natural-pdf 0.1.28__py3-none-any.whl → 0.1.31__py3-none-any.whl

natural_pdf/core/highlighting_service.py

@@ -70,16 +70,16 @@ class HighlightRenderer:
         page: Page,
         base_image: Image.Image,
         highlights: List[Highlight],
-        scale: float,
+        scale_factor: float,
         render_ocr: bool,
     ):
         self.page = page  # Keep page reference for OCR rendering
         self.base_image = base_image.convert("RGBA")  # Ensure RGBA
         self.highlights = highlights
-        self.scale = scale
+        self.scale_factor = scale_factor  # Renamed from scale to scale_factor for clarity
         self.render_ocr = render_ocr
         self.result_image = self.base_image.copy()
-        self.vertex_size = max(3, int(2 * self.scale))  # Size of corner markers
+        self.vertex_size = max(3, int(2 * self.scale_factor))  # Size of corner markers
 
     def render(self) -> Image.Image:
         """Executes the rendering process."""
@@ -98,7 +98,7 @@ class HighlightRenderer:
             scaled_bbox = None
 
             if highlight.is_polygon:
-                scaled_polygon = [(p[0] * self.scale, p[1] * self.scale) for p in highlight.polygon]
+                scaled_polygon = [(p[0] * self.scale_factor, p[1] * self.scale_factor) for p in highlight.polygon]
                 # Draw polygon fill and border
                 draw.polygon(
                     scaled_polygon, fill=highlight.color, outline=highlight.border_color, width=2
@@ -113,10 +113,10 @@ class HighlightRenderer:
             else:  # Rectangle
                 x0, top, x1, bottom = highlight.bbox
                 x0_s, top_s, x1_s, bottom_s = (
-                    x0 * self.scale,
-                    top * self.scale,
-                    x1 * self.scale,
-                    bottom * self.scale,
+                    x0 * self.scale_factor,
+                    top * self.scale_factor,
+                    x1 * self.scale_factor,
+                    bottom * self.scale_factor,
                 )
                 scaled_bbox = [x0_s, top_s, x1_s, bottom_s]
                 # Draw rectangle fill and border
@@ -159,15 +159,15 @@ class HighlightRenderer:
         """Draws attribute key-value pairs on the highlight."""
         try:
             # Slightly larger font, scaled
-            font_size = max(10, int(8 * self.scale))
+            font_size = max(10, int(8 * self.scale_factor))
             # Prioritize monospace fonts for better alignment
             font = ImageFont.truetype("Arial.ttf", font_size)  # Fallback sans-serif
         except IOError:
             font = ImageFont.load_default()
             font_size = 10  # Reset size for default font
 
-        line_height = font_size + int(4 * self.scale)  # Scaled line spacing
-        bg_padding = int(3 * self.scale)
+        line_height = font_size + int(4 * self.scale_factor)  # Scaled line spacing
+        bg_padding = int(3 * self.scale_factor)
         max_width = 0
         text_lines = []
 
@@ -191,8 +191,8 @@ class HighlightRenderer:
         total_height = line_height * len(text_lines)
 
         # Position near top-right corner with padding
-        x = bbox_scaled[2] - int(2 * self.scale) - max_width
-        y = bbox_scaled[1] + int(2 * self.scale)
+        x = bbox_scaled[2] - int(2 * self.scale_factor) - max_width
+        y = bbox_scaled[1] + int(2 * self.scale_factor)
 
         # Draw background rectangle (semi-transparent white)
         bg_x0 = x - bg_padding
@@ -244,10 +244,10 @@ class HighlightRenderer:
         for element in ocr_elements:
             x0, top, x1, bottom = element.bbox
             x0_s, top_s, x1_s, bottom_s = (
-                x0 * self.scale,
-                top * self.scale,
-                x1 * self.scale,
-                bottom * self.scale,
+                x0 * self.scale_factor,
+                top * self.scale_factor,
+                x1 * self.scale_factor,
+                bottom * self.scale_factor,
             )
             box_w, box_h = x1_s - x0_s, bottom_s - top_s
 
@@ -556,19 +556,62 @@ class HighlightingService:
         self._highlights_by_page[page_index].append(highlight)
         logger.debug(f"Added highlight to page {page_index}: {highlight}")
 
+        # --- Invalidate page-level image cache --------------------------------
+        # The Page.to_image method maintains an internal cache keyed by rendering
+        # parameters.  Because the cache key currently does **not** incorporate
+        # any information about the highlights themselves, it can return stale
+        # images after highlights are added or removed.  To ensure the next
+        # render reflects the new highlights, we clear the cache for the
+        # affected page here.
+        try:
+            page_obj = self._pdf[page_index]
+            if hasattr(page_obj, "_to_image_cache"):
+                page_obj._to_image_cache.clear()
+                logger.debug(
+                    f"Cleared cached to_image renders for page {page_index} after adding a highlight."
+                )
+        except Exception as cache_err:  # pragma: no cover – never fail highlight creation
+            logger.warning(
+                f"Failed to invalidate to_image cache for page {page_index}: {cache_err}",
+                exc_info=True,
+            )
+
     def clear_all(self):
         """Clears all highlights from all pages and resets the color manager."""
         self._highlights_by_page = {}
         self._color_manager.reset()
         logger.info("Cleared all highlights and reset ColorManager.")
 
+        # Clear cached images for *all* pages because their visual state may
+        # depend on highlight visibility.
+        for idx, page in enumerate(self._pdf.pages):
+            try:
+                if hasattr(page, "_to_image_cache"):
+                    page._to_image_cache.clear()
+            except Exception:
+                # Non-critical – keep going for remaining pages
+                continue
+
     def clear_page(self, page_index: int):
         """Clears all highlights from a specific page."""
         if page_index in self._highlights_by_page:
             del self._highlights_by_page[page_index]
             logger.debug(f"Cleared highlights for page {page_index}.")
-            # Note: We typically don't reset the color manager when clearing a single page
-            # to maintain color consistency if highlights are added back.
+        
+        # Also clear any cached rendered images for this page so the next render
+        # reflects the removal of highlights.
+        try:
+            page_obj = self._pdf[page_index]
+            if hasattr(page_obj, "_to_image_cache"):
+                page_obj._to_image_cache.clear()
+                logger.debug(
+                    f"Cleared cached to_image renders for page {page_index} after removing highlights."
+                )
+        except Exception as cache_err:  # pragma: no cover
+            logger.warning(
+                f"Failed to invalidate to_image cache for page {page_index}: {cache_err}",
+                exc_info=True,
+            )
 
     def get_highlights_for_page(self, page_index: int) -> List[Highlight]:
         """Returns a list of Highlight objects for a specific page."""
@@ -581,11 +624,10 @@ class HighlightingService:
     def render_page(
         self,
         page_index: int,
-        scale: float = 2.0,
+        resolution: float = 144,
         labels: bool = True,
         legend_position: str = "right",
         render_ocr: bool = False,
-        resolution: Optional[float] = None,
         **kwargs,  # Pass other args to pdfplumber.page.to_image if needed
     ) -> Optional[Image.Image]:
         """
@@ -594,12 +636,11 @@ class HighlightingService:
 
         Args:
             page_index: The 0-based index of the page to render.
-            scale: Scale factor for rendering highlights if width/height/resolution not in kwargs.
+            resolution: Resolution (DPI) for the base page image if width/height not in kwargs.
+                       Defaults to 144 DPI (equivalent to previous scale=2.0).
             labels: Whether to include a legend for highlights.
             legend_position: Position of the legend.
             render_ocr: Whether to render OCR text on the image.
-            resolution: Optional resolution (DPI) for the base page image if width/height not in kwargs.
-                       Defaults to scale * 72 if not otherwise specified.
             kwargs: Additional keyword arguments for pdfplumber's page.to_image (e.g., width, height).
 
         Returns:
@@ -625,13 +666,16 @@ class HighlightingService:
             logger.debug(f"Rendering page {page_index} with height={to_image_args['height']}.")
             # Actual scale will be calculated after image creation
         else:
-            # Use explicit resolution from kwargs if present, then the resolution param, then scale
-            render_resolution = to_image_args.pop(
-                "resolution", resolution
-            )  # Use and remove from kwargs if present
+            # Use explicit resolution if provided via kwargs, otherwise fallback to the
+            # `resolution` parameter (which might be None).  If we still end up with
+            # `None`, default to 144 DPI to avoid downstream errors.
+            render_resolution = to_image_args.pop("resolution", resolution)
             if render_resolution is None:
-                render_resolution = scale * 72
-            to_image_args["resolution"] = render_resolution  # Add it back for the call
+                render_resolution = 144
+
+            # Reinstate into kwargs for pdfplumber
+            to_image_args["resolution"] = render_resolution
+
             actual_scale_x = render_resolution / 72.0
             actual_scale_y = render_resolution / 72.0
             logger.debug(
@@ -657,11 +701,11 @@ class HighlightingService:
                 if page_obj.width > 0:
                     actual_scale_x = base_image_pil.width / page_obj.width
                 else:
-                    actual_scale_x = scale  # Fallback
+                    actual_scale_x = resolution / 72.0  # Fallback to resolution-based scale
                 if page_obj.height > 0:
                     actual_scale_y = base_image_pil.height / page_obj.height
                 else:
-                    actual_scale_y = scale  # Fallback
+                    actual_scale_y = resolution / 72.0  # Fallback to resolution-based scale
                 logger.debug(
                     f"Calculated actual scales for page {page_index}: x={actual_scale_x:.2f}, y={actual_scale_y:.2f}"
                 )
@@ -682,14 +726,20 @@ class HighlightingService:
                 page=page_obj,
                 base_image=base_image_pil,
                 highlights=highlights_on_page,
-                scale=renderer_scale,  # Use the determined actual scale
+                scale_factor=renderer_scale,  # Use the determined actual scale
                 render_ocr=render_ocr,
             )
             rendered_image = renderer.render()
         else:
             if render_ocr:
                 # Still render OCR even if no highlights, using the determined actual scale
-                renderer = HighlightRenderer(page_obj, base_image_pil, [], renderer_scale, True)
+                renderer = HighlightRenderer(
+                    page=page_obj,
+                    base_image=base_image_pil,
+                    highlights=[],
+                    scale_factor=renderer_scale,
+                    render_ocr=True,
+                )
                 rendered_image = renderer.render()
             else:
                 rendered_image = base_image_pil  # No highlights, no OCR requested
@@ -722,11 +772,10 @@ class HighlightingService:
         self,
         page_index: int,
         temporary_highlights: List[Dict],
-        scale: float = 2.0,
+        resolution: float = 144,
         labels: bool = True,
         legend_position: str = "right",
         render_ocr: bool = False,
-        resolution: Optional[float] = None,
         crop_bbox: Optional[Tuple[float, float, float, float]] = None,
         **kwargs,
     ) -> Optional[Image.Image]:
@@ -737,11 +786,11 @@ class HighlightingService:
         Args:
             page_index: Index of the page to render.
             temporary_highlights: List of highlight data dicts (from ElementCollection._prepare).
-            scale: Original scale factor for rendering, used if width/height are not provided.
+            resolution: Resolution (DPI) for base page image rendering if width/height not used.
+                       Defaults to 144 DPI (equivalent to previous scale=2.0).
             labels: Whether to include a legend.
             legend_position: Position of the legend.
             render_ocr: Whether to render OCR text.
-            resolution: Resolution for base page image rendering if width/height not used.
             crop_bbox: Optional bounding box (x0, top, x1, bottom) in PDF coordinate
                 space to crop the output image to, before legends or other overlays are
                 applied. If None, no cropping is performed.
@@ -777,9 +826,11 @@ class HighlightingService:
             # Resolution is implicitly handled by pdfplumber when height is set
             # after image is created, we will calculate actual_scale_x and actual_scale_y
         else:
-            # Neither width nor height is provided, use resolution or scale.
-            render_resolution = resolution if resolution is not None else scale * 72
+            # Neither width nor height is provided, rely on `resolution`.
+            # If `resolution` was explicitly passed as `None`, fall back to 144 DPI.
+            render_resolution = 144 if resolution is None else resolution
             to_image_args["resolution"] = render_resolution
+
             actual_scale_x = render_resolution / 72.0
             actual_scale_y = render_resolution / 72.0
             logger.debug(
@@ -804,11 +855,11 @@ class HighlightingService:
                 if page_obj.width > 0:
                     actual_scale_x = base_image_pil.width / page_obj.width
                 else:
-                    actual_scale_x = scale  # Fallback to original scale
+                    actual_scale_x = resolution / 72.0  # Fallback to resolution-based scale
                 if page_obj.height > 0:
                     actual_scale_y = base_image_pil.height / page_obj.height
                 else:
-                    actual_scale_y = scale  # Fallback to original scale
+                    actual_scale_y = resolution / 72.0  # Fallback to resolution-based scale
                 logger.debug(
                     f"Calculated actual scales for page {page_index}: x={actual_scale_x:.2f}, y={actual_scale_y:.2f} from image size {base_image_pil.size} and page size ({page_obj.width}, {page_obj.height})"
                 )
@@ -855,7 +906,11 @@ class HighlightingService:
             renderer_scale = actual_scale_x
 
             renderer = HighlightRenderer(
-                page_obj, base_image_pil, preview_highlights, renderer_scale, render_ocr
+                page=page_obj,
+                base_image=base_image_pil,
+                highlights=preview_highlights,
+                scale_factor=renderer_scale,
+                render_ocr=render_ocr,
             )
             rendered_image = renderer.render()
 

natural_pdf/core/page.py

@@ -867,6 +867,28 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
             >>> page.region(right=200, width=50)  # Region from x=150 to x=200
             >>> page.region(top=100, bottom=200, width="full") # Explicit full width
         """
+        # ------------------------------------------------------------------
+        # Percentage support – convert strings like "30%" to absolute values
+        # based on page dimensions.  X-axis params (left, right, width) use
+        # page.width; Y-axis params (top, bottom, height) use page.height.
+        # ------------------------------------------------------------------
+
+        def _pct_to_abs(val, axis: str):
+            if isinstance(val, str) and val.strip().endswith("%"):
+                try:
+                    pct = float(val.strip()[:-1]) / 100.0
+                except ValueError:
+                    return val  # leave unchanged if not a number
+                return pct * (self.width if axis == "x" else self.height)
+            return val
+
+        left = _pct_to_abs(left, "x")
+        right = _pct_to_abs(right, "x")
+        width = _pct_to_abs(width, "x")
+        top = _pct_to_abs(top, "y")
+        bottom = _pct_to_abs(bottom, "y")
+        height = _pct_to_abs(height, "y")
+
         # --- Type checking and basic validation ---
         is_width_numeric = isinstance(width, (int, float))
         is_width_string = isinstance(width, str)
@@ -1137,6 +1159,40 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
             user_kwargs=kwargs,  # Pass original user kwargs
         )
 
+        # --- Optional: apply Unicode BiDi algorithm for mixed RTL/LTR correctness ---
+        apply_bidi = kwargs.get("bidi", True)
+        if apply_bidi and result:
+            # Quick check for any RTL character
+            import unicodedata
+
+            def _contains_rtl(s):
+                return any(unicodedata.bidirectional(ch) in ("R", "AL", "AN") for ch in s)
+
+            if _contains_rtl(result):
+                try:
+                    from bidi.algorithm import get_display  # type: ignore
+                    from natural_pdf.utils.bidi_mirror import mirror_brackets
+
+                    result = "\n".join(
+                        mirror_brackets(
+                            get_display(
+                                line,
+                                base_dir=(
+                                    "R"
+                                    if any(
+                                        unicodedata.bidirectional(ch)
+                                        in ("R", "AL", "AN")
+                                        for ch in line
+                                    )
+                                    else "L"
+                                ),
+                            )
+                        )
+                        for line in result.split("\n")
+                    )
+                except ModuleNotFoundError:
+                    pass  # silently skip if python-bidi not available
+
         logger.debug(f"Page {self.number}: extract_text finished, result length: {len(result)}.")
         return result
 
@@ -1440,7 +1496,7 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
 
     def show(
         self,
-        scale: float = 2.0,
+        resolution: float = 144,
         width: Optional[int] = None,
         labels: bool = True,
         legend_position: str = "right",
@@ -1450,7 +1506,7 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
         Generates and returns an image of the page with persistent highlights rendered.
 
         Args:
-            scale: Scale factor for rendering.
+            resolution: Resolution in DPI for rendering (default: 144 DPI, equivalent to previous scale=2.0).
             width: Optional width for the output image.
             labels: Whether to include a legend for labels.
             legend_position: Position of the legend.
@@ -1460,7 +1516,7 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
             PIL Image object of the page with highlights, or None if rendering fails.
         """
         return self.to_image(
-            scale=scale,
+            resolution=resolution,
             width=width,
             labels=labels,
             legend_position=legend_position,
@@ -1471,13 +1527,12 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
     def save_image(
         self,
         filename: str,
-        scale: float = 2.0,
         width: Optional[int] = None,
         labels: bool = True,
         legend_position: str = "right",
         render_ocr: bool = False,
         include_highlights: bool = True,  # Allow saving without highlights
-        resolution: Optional[float] = None,
+        resolution: float = 144,
         **kwargs,
     ) -> "Page":
         """
@@ -1485,13 +1540,12 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
 
         Args:
             filename: Path to save the image to.
-            scale: Scale factor for rendering highlights.
             width: Optional width for the output image.
             labels: Whether to include a legend.
             legend_position: Position of the legend.
             render_ocr: Whether to render OCR text.
             include_highlights: Whether to render highlights.
-            resolution: Resolution for base image rendering.
+            resolution: Resolution in DPI for base image rendering (default: 144 DPI, equivalent to previous scale=2.0).
             **kwargs: Additional args for pdfplumber's to_image.
 
         Returns:
@@ -1500,7 +1554,6 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
         # Use to_image to generate and save the image
         self.to_image(
             path=filename,
-            scale=scale,
             width=width,
             labels=labels,
             legend_position=legend_position,
@@ -1554,7 +1607,6 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
     def to_image(
         self,
         path: Optional[str] = None,
-        scale: float = 2.0,
         width: Optional[int] = None,
         labels: bool = True,
         legend_position: str = "right",
@@ -1569,12 +1621,11 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
 
         Args:
             path: Optional path to save the image to.
-            scale: Scale factor for rendering highlights.
             width: Optional width for the output image.
             labels: Whether to include a legend for highlights.
             legend_position: Position of the legend.
             render_ocr: Whether to render OCR text on highlights.
-            resolution: Resolution in DPI for base page image (default: scale * 72).
+            resolution: Resolution in DPI for base page image. If None, uses global setting or defaults to 144 DPI.
             include_highlights: Whether to render highlights.
             exclusions: Accepts one of the following:
                         • None  – no masking (default)
@@ -1593,11 +1644,13 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
         # Use global options if parameters are not explicitly set
         if width is None:
             width = natural_pdf.options.image.width
-        if resolution is None and natural_pdf.options.image.resolution is not None:
-            resolution = natural_pdf.options.image.resolution
+        if resolution is None:
+            if natural_pdf.options.image.resolution is not None:
+                resolution = natural_pdf.options.image.resolution
+            else:
+                resolution = 144  # Default resolution when none specified
         # 1. Create cache key (excluding path)
         cache_key_parts = [
-            scale,
             width,
             labels,
             legend_position,
@@ -1641,7 +1694,7 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
             rendered_image_component: Optional[Image.Image] = (
                 None  # Renamed from 'image' in original
             )
-            render_resolution = resolution if resolution is not None else scale * 72
+            render_resolution = resolution
             thread_id = threading.current_thread().name
             logger.debug(
                 f"[{thread_id}] Page {self.index}: Attempting to acquire pdf_render_lock for to_image..."
@@ -1658,11 +1711,10 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
                         # Delegate rendering to the central service
                         rendered_image_component = self._highlighter.render_page(
                             page_index=self.index,
-                            scale=scale,
+                            resolution=render_resolution,
                             labels=labels,
                             legend_position=legend_position,
                             render_ocr=render_ocr,
-                            resolution=render_resolution,  # Pass the calculated resolution
                             **kwargs,
                         )
                     else:
@@ -2336,7 +2388,7 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
     def show_preview(
         self,
         temporary_highlights: List[Dict],
-        scale: float = 2.0,
+        resolution: float = 144,
         width: Optional[int] = None,
         labels: bool = True,
         legend_position: str = "right",
@@ -2349,7 +2401,7 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
         Args:
             temporary_highlights: List of highlight data dictionaries (as prepared by
                                   ElementCollection._prepare_highlight_data).
-            scale: Scale factor for rendering.
+            resolution: Resolution in DPI for rendering (default: 144 DPI, equivalent to previous scale=2.0).
             width: Optional width for the output image.
             labels: Whether to include a legend.
             legend_position: Position of the legend.
@@ -2363,7 +2415,7 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
             img = self._highlighter.render_preview(
                 page_index=self.index,
                 temporary_highlights=temporary_highlights,
-                scale=scale,
+                resolution=resolution,
                 labels=labels,
                 legend_position=legend_position,
                 render_ocr=render_ocr,
@@ -2897,3 +2949,17 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
             properties, and other details for each element
         """
         return self.find_all('*').inspect(limit=limit)
+
+    @property
+    def lines(self) -> List[Any]:
+        """Get all line elements on this page."""
+        return self._element_mgr.lines
+
+    # ------------------------------------------------------------------
+    # Image elements
+    # ------------------------------------------------------------------
+
+    @property
+    def images(self) -> List[Any]:
+        """Get all embedded raster images on this page."""
+        return self._element_mgr.images

natural_pdf/core/pdf.py

@@ -653,8 +653,6 @@ class PDF(ExtractionMixin, ExportMixin, ClassificationMixin):
             raise ValueError("Internal error: No selector or text provided.")
 
         selector_obj = parse_selector(effective_selector)
-        kwargs["regex"] = regex
-        kwargs["case"] = case
 
         # Search page by page
         for page in self.pages:

natural_pdf/describe/base.py

@@ -269,15 +269,28 @@ def _get_columns_for_type(element_type: str, show_page_column: bool) -> List[str
     base_columns = ['x0', 'top', 'x1', 'bottom']
     
     if element_type == 'word':
-        columns = ['text'] + base_columns + ['font_family', 'font_variant', 'size', 'bold', 'italic', 'source', 'confidence']
-        # Add color for text elements
+        columns = ['text'] + base_columns + [
+            'font_family',
+            'font_variant',
+            'size',
+            'bold',
+            'italic',
+            'strike',
+            'underline',
+            'highlight',
+            'source',
+            'confidence',
+        ]
+        # Add foreground text colour too
         columns.append('color')
     elif element_type == 'rect':
         columns = base_columns + ['width', 'height', 'stroke', 'fill', 'stroke_width']
     elif element_type == 'line':
         columns = base_columns + ['width', 'is_horizontal', 'is_vertical']  # LineElement properties
     elif element_type == 'region':
-        columns = base_columns + ['width', 'height', 'type']
+        columns = base_columns + ['width', 'height', 'type', 'color']
+    elif element_type == 'blob':
+        columns = base_columns + ['width', 'height', 'color']
     else:
         columns = base_columns + ['type']
     
@@ -325,19 +338,37 @@ def _extract_element_value(element: "Element", column: str) -> Any:
                 return fontname.split("+", 1)[0]
             return ''
         
-        elif column in ['bold', 'italic']:
+        elif column in ['bold', 'italic', 'strike', 'underline']:
             value = getattr(element, column, False)
             return value if isinstance(value, bool) else False
         
+        elif column == 'highlight':
+            # If element is highlighted, return its colour; otherwise blank
+            if getattr(element, 'highlight', False):
+                col_val = getattr(element, 'highlight_color', None)
+                if col_val is None:
+                    return 'True'  # fallback if colour missing
+                # Convert tuple to hex
+                if isinstance(col_val, (tuple, list)) and len(col_val) >= 3:
+                    try:
+                        r, g, b = [int(v * 255) if v <= 1 else int(v) for v in col_val[:3]]
+                        return f"#{r:02x}{g:02x}{b:02x}"
+                    except Exception:
+                        return str(col_val)
+                return str(col_val)
+            return ''
+        
         elif column in ['stroke', 'fill', 'color']:
-            # For rectangles and text, these return color tuples
             value = getattr(element, column, None)
+            # If already a string (e.g. '#ff00aa' or 'red') return as is
+            if isinstance(value, str):
+                return value
+            # If tuple/list convert to hex
             if value and isinstance(value, (tuple, list)) and len(value) >= 3:
-                # Convert to hex color for display
                 try:
                     r, g, b = [int(v * 255) if v <= 1 else int(v) for v in value[:3]]
                     return f"#{r:02x}{g:02x}{b:02x}"
-                except:
+                except Exception:
                     return str(value)
             return ""
         
@@ -406,7 +437,7 @@ def describe_element(element: "Element") -> "ElementSummary":
     
     # Add common text properties - use dict structure for proper list formatting
     text_props = {}
-    for prop in ['font_family', 'size', 'bold', 'italic', 'source', 'confidence']:
+    for prop in ['font_family', 'size', 'bold', 'italic', 'strike', 'underline', 'highlight', 'source', 'confidence']:
         if hasattr(element, prop):
             value = getattr(element, prop)
             if value is not None:
@@ -414,7 +445,7 @@ def describe_element(element: "Element") -> "ElementSummary":
                     text_props[prop] = round(value, 3)
                 elif prop == 'size' and isinstance(value, (int, float)):
                     text_props[prop] = round(value, 1)
-                elif prop in ['bold', 'italic']:
+                elif prop in ['bold', 'italic', 'strike', 'underline']:
                     text_props[prop] = value
                 else:
                     text_props[prop] = value