natural-pdf 0.1.37__py3-none-any.whl → 0.1.40__py3-none-any.whl

natural_pdf/flows/region.py

@@ -631,6 +631,12 @@ class FlowRegion:
         text_options: Optional[Dict] = None,
         cell_extraction_func: Optional[Callable[["PhysicalRegion"], Optional[str]]] = None,
         show_progress: bool = False,
+        # Optional row-level merge predicate. If provided, it decides whether
+        # the current row (first row of a segment/page) should be merged with
+        # the previous one (to handle multi-page spill-overs).
+        stitch_rows: Optional[
+            Callable[[List[Optional[str]], List[Optional[str]], int, "PhysicalRegion"], bool]
+        ] = None,
         **kwargs,
     ) -> TableResult:
         """Extracts a single logical table from the FlowRegion.
@@ -651,6 +657,15 @@ class FlowRegion:
             A TableResult object containing the aggregated table data.  Rows returned from
             consecutive constituent regions are appended in document order.  If
             no tables are detected in any region, an empty TableResult is returned.
+
+        stitch_rows parameter:
+            Controls whether the first rows of subsequent segments/regions should be merged
+            into the previous row (to handle spill-over across page breaks).
+
+            • None (default) – no merging (behaviour identical to previous versions).
+            • Callable – custom predicate taking
+                   (prev_row, cur_row, row_idx_in_segment, segment_object) → bool.
+               Return True to merge `cur_row` into `prev_row` (default column-wise merge is used).
         """
 
         if table_settings is None:
@@ -661,9 +676,26 @@ class FlowRegion:
         if not self.constituent_regions:
             return TableResult([])
 
+        # Resolve stitch_rows predicate -------------------------------------------------------
+        predicate: Optional[
+            Callable[[List[Optional[str]], List[Optional[str]], int, "PhysicalRegion"], bool]
+        ] = stitch_rows if callable(stitch_rows) else None
+
+        def _default_merge(prev_row: List[Optional[str]], cur_row: List[Optional[str]]) -> List[Optional[str]]:
+            """Column-wise merge – concatenates non-empty strings with a space."""
+            from itertools import zip_longest
+
+            merged: List[Optional[str]] = []
+            for p, c in zip_longest(prev_row, cur_row, fillvalue=""):
+                if (p or "").strip() and (c or "").strip():
+                    merged.append(f"{p} {c}".strip())
+                else:
+                    merged.append((p or "") + (c or ""))
+            return merged
+
         aggregated_rows: List[List[Optional[str]]] = []
 
-        for region in self.constituent_regions:
+        for region_idx, region in enumerate(self.constituent_regions):
             try:
                 region_result = region.extract_table(
                     method=method,
@@ -676,9 +708,25 @@ class FlowRegion:
                     **kwargs,
                 )
 
-                # region_result is now a TableResult object, extract the rows
-                if region_result:
-                    aggregated_rows.extend(region_result)
+                # Convert result to list of rows
+                if not region_result:
+                    continue
+
+                if isinstance(region_result, TableResult):
+                    segment_rows = list(region_result)
+                else:
+                    segment_rows = list(region_result)
+
+                for row_idx, row in enumerate(segment_rows):
+                    if (
+                        predicate is not None
+                        and aggregated_rows
+                        and predicate(aggregated_rows[-1], row, row_idx, region)
+                    ):
+                        # Merge with previous row
+                        aggregated_rows[-1] = _default_merge(aggregated_rows[-1], row)
+                    else:
+                        aggregated_rows.append(row)
             except Exception as e:
                 logger.error(
                     f"FlowRegion.extract_table: Error extracting table from constituent region {region}: {e}",

natural_pdf/selectors/parser.py

@@ -100,6 +100,12 @@ def safe_parse_color(value_str: str) -> tuple:
         ValueError: If the color cannot be parsed
     """
     value_str = value_str.strip()
+    
+    # Strip quotes first if it's a quoted string (same logic as safe_parse_value)
+    if (value_str.startswith('"') and value_str.endswith('"')) or (
+        value_str.startswith("'") and value_str.endswith("'")
+    ):
+        value_str = value_str[1:-1]
 
     # Try parsing as a Python literal (for RGB tuples)
     try:
@@ -504,6 +510,21 @@ def _is_approximate_match(value1, value2) -> bool:
     return value1 == value2
 
 
+def _is_exact_color_match(value1, value2) -> bool:
+    """
+    Check if two color values match exactly (with small tolerance for color variations).
+
+    For colors: Uses Delta E color difference with strict tolerance of 2.0
+    For non-colors: Falls back to exact equality
+    """
+    # First check if both values are colors
+    if _is_color_value(value1) and _is_color_value(value2):
+        return _color_distance(value1, value2) <= 2.0
+
+    # Default to exact match for non-colors
+    return value1 == value2
+
+
 PSEUDO_CLASS_FUNCTIONS = {
     "bold": lambda el: hasattr(el, "bold") and el.bold,
     "italic": lambda el: hasattr(el, "italic") and el.italic,
@@ -603,7 +624,19 @@ def _build_filter_list(selector: Dict[str, Any], **kwargs) -> List[Dict[str, Any
 
             # Determine compare_func based on op (reuse existing logic)
             if op == "=":
-                compare_func = lambda el_val, sel_val: el_val == sel_val
+                # For color attributes, use exact color matching with small tolerance
+                if name in [
+                    "color",
+                    "non_stroking_color",
+                    "fill",
+                    "stroke",
+                    "strokeColor",
+                    "fillColor",
+                ]:
+                    op_desc = f"= {value!r} (exact color)"
+                    compare_func = lambda el_val, sel_val: _is_exact_color_match(el_val, sel_val)
+                else:
+                    compare_func = lambda el_val, sel_val: el_val == sel_val
             elif op == "!=":
                 compare_func = lambda el_val, sel_val: el_val != sel_val
             elif op == "~=":

natural_pdf/text_mixin.py

@@ -0,0 +1,97 @@
+from __future__ import annotations
+
+import logging
+from typing import Any, Callable, Optional
+
+logger = logging.getLogger(__name__)
+
+
+class TextMixin:  # pylint: disable=too-few-public-methods
+    """Mixin that adds general text-replacement capabilities.
+
+    Two public entry points are exposed to any class that inherits this mix-in:
+
+    1. ``update_text`` (preferred) –  iterate over text elements selected via the
+       ``selector`` argument (default: ``"text"``) and apply a *correction* callback
+       which optionally returns replacement text.  If the callback returns a
+       non-``None`` string that differs from the current value, the element's
+       ``text`` attribute is updated in-place.
+
+    2. ``correct_ocr`` – legacy name kept for backward compatibility.  It simply
+       forwards to :py:meth:`update_text` while forcing
+       ``selector="text[source=ocr]"`` so that the historic behaviour (acting only
+       on OCR-generated elements) is preserved.
+    """
+
+    # ---------------------------------------------------------------------
+    # Back-compat shim
+    # ---------------------------------------------------------------------
+    def correct_ocr(self, *args, selector: str = "text[source=ocr]", **kwargs):  # type: ignore[override]
+        """Backward-compatibility wrapper that forwards to *update_text*.
+
+        Parameters
+        ----------
+        *args, **kwargs
+            Forwarded verbatim to :py:meth:`update_text` (after injecting the
+            ``selector`` default shown above).
+        """
+
+        # Delegate – subclasses may have overridden *update_text* with a richer
+        # signature so we pass everything through untouched.
+        return self.update_text(*args, selector=selector, **kwargs)  # type: ignore[arg-type]
+
+    # ------------------------------------------------------------------
+    # Generic fallback implementation
+    # ------------------------------------------------------------------
+    def update_text(  # type: ignore[override]
+        self,
+        transform: Callable[[Any], Optional[str]],
+        *,
+        selector: str = "text",
+        apply_exclusions: bool = False,
+        **_,
+    ):
+        """Generic implementation that works for any object exposing *find_all*.
+
+        Classes that require more sophisticated behaviour (parallelism, page
+        delegation, etc.) are expected to *override* this method while keeping
+        the same public contract.
+        """
+
+        if not callable(transform):
+            raise TypeError("transform must be callable")
+
+        # We rely on the presence of *find_all* to obtain elements.  If the
+        # subclass does not implement it then it *must* override update_text.
+        if not hasattr(self, "find_all"):
+            raise NotImplementedError(
+                f"{self.__class__.__name__} must implement `update_text` explicitly "
+                "(no `find_all` method found)."
+            )
+
+        try:
+            elements_collection = self.find_all(selector=selector, apply_exclusions=apply_exclusions)
+        except Exception as exc:  # pragma: no cover – defensive
+            raise RuntimeError(f"Failed to gather elements with selector '{selector}': {exc}") from exc
+
+        # `find_all` returns an ElementCollection; fall back gracefully otherwise.
+        elements_iter = getattr(elements_collection, "elements", elements_collection)
+        updated = 0
+
+        for element in elements_iter:
+            if not hasattr(element, "text"):
+                continue
+
+            new_text = transform(element)
+            if new_text is not None and isinstance(new_text, str) and new_text != element.text:
+                element.text = new_text
+                updated += 1
+
+        logger.info(
+            "%s.update_text – processed %d element(s); updated %d.",
+            self.__class__.__name__,
+            len(elements_iter),
+            updated,
+        )
+
+        return self 

{natural_pdf-0.1.37.dist-info → natural_pdf-0.1.40.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: natural-pdf
-Version: 0.1.37
+Version: 0.1.40
 Summary: A more intuitive interface for working with PDFs
 Author-email: Jonathan Soma <jonathan.soma@gmail.com>
 License-Expression: MIT

{natural_pdf-0.1.37.dist-info → natural_pdf-0.1.40.dist-info}/RECORD

@@ -1,5 +1,6 @@
-natural_pdf/__init__.py,sha256=qDFJNF8sbEDO-2WSFAxoWEM8updOUP6dB-ckya0kxfs,3275
+natural_pdf/__init__.py,sha256=ACwgLmWAgifAO7NtatP5c57u60g3j6YfYZd-uSL9Ly4,3455
 natural_pdf/cli.py,sha256=SkPwhhMM-GhLsj3O1n1Agxz4KOxcZ08sj8hVQSFJB5c,4064
+natural_pdf/text_mixin.py,sha256=O-RECzpjcmbf5HF7LSqL5VmyexVNF8GIwj9LyJNqLgQ,3952
 natural_pdf/analyzers/__init__.py,sha256=M5oD4oEsIBWrzgTaeg4uBYrC4OPd7Mp7tz8ootKN_l8,1134
 natural_pdf/analyzers/guides.py,sha256=D55ov45PE7mhqvEnarn82y5hG6gmDzk7tYw233LnluA,141896
 natural_pdf/analyzers/shape_detection_mixin.py,sha256=Ef1o73QYVXQ2QcQMM_W9XRwY6vaIQHgxzD7etJ6LbiM,62820
@@ -27,8 +28,8 @@ natural_pdf/collections/pdf_collection.py,sha256=sDVEbFMNME_2OaHIsCoR_W7V1cAATNw
 natural_pdf/core/__init__.py,sha256=QC8H4M3KbXwMFiQORZ0pdPlzx1Ix6oKKQSS7Ib2KEaA,38
 natural_pdf/core/element_manager.py,sha256=DRZvntd99wjXy6KeDjCq5uRhjMftZop9QklOZqlUH8M,55349
 natural_pdf/core/highlighting_service.py,sha256=2tBrrEq6d6hz5f6Yf7z5TysJdlTyuHTURBnQxokJnDM,40645
-natural_pdf/core/page.py,sha256=MwIENkMjEKStC6RlD3SBrqmyZt_MKzrIY7vLBFIvrwY,142529
-natural_pdf/core/pdf.py,sha256=2hK3yRVRxEQMVy1v4w6P26VGoDpCu_3FNkYgN-LO4hA,93221
+natural_pdf/core/page.py,sha256=6QaeOZqh57dWNs9-JlJl-lQWkV1cQ7QqnUxUHfpvKXE,145377
+natural_pdf/core/pdf.py,sha256=-ezlEPm-D_8U2UJe_Msb0v2wCr7I2nf88bVA3im8Spo,98945
 natural_pdf/describe/__init__.py,sha256=kIV7ORmWWB1SAur7nK2aAwR-wHqSedhKfUsaUl4hG0A,586
 natural_pdf/describe/base.py,sha256=CLhZXYQO6SOPUVWLt6VwZ7MK48t_6wgPMyFMLtTCKRc,18166
 natural_pdf/describe/elements.py,sha256=JicXC9SJmmasqxalpCXA47-kVwv-6JnR3Xiu778aNHM,12634
@@ -36,11 +37,11 @@ natural_pdf/describe/mixin.py,sha256=rkX14aGrSz7Jvxx8Rbxv3eSfbO-_29DipwpstrV2pDQ
 natural_pdf/describe/summary.py,sha256=cfT4ZQkeatCDAOwWPwhtEVXisNgk6E57fAXAnoRysSU,7645
 natural_pdf/elements/__init__.py,sha256=S8XeiNWJ1WcgnyYKdYV1yxQlAxCCO3FfITT8MQwNbyk,41
 natural_pdf/elements/base.py,sha256=-ZAcc8lb2aSWTKcprwKTvnR6hsDGDm7T8a1Y9V38E_A,52042
-natural_pdf/elements/collections.py,sha256=_B03lJA1n147alE4xvn6qQ9uZWI8kb8VGxpchghqxqg,131834
+natural_pdf/elements/collections.py,sha256=ko9tzZanbHsDtn0RP6p1Ah1kWH35BBxLOyrRjTokYLU,138948
 natural_pdf/elements/image.py,sha256=zu-P2Y8fRoEXf6IeZU0EYRWsgZ6I_a5vy1FA3VXTGkQ,1424
 natural_pdf/elements/line.py,sha256=mHSeV-ZABY-Cc_K_NpFL53OGtTWlexYDlMvZc8_Vrx8,3845
 natural_pdf/elements/rect.py,sha256=QuQg0Qo7XYQKBac-3Ss0n0ELV6icdPcrygWM2VWzeX8,3325
-natural_pdf/elements/region.py,sha256=ewY9HmV_VN6tN_VKtHj7dtk6nh7hrot-pW5Soz5iMg0,148150
+natural_pdf/elements/region.py,sha256=wpCqK7zLXiwKgH2SQ1-VKjQgcNQsc3Nhmy-bRmJ6Ds4,153201
 natural_pdf/elements/text.py,sha256=giPJQaXuOBCviQ7QKVx_ZMrKFVpgQAsaCS2-kn-8mp0,20530
 natural_pdf/export/mixin.py,sha256=L1q3MIEFWuvie4j4_EmW7GT3NerbZ1as0XMUoqTS7gM,5083
 natural_pdf/exporters/__init__.py,sha256=QffoARekR6WzXEd05oxOytly4qPdBizuIF-SUkeFpig,643
@@ -58,9 +59,9 @@ natural_pdf/extraction/mixin.py,sha256=z0HNRs4x4RoioNjzg3slDeqoHbiPug0HB37bUHehq
 natural_pdf/extraction/result.py,sha256=PDaCCN2LQBbHsZy0_lrQ0ROeMsnmH1WRoXWOjk9M2o4,1825
 natural_pdf/flows/__init__.py,sha256=cUN4A8hTDLZSRr4PO2W_lR4z6hWpbNG8Seox-IIcrLU,277
 natural_pdf/flows/collections.py,sha256=iF8SsfKKb-YVIGi3m-yMRnfKgo_0n_EGhojnYK24h-Q,28493
-natural_pdf/flows/element.py,sha256=mKzk3B7A7sWNvu4CDvAjLr3_ZFLt--ktrSNoLfLpFxU,23940
-natural_pdf/flows/flow.py,sha256=ukkUqXsZmEw-QJEiVqEBLC8ktfBG2Bw56_RR1OEsd24,12802
-natural_pdf/flows/region.py,sha256=jRenBFh2ZmFNklNnGkzCsAM0OfMjbP8fo0p7BiVCl_k,31795
+natural_pdf/flows/element.py,sha256=Tlxjhe3-4i3V6ondGTPWHHE3B3z7A3_CSamLe4X45SE,24932
+natural_pdf/flows/flow.py,sha256=h_tNE0bUQPIC554rprLYoJ5ct1hYPz82FYViNZIGmMw,44068
+natural_pdf/flows/region.py,sha256=QwJ5E0LplIzPf9TbdwivQZqouyVlozITzBkC2SI9gDM,34201
 natural_pdf/ocr/__init__.py,sha256=VY8hhvDPf7Gh2lB-d2QRmghLLyTy6ydxlgo1cS4dOSk,2482
 natural_pdf/ocr/engine.py,sha256=SwNlWydtHbrIghV5JD_j5B4-rnjCMYIWUIEARag-zHw,11839
 natural_pdf/ocr/engine_doctr.py,sha256=ptKrupMWoulZb-R93zr9btoe94JPWU7vlJuN7OBJEIM,17740
@@ -81,7 +82,7 @@ natural_pdf/search/search_options.py,sha256=sq_e8_jSROicD94b_xtDtLnjEr_Zsy4icjzP
 natural_pdf/search/search_service_protocol.py,sha256=Dl-Q-CrutkhZwI69scbW9EWPeYM63qxB60_EA7YqIYo,6699
 natural_pdf/search/searchable_mixin.py,sha256=hqQ_AuID5eTGRCtKYdFLZ1zF35y73uk3x1M1VW9Il8U,23514
 natural_pdf/selectors/__init__.py,sha256=oZGeqSv53EqmIZOhcnawuaGGlRg1h79vArXuZCWKm4A,123
-natural_pdf/selectors/parser.py,sha256=W1gZuBhGy2uHqCoExzCAFbsiVMKYSgUfCc9cr4rO1V0,37540
+natural_pdf/selectors/parser.py,sha256=oa5tzX8UnTwxKbhVrH46lzdILfDObWWRWlQgy2tuwFQ,38828
 natural_pdf/tables/__init__.py,sha256=sCvCGbGsL6BiqlNxAYfVv003bIDLI11FmjHhaWfcU6w,104
 natural_pdf/tables/result.py,sha256=hrGIWDkImpdxsGzugcQKU-qrTgHwwfOigJDFdYl8aUc,3994
 natural_pdf/templates/__init__.py,sha256=jYBxzfi73vew0f6yhIh1MlRxw4F_TVN2hKQR0YXOFe0,20
@@ -98,7 +99,7 @@ natural_pdf/utils/text_extraction.py,sha256=HYWlYGPfafwzsuMyfL5oQhvcD4NobbvC_aCp
 natural_pdf/utils/visualization.py,sha256=olDkWtuVzP0NxRg0CP0DL-eXNCY7Bs-SH-2Xn-cjbo0,9370
 natural_pdf/widgets/__init__.py,sha256=QTVaUmsw__FCweFYZebwPssQxxUFUMd0wpm_cUbGZJY,181
 natural_pdf/widgets/viewer.py,sha256=KW3JogdR2TMg2ECUMYp8hwd060hfg8EsYBWxb5IEzBY,24942
-natural_pdf-0.1.37.dist-info/licenses/LICENSE,sha256=9zfwINwJlarbDmdh6iJV4QUG54QSJlSAUcnC1YiC_Ns,1074
+natural_pdf-0.1.40.dist-info/licenses/LICENSE,sha256=9zfwINwJlarbDmdh6iJV4QUG54QSJlSAUcnC1YiC_Ns,1074
 optimization/memory_comparison.py,sha256=F90D_5WhliSGAct_lyx93xd4q4F-jeo8QpGyDr8tmNw,6543
 optimization/pdf_analyzer.py,sha256=xf6h-FNlqCpsm8NriXcs_bQZOB8eQkxgGGKVRL_jgCM,19347
 optimization/performance_analysis.py,sha256=RjAqeE3YS1r_7qTWkY6Ng5YMbb6MXJXfXX6LoVjg_xQ,13035
@@ -115,8 +116,8 @@ tools/bad_pdf_eval/llm_enrich.py,sha256=mCh4KGi1HmIkzGjj5rrHz1Osd7sEX1IZ_FW08H1t
 tools/bad_pdf_eval/llm_enrich_with_retry.py,sha256=XUtPF1hUvqd3frDXT0wDTXoonuAivhjM5vgFdZ-tm0A,9373
 tools/bad_pdf_eval/reporter.py,sha256=e1g__mkSB4q02p3mGWOwMhvFs7F2HJosNBxup0-LkyU,400
 tools/bad_pdf_eval/utils.py,sha256=hR95XQ7qf7Cu6BdyX0L7ggGVx-ah5sK0jHWblTJUUic,4896
-natural_pdf-0.1.37.dist-info/METADATA,sha256=1POawL7Edgjod2Qt1TO-2DhUkVesip-OnB0KkQCgGQ0,6739
-natural_pdf-0.1.37.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-natural_pdf-0.1.37.dist-info/entry_points.txt,sha256=1R_KMv7g60UBBpRqGfw7bppsMNGdayR-iJlb9ohEk_8,81
-natural_pdf-0.1.37.dist-info/top_level.txt,sha256=oZlRzSc3nZ9sV3L6kD_Di734Pp62ANrm46imFVa51qQ,58
-natural_pdf-0.1.37.dist-info/RECORD,,
+natural_pdf-0.1.40.dist-info/METADATA,sha256=igAI5dRWEeocwR0_BpNgtA8Rw5dOVuoGZ0kJwh08q3Y,6739
+natural_pdf-0.1.40.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+natural_pdf-0.1.40.dist-info/entry_points.txt,sha256=1R_KMv7g60UBBpRqGfw7bppsMNGdayR-iJlb9ohEk_8,81
+natural_pdf-0.1.40.dist-info/top_level.txt,sha256=oZlRzSc3nZ9sV3L6kD_Di734Pp62ANrm46imFVa51qQ,58
+natural_pdf-0.1.40.dist-info/RECORD,,