtqok,11413
 kreuzberg/_registry.py,sha256=8XYT-vPhNYMAbB5RBIUKz-1Zdg48OCnBcdVZzBq6YwY,3307
-kreuzberg/_types.py,sha256=yw8ZzCgwp8T4byh00gdSlABDtRwro6H1pemQsO5IZMQ,39132
+kreuzberg/_types.py,sha256=BEMTnA8fvHL0dDCnjq7g9Jjd2Ze8NFq988YkMH4zQ9g,39163
 kreuzberg/cli.py,sha256=Ob0IfqWcaiM09pFdC6wTpdSeql0SGZDxBxfrEhJAGmo,13501
 kreuzberg/exceptions.py,sha256=PTiAZgQwcG9hXbgYg2W7sfxksFhq5_wzOFgZGnTJAoc,2991
 kreuzberg/extraction.py,sha256=qT-Ziw5FmMqcPT88VrglikL1RASSJCf5W7xP6L9Vi5s,17673
 kreuzberg/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 kreuzberg/_api/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-kreuzberg/_api/main.py,sha256=bZLaQpW8eoTFGvCGJgFodALy4rDfe9kuY1oj9OKPQpU,10792
+kreuzberg/_api/main.py,sha256=8g_8j8Dp2e70_yYYUUrJNC5Ku9fuyNgyjUuIgJTRUW8,12500
 kreuzberg/_extractors/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 kreuzberg/_extractors/_base.py,sha256=i2FvAhRnamEtBb4a-C7pfcdWIXnkEBw0saMQu7h1_RQ,2069
 kreuzberg/_extractors/_email.py,sha256=jn_8J4BASKJ7zFHBG0PgxNe3OT4pjmEM2tTKX8y_0AE,5887
@@ -51,8 +51,8 @@ kreuzberg/_utils/_string.py,sha256=wVyvEHByHBeu_6evmqJGv9Ml-NAwkyz60n8l-7L5Cw0,4
 kreuzberg/_utils/_sync.py,sha256=OWiciXPTGHIxgiGoHI2AglZ1siTNT-nU_JCgHPNzzHk,2196
 kreuzberg/_utils/_table.py,sha256=OVg6T2QnerMhVNb1juLTBSIjyjFiE5-OrUWr5NSCgnQ,6493
 kreuzberg/_utils/_tmp.py,sha256=wnOInBkcuQoxI1vBLvNv9NqbRCEu9Y03qfOjqQuAk3s,841
-kreuzberg-3.14.0.dist-info/METADATA,sha256=68rRivXnf8n_F9lqekOydDOd8sehWpHpbbKzRup7XDc,12127
-kreuzberg-3.14.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-kreuzberg-3.14.0.dist-info/entry_points.txt,sha256=GplGhFryCP7kyAG_k-Mdahznvo2fwi73qLFg5yQfH_A,91
-kreuzberg-3.14.0.dist-info/licenses/LICENSE,sha256=-8caMvpCK8SgZ5LlRKhGCMtYDEXqTKH9X8pFEhl91_4,1066
-kreuzberg-3.14.0.dist-info/RECORD,,
+kreuzberg-3.14.1.dist-info/METADATA,sha256=4sG9L9AtvBHFxjv84obrcaYNToc_sO0-AHnnpo1-ZGY,12127
+kreuzberg-3.14.1.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+kreuzberg-3.14.1.dist-info/entry_points.txt,sha256=GplGhFryCP7kyAG_k-Mdahznvo2fwi73qLFg5yQfH_A,91
+kreuzberg-3.14.1.dist-info/licenses/LICENSE,sha256=-8caMvpCK8SgZ5LlRKhGCMtYDEXqTKH9X8pFEhl91_4,1066
+kreuzberg-3.14.1.dist-info/RECORD,,