natural-pdf 0.1.38__py3-none-any.whl → 0.2.0__py3-none-any.whl

natural_pdf/elements/text.py

@@ -230,7 +230,9 @@ class TextElement(Element):
         # Default to black
         return (0, 0, 0)
 
-    def extract_text(self, keep_blank_chars=True, strip: Optional[bool] = True, content_filter=None, **kwargs) -> str:
+    def extract_text(
+        self, keep_blank_chars=True, strip: Optional[bool] = True, content_filter=None, **kwargs
+    ) -> str:
         """
         Extract text from this element.
 
@@ -253,22 +255,22 @@ class TextElement(Element):
         # Apply content filtering if provided
         if content_filter is not None and result:
             import re
-            
+
             if isinstance(content_filter, str):
                 # Single regex pattern - remove matching characters
                 try:
-                    result = re.sub(content_filter, '', result)
+                    result = re.sub(content_filter, "", result)
                 except re.error:
                     pass  # Invalid regex, skip filtering
-                    
+
             elif isinstance(content_filter, list):
                 # List of regex patterns - remove characters matching ANY pattern
                 try:
                     for pattern in content_filter:
-                        result = re.sub(pattern, '', result)
+                        result = re.sub(pattern, "", result)
                 except re.error:
                     pass  # Invalid regex, skip filtering
-                    
+
             elif callable(content_filter):
                 # Callable filter - apply to individual characters
                 try:
@@ -276,7 +278,7 @@ class TextElement(Element):
                     for char in result:
                         if content_filter(char):
                             filtered_chars.append(char)
-                    result = ''.join(filtered_chars)
+                    result = "".join(filtered_chars)
                 except Exception:
                     pass  # Function error, skip filtering
 

natural_pdf/exporters/base.py

@@ -3,8 +3,8 @@ import logging
 from typing import TYPE_CHECKING, List, Union
 
 if TYPE_CHECKING:
-    from natural_pdf.collections.pdf_collection import PDFCollection
     from natural_pdf.core.pdf import PDF
+    from natural_pdf.core.pdf_collection import PDFCollection
 
 logger = logging.getLogger(__name__)
 
@@ -40,8 +40,8 @@ class FinetuneExporter(abc.ABC):
         """
         Helper to consistently resolve the input source to a list of PDF objects.
         """
-        from natural_pdf.collections.pdf_collection import PDFCollection  # Avoid circular import
         from natural_pdf.core.pdf import PDF  # Avoid circular import at module level
+        from natural_pdf.core.pdf_collection import PDFCollection  # Avoid circular import
 
         pdfs_to_process: List["PDF"] = []
         if isinstance(source, PDF):

natural_pdf/exporters/original_pdf.py

@@ -17,8 +17,8 @@ except ImportError:
 
 if TYPE_CHECKING:
     from natural_pdf.core.page import Page
+    from natural_pdf.core.page_collection import PageCollection
     from natural_pdf.core.pdf import PDF
-    from natural_pdf.elements.collections import PageCollection
 
 logger = logging.getLogger(__name__)
 

natural_pdf/exporters/paddleocr.py

@@ -13,8 +13,8 @@ from natural_pdf.exporters.base import FinetuneExporter
 from natural_pdf.utils.identifiers import generate_short_path_hash
 
 if TYPE_CHECKING:
-    from natural_pdf.collections.pdf_collection import PDFCollection
     from natural_pdf.core.pdf import PDF
+    from natural_pdf.core.pdf_collection import PDFCollection
     from natural_pdf.elements.text import TextElement
 
 logger = logging.getLogger(__name__)
@@ -216,9 +216,7 @@ class PaddleOCRRecognitionExporter(FinetuneExporter):
                     try:
                         # Expand region, render, and save image
                         region = element.expand(self.padding)
-                        img = region.to_image(
-                            resolution=self.resolution, crop=True, include_highlights=False
-                        )
+                        img = region.render(resolution=self.resolution, crop=True)
                         img.save(absolute_image_path, "PNG")
 
                         # Add to labels and character set

natural_pdf/exporters/searchable_pdf.py

@@ -28,8 +28,8 @@ except ImportError:
 
 if TYPE_CHECKING:
     from natural_pdf.core.page import Page
+    from natural_pdf.core.page_collection import PageCollection
     from natural_pdf.core.pdf import PDF
-    from natural_pdf.elements.collections import PageCollection
 
 
 logger = logging.getLogger(__name__)
@@ -345,7 +345,8 @@ def create_searchable_pdf(
                 # 1. Render page image at target DPI
                 logger.debug(f"  Rendering page {i} to image ({dpi} DPI)...")
                 # Use the Page's to_image method
-                pil_image = page.to_image(resolution=dpi, include_highlights=False)
+                # Use render() for clean image without highlights
+                pil_image = page.render(resolution=dpi)
                 pil_image.save(img_path, format="PNG")
                 img_width, img_height = pil_image.size
                 logger.debug(f"  Image saved to {img_path} ({img_width}x{img_height})")

natural_pdf/extraction/mixin.py

@@ -94,10 +94,8 @@ class ExtractionMixin(ABC):
                 resolution = kwargs.pop("resolution", 72)
                 include_highlights = kwargs.pop("include_highlights", False)
                 labels = kwargs.pop("labels", False)
-                return self.to_image(
+                return self.render(
                     resolution=resolution,
-                    include_highlights=include_highlights,
-                    labels=labels,
                     **kwargs,
                 )
             else:

natural_pdf/flows/collections.py

@@ -8,7 +8,7 @@ if TYPE_CHECKING:
     # from PIL.Image import Image as PIL_Image # No longer needed with Image.Image type hint
     from natural_pdf.core.page import Page as PhysicalPage
     from natural_pdf.elements.base import Element as PhysicalElement
-    from natural_pdf.elements.collections import ElementCollection
+    from natural_pdf.elements.element_collection import ElementCollection
 
     from .element import FlowElement
     from .flow import Flow  # Though not directly used in __init__, FlowRegion needs it.
@@ -632,73 +632,5 @@ class FlowRegionCollection(MutableSequence[T_FRC]):
             all_cropped_images.extend(fr.to_images(resolution=resolution, **kwargs))
         return all_cropped_images
 
-    def to_image(
-        self,
-        stack_direction: str = "vertical",
-        background_color=(255, 255, 255),
-        gap: int = 5,
-        **kwargs_for_constituent_to_image,
-    ) -> Optional[Image.Image]:
-        """
-        Creates a single composite image by stacking the composite images of each FlowRegion.
-        Each FlowRegion's composite is generated by its own .to_image() method.
-        These are then stacked.
-
-        Args:
-            stack_direction: "vertical" or "horizontal".
-            background_color: Background for the final composite.
-            gap: Gap in pixels between stacked FlowRegion images.
-            **kwargs_for_constituent_to_image: Passed to each FlowRegion.to_image().
-        """
-        if not self._flow_regions:
-            return None
-
-        region_composites: List[Image.Image] = []
-        for fr in self._flow_regions:
-            img = fr.to_image(background_color=background_color, **kwargs_for_constituent_to_image)
-            if img:
-                region_composites.append(img)
-
-        if not region_composites:
-            return None
-        if len(region_composites) == 1:
-            return region_composites[0]
-
-        if stack_direction == "vertical":
-            final_width = max(img.width for img in region_composites)
-            final_height = (
-                sum(img.height for img in region_composites) + (len(region_composites) - 1) * gap
-            )
-            if final_width == 0 or final_height == 0:
-                return None
-
-            new_image = Image.new("RGB", (final_width, final_height), background_color)
-            current_y = 0
-            for img in region_composites:
-                # Align to left for vertical stacking
-                new_image.paste(img, (0, current_y))
-                current_y += img.height + gap
-            return new_image
-        elif stack_direction == "horizontal":
-            final_width = (
-                sum(img.width for img in region_composites) + (len(region_composites) - 1) * gap
-            )
-            final_height = max(img.height for img in region_composites)
-            if final_width == 0 or final_height == 0:
-                return None
-
-            new_image = Image.new("RGB", (final_width, final_height), background_color)
-            current_x = 0
-            for img in region_composites:
-                # Align to top for horizontal stacking
-                new_image.paste(img, (current_x, 0))
-                current_x += img.width + gap
-            return new_image
-        else:
-            logger.warning(
-                f"Invalid stack_direction: {stack_direction}. Must be 'vertical' or 'horizontal'."
-            )
-            return None  # Or perhaps return the list of images?
-
     def apply(self, func: Callable[["FlowRegion"], Any]) -> List[Any]:
         return [func(fr) for fr in self._flow_regions]

natural_pdf/flows/element.py

@@ -73,6 +73,31 @@ class FlowElement:
         """Returns the physical page of the underlying element."""
         return getattr(self.physical_object, "page", None)
 
+    def __getattr__(self, name: str) -> Any:
+        """
+        Delegate unknown attribute access to the physical_object.
+
+        This ensures that attributes like 'type', 'region_type', 'source', 'model', etc.
+        from the physical element are accessible on the FlowElement wrapper.
+
+        Args:
+            name: The attribute name being accessed
+
+        Returns:
+            The attribute value from physical_object
+
+        Raises:
+            AttributeError: If the attribute doesn't exist on physical_object either
+        """
+        try:
+            return getattr(self.physical_object, name)
+        except AttributeError:
+            # Provide a helpful error message that mentions both FlowElement and physical_object
+            raise AttributeError(
+                f"'{type(self).__name__}' object has no attribute '{name}' "
+                f"(also not found on underlying {type(self.physical_object).__name__})"
+            )
+
     def _flow_direction(
         self,
         direction: str,  # "above", "below", "left", "right"