natural-pdf 0.2.10__py3-none-any.whl → 0.2.12__py3-none-any.whl

natural_pdf/analyzers/guides.py

@@ -185,7 +185,9 @@ class GuidesList(UserList):
         self,
         markers: Union[str, List[str], "ElementCollection", Callable, None],
         obj: Optional[Union["Page", "Region", "FlowRegion"]] = None,
-        align: Literal["left", "right", "center", "between"] = "left",
+        align: Union[
+            Literal["left", "right", "center", "between"], Literal["top", "bottom"]
+        ] = "left",
         outer: bool = True,
         tolerance: float = 5,
         *,
@@ -203,7 +205,10 @@ class GuidesList(UserList):
                 - Callable: function that takes a page and returns markers
                 - None: no markers
             obj: Page/Region/FlowRegion to search (uses parent's context if None)
-            align: How to align guides relative to found elements
+            align: How to align guides relative to found elements:
+                - For vertical guides: 'left', 'right', 'center', 'between'
+                - For horizontal guides: 'top', 'bottom', 'center', 'between'
+                - Note: 'left'/'right' also work for horizontal (mapped to top/bottom)
             outer: Whether to add outer boundary guides
             tolerance: Tolerance for snapping to element edges
             apply_exclusions: Whether to apply exclusion zones when searching for text
@@ -224,19 +229,25 @@ class GuidesList(UserList):
             self._callable = None
             actual_markers = markers
 
+        # Normalize alignment for horizontal guides
+        if self._axis == "horizontal":
+            if align == "top":
+                align = "left"
+            elif align == "bottom":
+                align = "right"
+
         # Check if parent is in flow mode
         if self._parent.is_flow_region:
             # Create guides across all constituent regions
             all_guides = []
             for region in self._parent.context.constituent_regions:
-                # Normalize markers for this region
-                marker_texts = _normalize_markers(actual_markers, region)
+                # Pass markers directly - from_content will handle them properly
 
                 # Create guides for this region
                 region_guides = Guides.from_content(
                     obj=region,
                     axis=self._axis,
-                    markers=marker_texts,
+                    markers=actual_markers,  # Pass original markers, not normalized text
                     align=align,
                     outer=outer,
                     tolerance=tolerance,
@@ -312,14 +323,14 @@ class GuidesList(UserList):
             return self._parent
 
         # Original single-region logic
-        # Normalize markers to list of text strings
-        marker_texts = _normalize_markers(actual_markers, target_obj)
+        # Pass markers directly to from_content which will handle them properly
+        # (no need to normalize here since from_content now handles ElementCollection)
 
         # Create guides for this axis
         new_guides = Guides.from_content(
             obj=target_obj,
             axis=self._axis,
-            markers=marker_texts,
+            markers=actual_markers,  # Pass original markers, not normalized text
             align=align,
             outer=outer,
             tolerance=tolerance,
@@ -930,6 +941,82 @@ class GuidesList(UserList):
         self.data.clear()
         return self._parent
 
+    def from_stripes(
+        self,
+        stripes=None,
+        color=None,  # Explicitly specify stripe color
+    ) -> "Guides":
+        """Create guides from striped table rows or columns.
+
+        Creates guides at both edges of stripe elements (e.g., colored table rows).
+        Perfect for zebra-striped tables where you need guides at every row boundary.
+
+        Args:
+            stripes: Elements representing stripes. If None, auto-detects.
+            color: Specific color to look for (e.g., '#00ffff'). If None, finds most common.
+
+        Examples:
+            # Auto-detect zebra stripes
+            guides.horizontal.from_stripes()
+
+            # Specific color
+            guides.horizontal.from_stripes(color='#00ffff')
+
+            # Manual selection
+            stripes = page.find_all('rect[fill=#00ffff]')
+            guides.horizontal.from_stripes(stripes)
+
+            # Vertical stripes
+            guides.vertical.from_stripes(color='#e0e0e0')
+
+        Returns:
+            Parent Guides object for chaining
+        """
+        from collections import defaultdict
+
+        target_obj = self._parent.context
+        if target_obj is None:
+            raise ValueError("No context available for stripe detection")
+
+        if stripes is None:
+            if color:
+                # User specified color
+                stripes = target_obj.find_all(f"rect[fill={color}]")
+            else:
+                # Auto-detect most common non-white fill
+                all_rects = target_obj.find_all("rect[fill]")
+
+                # Group by fill color
+                fill_counts = defaultdict(list)
+                for rect in all_rects:
+                    if rect.fill and rect.fill not in ["#ffffff", "white", "none", "transparent"]:
+                        fill_counts[rect.fill].append(rect)
+
+                if not fill_counts:
+                    return self._parent  # No stripes found
+
+                # Find most common fill color
+                stripes = max(fill_counts.values(), key=len)
+
+        if not stripes:
+            return self._parent
+
+        # Get both edges of each stripe
+        edges = []
+        if self._axis == "horizontal":
+            for stripe in stripes:
+                edges.extend([stripe.top, stripe.bottom])
+        else:
+            for stripe in stripes:
+                edges.extend([stripe.x0, stripe.x1])
+
+        # Remove duplicates and sort
+        edges = sorted(set(edges))
+
+        # Add guides
+        self.extend(edges)
+        return self._parent
+
     def __add__(self, other):
         """Handle addition of GuidesList objects by returning combined data."""
         if isinstance(other, GuidesList):
@@ -1459,7 +1546,9 @@ class Guides:
         obj: Union["Page", "Region", "FlowRegion"],
         axis: Literal["vertical", "horizontal"] = "vertical",
         markers: Union[str, List[str], "ElementCollection", None] = None,
-        align: Literal["left", "right", "center", "between"] = "left",
+        align: Union[
+            Literal["left", "right", "center", "between"], Literal["top", "bottom"]
+        ] = "left",
         outer: bool = True,
         tolerance: float = 5,
         apply_exclusions: bool = True,
@@ -1475,7 +1564,9 @@ class Guides:
                 - List[str]: list of selectors or literal text strings
                 - ElementCollection: collection of elements to extract text from
                 - None: no markers
-            align: Where to place guides relative to found text
+            align: Where to place guides relative to found text:
+                - For vertical guides: 'left', 'right', 'center', 'between'
+                - For horizontal guides: 'top', 'bottom', 'center', 'between'
             outer: Whether to add guides at the boundaries
             tolerance: Maximum distance to search for text
             apply_exclusions: Whether to apply exclusion zones when searching for text
@@ -1483,6 +1574,13 @@ class Guides:
         Returns:
             New Guides object aligned to text content
         """
+        # Normalize alignment for horizontal guides
+        if axis == "horizontal":
+            if align == "top":
+                align = "left"
+            elif align == "bottom":
+                align = "right"
+
         # Handle FlowRegion
         if hasattr(obj, "constituent_regions"):
             guides = cls(context=obj)
@@ -1530,39 +1628,51 @@ class Guides:
         elif hasattr(obj, "width"):
             bounds = (0, 0, obj.width, obj.height)
 
-        # Normalize markers to list of text strings
-        marker_texts = _normalize_markers(markers, obj)
+        # Handle different marker types
+        elements_to_process = []
 
-        # Find each marker and determine guide position
-        for marker in marker_texts:
-            if hasattr(obj, "find"):
-                element = obj.find(f'text:contains("{marker}")', apply_exclusions=apply_exclusions)
-                if element:
-                    if axis == "vertical":
-                        if align == "left":
-                            guides_coords.append(element.x0)
-                        elif align == "right":
-                            guides_coords.append(element.x1)
-                        elif align == "center":
-                            guides_coords.append((element.x0 + element.x1) / 2)
-                        elif align == "between":
-                            # For between, collect left edges for processing later
-                            guides_coords.append(element.x0)
-                    else:  # horizontal
-                        if align == "left":  # top for horizontal
-                            guides_coords.append(element.top)
-                        elif align == "right":  # bottom for horizontal
-                            guides_coords.append(element.bottom)
-                        elif align == "center":
-                            guides_coords.append((element.top + element.bottom) / 2)
-                        elif align == "between":
-                            # For between, collect top edges for processing later
-                            guides_coords.append(element.top)
+        # Check if markers is an ElementCollection or has elements attribute
+        if hasattr(markers, "elements") or hasattr(markers, "_elements"):
+            # It's an ElementCollection - use elements directly
+            elements_to_process = getattr(markers, "elements", getattr(markers, "_elements", []))
+        elif hasattr(markers, "__iter__") and not isinstance(markers, str):
+            # Check if it's an iterable of elements (not strings)
+            try:
+                markers_list = list(markers)
+                if markers_list and hasattr(markers_list[0], "x0"):
+                    # It's a list of elements
+                    elements_to_process = markers_list
+            except:
+                pass
 
-        # Handle 'between' alignment - find midpoints between adjacent markers
-        if align == "between" and len(guides_coords) >= 2:
-            # We need to get the right and left edges of each marker
-            marker_bounds = []
+        if elements_to_process:
+            # Process elements directly without text search
+            for element in elements_to_process:
+                if axis == "vertical":
+                    if align == "left":
+                        guides_coords.append(element.x0)
+                    elif align == "right":
+                        guides_coords.append(element.x1)
+                    elif align == "center":
+                        guides_coords.append((element.x0 + element.x1) / 2)
+                    elif align == "between":
+                        # For between, collect left edges for processing later
+                        guides_coords.append(element.x0)
+                else:  # horizontal
+                    if align == "left":  # top for horizontal
+                        guides_coords.append(element.top)
+                    elif align == "right":  # bottom for horizontal
+                        guides_coords.append(element.bottom)
+                    elif align == "center":
+                        guides_coords.append((element.top + element.bottom) / 2)
+                    elif align == "between":
+                        # For between, collect top edges for processing later
+                        guides_coords.append(element.top)
+        else:
+            # Fall back to text-based search
+            marker_texts = _normalize_markers(markers, obj)
+
+            # Find each marker and determine guide position
             for marker in marker_texts:
                 if hasattr(obj, "find"):
                     element = obj.find(
@@ -1570,9 +1680,52 @@ class Guides:
                     )
                     if element:
                         if axis == "vertical":
-                            marker_bounds.append((element.x0, element.x1))
+                            if align == "left":
+                                guides_coords.append(element.x0)
+                            elif align == "right":
+                                guides_coords.append(element.x1)
+                            elif align == "center":
+                                guides_coords.append((element.x0 + element.x1) / 2)
+                            elif align == "between":
+                                # For between, collect left edges for processing later
+                                guides_coords.append(element.x0)
                         else:  # horizontal
-                            marker_bounds.append((element.top, element.bottom))
+                            if align == "left":  # top for horizontal
+                                guides_coords.append(element.top)
+                            elif align == "right":  # bottom for horizontal
+                                guides_coords.append(element.bottom)
+                            elif align == "center":
+                                guides_coords.append((element.top + element.bottom) / 2)
+                            elif align == "between":
+                                # For between, collect top edges for processing later
+                                guides_coords.append(element.top)
+
+        # Handle 'between' alignment - find midpoints between adjacent markers
+        if align == "between" and len(guides_coords) >= 2:
+            # We need to get the right and left edges of each marker
+            marker_bounds = []
+
+            if elements_to_process:
+                # Use elements directly
+                for element in elements_to_process:
+                    if axis == "vertical":
+                        marker_bounds.append((element.x0, element.x1))
+                    else:  # horizontal
+                        marker_bounds.append((element.top, element.bottom))
+            else:
+                # Fall back to text search
+                if "marker_texts" not in locals():
+                    marker_texts = _normalize_markers(markers, obj)
+                for marker in marker_texts:
+                    if hasattr(obj, "find"):
+                        element = obj.find(
+                            f'text:contains("{marker}")', apply_exclusions=apply_exclusions
+                        )
+                        if element:
+                            if axis == "vertical":
+                                marker_bounds.append((element.x0, element.x1))
+                            else:  # horizontal
+                                marker_bounds.append((element.top, element.bottom))
 
             # Sort markers by their left edge (or top edge for horizontal)
             marker_bounds.sort(key=lambda x: x[0])
@@ -4246,12 +4399,26 @@ class _ColumnAccessor:
         """Return number of columns (vertical guides - 1)."""
         return max(0, len(self._guides.vertical) - 1)
 
-    def __getitem__(self, index: int) -> "Region":
-        """Get column at the specified index."""
-        # Handle negative indexing
-        if index < 0:
-            index = len(self) + index
-        return self._guides.column(index)
+    def __getitem__(self, index: Union[int, slice]) -> Union["Region", "ElementCollection"]:
+        """Get column at the specified index or slice."""
+        from natural_pdf.elements.element_collection import ElementCollection
+
+        if isinstance(index, slice):
+            # Handle slice notation - return multiple columns
+            columns = []
+            num_cols = len(self)
+
+            # Convert slice to range of indices
+            start, stop, step = index.indices(num_cols)
+            for i in range(start, stop, step):
+                columns.append(self._guides.column(i))
+
+            return ElementCollection(columns)
+        else:
+            # Handle negative indexing
+            if index < 0:
+                index = len(self) + index
+            return self._guides.column(index)
 
 
 class _RowAccessor:
@@ -4264,12 +4431,26 @@ class _RowAccessor:
         """Return number of rows (horizontal guides - 1)."""
         return max(0, len(self._guides.horizontal) - 1)
 
-    def __getitem__(self, index: int) -> "Region":
-        """Get row at the specified index."""
-        # Handle negative indexing
-        if index < 0:
-            index = len(self) + index
-        return self._guides.row(index)
+    def __getitem__(self, index: Union[int, slice]) -> Union["Region", "ElementCollection"]:
+        """Get row at the specified index or slice."""
+        from natural_pdf.elements.element_collection import ElementCollection
+
+        if isinstance(index, slice):
+            # Handle slice notation - return multiple rows
+            rows = []
+            num_rows = len(self)
+
+            # Convert slice to range of indices
+            start, stop, step = index.indices(num_rows)
+            for i in range(start, stop, step):
+                rows.append(self._guides.row(i))
+
+            return ElementCollection(rows)
+        else:
+            # Handle negative indexing
+            if index < 0:
+                index = len(self) + index
+            return self._guides.row(index)
 
 
 class _CellAccessor:
@@ -4278,33 +4459,82 @@ class _CellAccessor:
     def __init__(self, guides: "Guides"):
         self._guides = guides
 
-    def __getitem__(self, key) -> Union["Region", "_CellRowAccessor"]:
+    def __getitem__(self, key) -> Union["Region", "_CellRowAccessor", "ElementCollection"]:
         """
         Get cell(s) at the specified position.
 
         Supports:
-        - guides.cells[row, col] - tuple indexing
-        - guides.cells[row][col] - nested indexing
+        - guides.cells[row, col] - single cell
+        - guides.cells[row][col] - single cell (nested)
+        - guides.cells[row, :] - all cells in a row
+        - guides.cells[:, col] - all cells in a column
+        - guides.cells[:, :] - all cells
+        - guides.cells[row][:] - all cells in a row (nested)
         """
+        from natural_pdf.elements.element_collection import ElementCollection
+
         if isinstance(key, tuple) and len(key) == 2:
-            # Direct tuple access: guides.cells[row, col]
             row, col = key
-            # Handle negative indexing for both row and col
-            if row < 0:
-                row = len(self._guides.rows) + row
-            if col < 0:
-                col = len(self._guides.columns) + col
-            return self._guides.cell(row, col)
+
+            # Handle slices for row and/or column
+            if isinstance(row, slice) or isinstance(col, slice):
+                cells = []
+                num_rows = len(self._guides.rows)
+                num_cols = len(self._guides.columns)
+
+                # Convert slices to ranges
+                if isinstance(row, slice):
+                    row_indices = range(*row.indices(num_rows))
+                else:
+                    # Single row index
+                    if row < 0:
+                        row = num_rows + row
+                    row_indices = [row]
+
+                if isinstance(col, slice):
+                    col_indices = range(*col.indices(num_cols))
+                else:
+                    # Single column index
+                    if col < 0:
+                        col = num_cols + col
+                    col_indices = [col]
+
+                # Collect all cells in the specified ranges
+                for r in row_indices:
+                    for c in col_indices:
+                        cells.append(self._guides.cell(r, c))
+
+                return ElementCollection(cells)
+            else:
+                # Both are integers - single cell access
+                # Handle negative indexing for both row and col
+                if row < 0:
+                    row = len(self._guides.rows) + row
+                if col < 0:
+                    col = len(self._guides.columns) + col
+                return self._guides.cell(row, col)
+        elif isinstance(key, slice):
+            # First level slice: guides.cells[:] - return all rows as accessors
+            # For now, let's return all cells flattened
+            cells = []
+            num_rows = len(self._guides.rows)
+            row_indices = range(*key.indices(num_rows))
+
+            for r in row_indices:
+                for c in range(len(self._guides.columns)):
+                    cells.append(self._guides.cell(r, c))
+
+            return ElementCollection(cells)
         elif isinstance(key, int):
             # First level of nested access: guides.cells[row]
             # Handle negative indexing for row
             if key < 0:
                 key = len(self._guides.rows) + key
-            # Return a row accessor that allows [col] indexing
+            # Return a row accessor that allows [col] or [:] indexing
             return _CellRowAccessor(self._guides, key)
         else:
             raise TypeError(
-                f"Cell indices must be integers or tuple of two integers, got {type(key)}"
+                f"Cell indices must be integers, slices, or tuple of two integers/slices, got {type(key)}"
             )
 
 
@@ -4315,9 +4545,24 @@ class _CellRowAccessor:
         self._guides = guides
         self._row = row
 
-    def __getitem__(self, col: int) -> "Region":
-        """Get cell at [row][col]."""
-        # Handle negative indexing for column
-        if col < 0:
-            col = len(self._guides.columns) + col
-        return self._guides.cell(self._row, col)
+    def __getitem__(self, col: Union[int, slice]) -> Union["Region", "ElementCollection"]:
+        """Get cell at [row][col] or all cells in row with [row][:]."""
+        from natural_pdf.elements.element_collection import ElementCollection
+
+        if isinstance(col, slice):
+            # Handle slice notation - return all cells in this row
+            cells = []
+            num_cols = len(self._guides.columns)
+
+            # Convert slice to range of indices
+            start, stop, step = col.indices(num_cols)
+            for c in range(start, stop, step):
+                cells.append(self._guides.cell(self._row, c))
+
+            return ElementCollection(cells)
+        else:
+            # Handle single column index
+            # Handle negative indexing for column
+            if col < 0:
+                col = len(self._guides.columns) + col
+            return self._guides.cell(self._row, col)

natural_pdf/core/page.py

@@ -815,11 +815,38 @@ class Page(
                             if debug:
                                 print(f"    ✗ Empty iterable returned from callable '{label}'")
                     elif region_result:
-                        logger.warning(
-                            f"Callable exclusion '{exclusion_label}' returned non-Region object: {type(region_result)}. Skipping."
-                        )
-                        if debug:
-                            print(f"    ✗ Callable returned non-Region/None: {type(region_result)}")
+                        # Check if it's a single Element that can be converted to a Region
+                        from natural_pdf.elements.base import Element
+
+                        if isinstance(region_result, Element) or (
+                            hasattr(region_result, "bbox") and hasattr(region_result, "expand")
+                        ):
+                            try:
+                                # Convert Element to Region using expand()
+                                expanded_region = region_result.expand()
+                                if isinstance(expanded_region, Region):
+                                    expanded_region.label = label
+                                    regions.append(expanded_region)
+                                    if debug:
+                                        print(
+                                            f"    ✓ Converted Element to Region from callable '{label}': {expanded_region}"
+                                        )
+                                else:
+                                    if debug:
+                                        print(
+                                            f"    ✗ Element.expand() did not return a Region: {type(expanded_region)}"
+                                        )
+                            except Exception as e:
+                                if debug:
+                                    print(f"    ✗ Failed to convert Element to Region: {e}")
+                        else:
+                            logger.warning(
+                                f"Callable exclusion '{exclusion_label}' returned non-Region object: {type(region_result)}. Skipping."
+                            )
+                            if debug:
+                                print(
+                                    f"    ✗ Callable returned non-Region/None: {type(region_result)}"
+                                )
                     else:
                         if debug:
                             print(
@@ -839,6 +866,27 @@ class Page(
                 if debug:
                     print(f"  - Added direct region '{label}': {exclusion_item}")
 
+            # Process direct Element objects - convert to Region
+            elif hasattr(exclusion_item, "bbox") and hasattr(exclusion_item, "expand"):
+                try:
+                    # Convert Element to Region using expand()
+                    expanded_region = exclusion_item.expand()
+                    if isinstance(expanded_region, Region):
+                        expanded_region.label = label
+                        regions.append(expanded_region)
+                        if debug:
+                            print(
+                                f"  - Converted direct Element to Region '{label}': {expanded_region}"
+                            )
+                    else:
+                        if debug:
+                            print(
+                                f"  - Element.expand() did not return a Region: {type(expanded_region)}"
+                            )
+                except Exception as e:
+                    if debug:
+                        print(f"  - Failed to convert Element to Region: {e}")
+
             # Process string selectors (from PDF-level exclusions)
             elif isinstance(exclusion_item, str):
                 selector_str = exclusion_item
@@ -1081,7 +1129,7 @@ class Page(
         )  # _apply_selector doesn't filter
 
         # Filter the results based on exclusions if requested
-        if apply_exclusions and self._exclusions and results_collection:
+        if apply_exclusions and results_collection:
             filtered_elements = self._filter_elements_by_exclusions(results_collection.elements)
             # Return the first element from the filtered list
             return filtered_elements[0] if filtered_elements else None
@@ -1176,7 +1224,7 @@ class Page(
         )  # _apply_selector doesn't filter
 
         # Filter the results based on exclusions if requested
-        if apply_exclusions and self._exclusions and results_collection:
+        if apply_exclusions and results_collection:
             filtered_elements = self._filter_elements_by_exclusions(results_collection.elements)
             return ElementCollection(filtered_elements)
         else:
@@ -1548,7 +1596,7 @@ class Page(
         all_elements = self._element_mgr.get_all_elements()
 
         # Apply exclusions if requested
-        if apply_exclusions and self._exclusions:
+        if apply_exclusions:
             return self._filter_elements_by_exclusions(
                 all_elements, debug_exclusions=debug_exclusions
             )

natural_pdf/elements/region.py

@@ -1270,7 +1270,8 @@ class Region(
         # 3. Get Relevant Exclusions (overlapping this region)
         apply_exclusions_flag = kwargs.get("apply_exclusions", apply_exclusions)
         exclusion_regions = []
-        if apply_exclusions_flag and self._page._exclusions:
+        if apply_exclusions_flag:
+            # Always call _get_exclusion_regions to get both page and PDF level exclusions
             all_page_exclusions = self._page._get_exclusion_regions(
                 include_callable=True, debug=debug
             )
@@ -1281,10 +1282,11 @@ class Region(
             exclusion_regions = overlapping_exclusions
             if debug:
                 logger.debug(
-                    f"Region {self.bbox}: Applying {len(exclusion_regions)} overlapping exclusions."
+                    f"Region {self.bbox}: Found {len(all_page_exclusions)} total exclusions, "
+                    f"{len(exclusion_regions)} overlapping this region."
                 )
         elif debug:
-            logger.debug(f"Region {self.bbox}: Not applying exclusions.")
+            logger.debug(f"Region {self.bbox}: Not applying exclusions (apply_exclusions=False).")
 
         # 4. Spatially Filter Characters using Utility
         # Pass self as the target_region for precise polygon checks etc.

{natural_pdf-0.2.10.dist-info → natural_pdf-0.2.12.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: natural-pdf
-Version: 0.2.10
+Version: 0.2.12
 Summary: A more intuitive interface for working with PDFs
 Author-email: Jonathan Soma <jonathan.soma@gmail.com>
 License-Expression: MIT