natural-pdf 0.1.34__tar.gz → 0.1.36__tar.gz

{natural_pdf-0.1.34/natural_pdf.egg-info → natural_pdf-0.1.36}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: natural-pdf
-Version: 0.1.34
+Version: 0.1.36
 Summary: A more intuitive interface for working with PDFs
 Author-email: Jonathan Soma <jonathan.soma@gmail.com>
 License-Expression: MIT
@@ -11,6 +11,7 @@ Classifier: Operating System :: OS Independent
 Requires-Python: >=3.9
 Description-Content-Type: text/markdown
 License-File: LICENSE
+Requires-Dist: scikit-learn
 Requires-Dist: markdown
 Requires-Dist: pandas
 Requires-Dist: pdfplumber

natural_pdf-0.1.36/natural_pdf/analyzers/__init__.py

@@ -0,0 +1,33 @@
+"""
+Analyzers for natural_pdf.
+"""
+
+# Import these directly as they don't depend on Region
+from natural_pdf.analyzers.guides import Guides
+from natural_pdf.analyzers.shape_detection_mixin import ShapeDetectionMixin
+from natural_pdf.analyzers.text_options import TextStyleOptions
+from natural_pdf.analyzers.text_structure import TextStyleAnalyzer
+
+# Lazy imports to avoid circular dependencies
+# These will be imported when actually accessed
+def __getattr__(name):
+    if name == "LayoutAnalyzer":
+        from natural_pdf.analyzers.layout.layout_analyzer import LayoutAnalyzer
+        return LayoutAnalyzer
+    elif name == "LayoutManager":
+        from natural_pdf.analyzers.layout.layout_manager import LayoutManager
+        return LayoutManager
+    elif name == "LayoutOptions":
+        from natural_pdf.analyzers.layout.layout_options import LayoutOptions
+        return LayoutOptions
+    raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
+
+__all__ = [
+    "LayoutAnalyzer",
+    "LayoutManager", 
+    "LayoutOptions",
+    "ShapeDetectionMixin",
+    "TextStyleOptions",
+    "TextStyleAnalyzer",
+    "Guides",
+]

{natural_pdf-0.1.34 → natural_pdf-0.1.36}/natural_pdf/analyzers/guides.py

@@ -119,6 +119,8 @@ class GuidesList(UserList):
         align: Literal["left", "right", "center", "between"] = "left",
         outer: bool = True,
         tolerance: float = 5,
+        *,
+        append: bool = False,
     ) -> "Guides":
         """
         Create guides from content markers and add to this axis.
@@ -154,11 +156,17 @@ class GuidesList(UserList):
             tolerance=tolerance,
         )
 
-        # Add to our list
-        if self._axis == "vertical":
-            self.extend(new_guides.vertical)
+        # Replace or append based on parameter
+        if append:
+            if self._axis == "vertical":
+                self.extend(new_guides.vertical)
+            else:
+                self.extend(new_guides.horizontal)
         else:
-            self.extend(new_guides.horizontal)
+            if self._axis == "vertical":
+                self.data = list(new_guides.vertical)
+            else:
+                self.data = list(new_guides.horizontal)
 
         # Remove duplicates while preserving order
         seen = set()
@@ -183,6 +191,7 @@ class GuidesList(UserList):
         *,
         n: Optional[int] = None,
         min_gap: Optional[int] = None,
+        append: bool = False,
         **detect_kwargs,
     ) -> "Guides":
         """
@@ -241,11 +250,17 @@ class GuidesList(UserList):
             **detect_kwargs,
         )
 
-        # Add to our list
-        if self._axis == "vertical":
-            self.extend(new_guides.vertical)
+        # Replace or append based on parameter
+        if append:
+            if self._axis == "vertical":
+                self.extend(new_guides.vertical)
+            else:
+                self.extend(new_guides.horizontal)
         else:
-            self.extend(new_guides.horizontal)
+            if self._axis == "vertical":
+                self.data = list(new_guides.vertical)
+            else:
+                self.data = list(new_guides.horizontal)
 
         # Remove duplicates
         seen = set()
@@ -259,7 +274,8 @@ class GuidesList(UserList):
         return self._parent
 
     def from_whitespace(
-        self, obj: Optional[Union["Page", "Region"]] = None, min_gap: float = 10
+        self, obj: Optional[Union["Page", "Region"]] = None, min_gap: float = 10,
+        *, append: bool = False
     ) -> "Guides":
         """
         Create guides from whitespace gaps.
@@ -278,11 +294,17 @@ class GuidesList(UserList):
         # Create guides for this axis
         new_guides = Guides.from_whitespace(obj=target_obj, axis=self._axis, min_gap=min_gap)
 
-        # Add to our list
-        if self._axis == "vertical":
-            self.extend(new_guides.vertical)
+        # Replace or append
+        if append:
+            if self._axis == "vertical":
+                self.extend(new_guides.vertical)
+            else:
+                self.extend(new_guides.horizontal)
         else:
-            self.extend(new_guides.horizontal)
+            if self._axis == "vertical":
+                self.data = list(new_guides.vertical)
+            else:
+                self.data = list(new_guides.horizontal)
 
         # Remove duplicates
         seen = set()
@@ -313,11 +335,11 @@ class GuidesList(UserList):
         # Create guides using divide
         new_guides = Guides.divide(obj=target_obj, n=n, axis=self._axis)
 
-        # Add to our list
+        # Replace existing guides instead of extending (no append option here)
         if self._axis == "vertical":
-            self.extend(new_guides.vertical)
+            self.data = list(new_guides.vertical)
         else:
-            self.extend(new_guides.horizontal)
+            self.data = list(new_guides.horizontal)
 
         # Remove duplicates
         seen = set()
@@ -1982,6 +2004,53 @@ class Guides:
         row_boundaries = sorted(list(set(row_boundaries)))
         col_boundaries = sorted(list(set(col_boundaries)))
 
+        # ------------------------------------------------------------------
+        # Clean-up: remove any previously created grid regions (table, rows,
+        # columns, cells) that were generated by the same `source` label and
+        # overlap the area we are about to populate.  This prevents the page's
+        # `ElementManager` from accumulating stale/duplicate regions when the
+        # user rebuilds the grid multiple times.
+        # ------------------------------------------------------------------
+        try:
+            # Bounding box of the grid we are about to create
+            if row_boundaries and col_boundaries:
+                grid_bbox = (
+                    col_boundaries[0],  # x0
+                    row_boundaries[0],  # top
+                    col_boundaries[-1],  # x1
+                    row_boundaries[-1],  # bottom
+                )
+
+                def _bbox_overlap(b1, b2):
+                    """Return True if two (x0, top, x1, bottom) bboxes overlap."""
+                    return not (
+                        b1[2] <= b2[0]  # b1 right ≤ b2 left
+                        or b1[0] >= b2[2]  # b1 left ≥ b2 right
+                        or b1[3] <= b2[1]  # b1 bottom ≤ b2 top
+                        or b1[1] >= b2[3]  # b1 top ≥ b2 bottom
+                    )
+
+                # Collect existing regions that match the source & region types
+                regions_to_remove = [
+                    r
+                    for r in element_manager.regions
+                    if getattr(r, "source", None) == source
+                    and getattr(r, "region_type", None)
+                    in {"table", "table_row", "table_column", "table_cell"}
+                    and hasattr(r, "bbox")
+                    and _bbox_overlap(r.bbox, grid_bbox)
+                ]
+
+                for r in regions_to_remove:
+                    element_manager.remove_element(r, element_type="regions")
+
+                if regions_to_remove:
+                    logger.debug(
+                        f"Removed {len(regions_to_remove)} existing grid region(s) prior to rebuild"
+                    )
+        except Exception as cleanup_err:  # pragma: no cover – cleanup must never crash
+            logger.warning(f"Grid cleanup failed: {cleanup_err}")
+
         logger.debug(
             f"Building grid with {len(row_boundaries)} row and {len(col_boundaries)} col boundaries"
         )

{natural_pdf-0.1.34 → natural_pdf-0.1.36}/natural_pdf/core/page.py

@@ -1780,8 +1780,14 @@ class Page(ClassificationMixin, ExtractionMixin, ShapeDetectionMixin, DescribeMi
         # Apply global options as defaults, but allow explicit parameters to override
         import natural_pdf
 
+        # Determine if this is likely a computational use (OCR, analysis, etc.)
+        # If resolution is explicitly provided but width is not, assume computational use
+        # and don't apply global display width settings
+        is_computational_use = (resolution is not None and width is None and 
+                               kwargs.get('include_highlights', True) is False)
+
         # Use global options if parameters are not explicitly set
-        if width is None:
+        if width is None and not is_computational_use:
             width = natural_pdf.options.image.width
         if resolution is None:
             if natural_pdf.options.image.resolution is not None:

{natural_pdf-0.1.34 → natural_pdf-0.1.36/natural_pdf.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: natural-pdf
-Version: 0.1.34
+Version: 0.1.36
 Summary: A more intuitive interface for working with PDFs
 Author-email: Jonathan Soma <jonathan.soma@gmail.com>
 License-Expression: MIT
@@ -11,6 +11,7 @@ Classifier: Operating System :: OS Independent
 Requires-Python: >=3.9
 Description-Content-Type: text/markdown
 License-File: LICENSE
+Requires-Dist: scikit-learn
 Requires-Dist: markdown
 Requires-Dist: pandas
 Requires-Dist: pdfplumber

{natural_pdf-0.1.34 → natural_pdf-0.1.36}/natural_pdf.egg-info/requires.txt

@@ -1,3 +1,4 @@
+scikit-learn
 markdown
 pandas
 pdfplumber

{natural_pdf-0.1.34 → natural_pdf-0.1.36}/pyproject.toml

@@ -19,6 +19,7 @@ classifiers = [
     "Operating System :: OS Independent",
 ]
 dependencies = [
+    "scikit-learn",
     "markdown",
     "pandas",
     "pdfplumber",

natural_pdf-0.1.34/natural_pdf/analyzers/__init__.py

@@ -1,21 +0,0 @@
-"""
-Analyzers for natural_pdf.
-"""
-
-from natural_pdf.analyzers.guides import Guides
-from natural_pdf.analyzers.layout.layout_analyzer import LayoutAnalyzer
-from natural_pdf.analyzers.layout.layout_manager import LayoutManager
-from natural_pdf.analyzers.layout.layout_options import LayoutOptions
-from natural_pdf.analyzers.shape_detection_mixin import ShapeDetectionMixin
-from natural_pdf.analyzers.text_options import TextStyleOptions
-from natural_pdf.analyzers.text_structure import TextStyleAnalyzer
-
-__all__ = [
-    "LayoutAnalyzer",
-    "LayoutManager",
-    "LayoutOptions",
-    "ShapeDetectionMixin",
-    "TextStyleOptions",
-    "TextStyleAnalyzer",
-    "Guides",
-]