natural-pdf 0.2.17__py3-none-any.whl → 0.2.19__py3-none-any.whl

natural_pdf/selectors/parser.py

@@ -7,11 +7,12 @@ selectors with extensions for PDF-specific attributes and spatial relationships.
 The parser handles:
 - Basic element selectors (text, rect, line, image)
 - Attribute selectors with comparisons ([size>12], [color="red"])
-- Pseudo-selectors for text content (:contains(), :regex())
+- Pseudo-selectors for text content (:contains(), :regex(), :closest())
 - Spatial relationship selectors (:above(), :below(), :near())
 - Color matching with Delta E distance calculations
 - Logical operators (AND, OR) and grouping
 - Complex nested expressions with proper precedence
+- Fuzzy text matching for OCR errors (:closest())
 
 Key features:
 - Safe value parsing without eval() for security
@@ -25,9 +26,12 @@ This enables powerful document navigation like:
 - page.find_all('rect[color~="red"]:above(text:contains("Total"))')
 - page.find('text:regex("[0-9]{4}-[0-9]{2}-[0-9]{2}")')
 - page.find('text:regex("[\u2500-\u257f]")')  # Box drawing characters
+- page.find('text:closest("Date(s) of Review")')  # Fuzzy match for OCR errors
+- page.find('text:closest("Invoice Date@0.9")')   # 90% similarity threshold
 """
 
 import ast
+import difflib
 import logging
 import re
 from collections import Counter
@@ -691,6 +695,9 @@ def _build_filter_list(
                         return getattr(element, "region_type", "").lower().replace(" ", "_")
                 elif name == "model":
                     return getattr(element, "model", None)
+                elif name == "checked":
+                    # Map 'checked' attribute to is_checked for checkboxes
+                    return getattr(element, "is_checked", None)
                 else:
                     return getattr(element, python_name, None)
             else:
@@ -724,6 +731,29 @@ def _build_filter_list(
                 ]:
                     op_desc = f"= {value!r} (exact color)"
                     compare_func = lambda el_val, sel_val: _is_exact_color_match(el_val, sel_val)
+                # For boolean attributes, handle string/bool comparison
+                elif name in ["checked", "is_checked", "bold", "italic"]:
+
+                    def bool_compare(el_val, sel_val):
+                        # Convert both to boolean for comparison
+                        if isinstance(el_val, bool):
+                            el_bool = el_val
+                        else:
+                            el_bool = str(el_val).lower() in ("true", "1", "yes")
+
+                        if isinstance(sel_val, bool):
+                            sel_bool = sel_val
+                        else:
+                            sel_bool = str(sel_val).lower() in ("true", "1", "yes")
+
+                        # Debug logging
+                        logger.debug(
+                            f"Boolean comparison: el_val={el_val} ({type(el_val)}) -> {el_bool}, sel_val={sel_val} ({type(sel_val)}) -> {sel_bool}"
+                        )
+
+                        return el_bool == sel_bool
+
+                    compare_func = bool_compare
                 else:
                     compare_func = lambda el_val, sel_val: el_val == sel_val
             elif op == "!=":
@@ -894,6 +924,13 @@ def _build_filter_list(
 
             filter_lambda = regex_check
 
+        # --- Handle :closest pseudo-class for fuzzy text matching --- #
+        elif name == "closest" and args is not None:
+            # Note: :closest is handled specially in the page._apply_selector method
+            # It doesn't filter elements here, but marks them for special processing
+            # This allows us to first check :contains matches, then sort by similarity
+            filter_lambda = lambda el: True  # Accept all elements for now
+
         # --- Handle :startswith and :starts-with (alias) --- #
         elif name in ("starts-with", "startswith") and args is not None:
             filter_name = f"pseudo-class :{name}({args!r})"
@@ -936,6 +973,10 @@ def _build_filter_list(
             filter_lambda = lambda el: hasattr(el, "is_horizontal") and el.is_horizontal
         elif name == "vertical":
             filter_lambda = lambda el: hasattr(el, "is_vertical") and el.is_vertical
+        elif name == "checked":
+            filter_lambda = lambda el: hasattr(el, "is_checked") and el.is_checked
+        elif name == "unchecked":
+            filter_lambda = lambda el: hasattr(el, "is_checked") and not el.is_checked
 
         # --- New: :strike / :strikethrough / :strikeout pseudo-classes --- #
         elif name in ("strike", "strikethrough", "strikeout"):

natural_pdf/utils/spatial.py

@@ -10,11 +10,11 @@ with include_boundaries='none'.
 
 Example:
     from natural_pdf.utils.spatial import is_element_in_region
-    
+
     # Check if element is in region using center-based logic (default)
     if is_element_in_region(element, region):
         print("Element is in region")
-    
+
     # Use different strategies
     if is_element_in_region(element, region, strategy="intersects"):
         print("Element overlaps with region")
@@ -35,16 +35,16 @@ InclusionStrategy = Literal["center", "intersects", "contains"]
 
 def is_element_in_region(
     element: "Element",
-    region: "Region", 
+    region: "Region",
     strategy: InclusionStrategy = "center",
-    check_page: bool = True
+    check_page: bool = True,
 ) -> bool:
     """
     Unified function to check if an element is inside a region.
-    
+
     This centralizes the logic used across Region, Page, and Flow to ensure
     consistent behavior throughout the library.
-    
+
     Args:
         element: The element to check
         region: The region to check against
@@ -53,7 +53,7 @@ def is_element_in_region(
             - "intersects": Element belongs if any part overlaps
             - "contains": Element belongs only if fully contained
         check_page: Whether to verify element and region are on the same page
-        
+
     Returns:
         bool: True if element is in region according to the strategy
     """
@@ -61,18 +61,18 @@ def is_element_in_region(
     if not hasattr(element, "bbox") or not element.bbox:
         logger.debug(f"Element lacks bbox attributes: {element}")
         return False
-        
+
     if not hasattr(region, "bbox") or not region.bbox:
         logger.debug(f"Region lacks bbox attributes: {region}")
         return False
-    
+
     # Check page membership if requested
     if check_page:
         if not hasattr(element, "page") or not hasattr(region, "page"):
             return False
         if element.page != region.page:
             return False
-    
+
     # Apply the appropriate strategy
     if strategy == "center":
         # Use existing region method if available
@@ -82,37 +82,43 @@ def is_element_in_region(
             # Fallback calculation
             elem_center_x = (element.x0 + element.x1) / 2
             elem_center_y = (element.top + element.bottom) / 2
-            
+
             # Use region's is_point_inside if available
             if hasattr(region, "is_point_inside"):
                 return region.is_point_inside(elem_center_x, elem_center_y)
             else:
                 # Simple bounds check
-                return (region.x0 <= elem_center_x <= region.x1 and 
-                       region.top <= elem_center_y <= region.bottom)
-    
+                return (
+                    region.x0 <= elem_center_x <= region.x1
+                    and region.top <= elem_center_y <= region.bottom
+                )
+
     elif strategy == "intersects":
         # Use existing region method if available
         if hasattr(region, "intersects"):
             return region.intersects(element)
         else:
             # Simple bbox overlap check
-            return not (element.x1 < region.x0 or 
-                       element.x0 > region.x1 or 
-                       element.bottom < region.top or 
-                       element.top > region.bottom)
-    
+            return not (
+                element.x1 < region.x0
+                or element.x0 > region.x1
+                or element.bottom < region.top
+                or element.top > region.bottom
+            )
+
     elif strategy == "contains":
         # Use existing region method if available
         if hasattr(region, "contains"):
             return region.contains(element)
         else:
             # Simple full containment check
-            return (region.x0 <= element.x0 and 
-                   element.x1 <= region.x1 and 
-                   region.top <= element.top and 
-                   element.bottom <= region.bottom)
-    
+            return (
+                region.x0 <= element.x0
+                and element.x1 <= region.x1
+                and region.top <= element.top
+                and element.bottom <= region.bottom
+            )
+
     else:
         raise ValueError(f"Unknown inclusion strategy: {strategy}")
 
@@ -120,10 +126,10 @@ def is_element_in_region(
 def get_inclusion_strategy() -> InclusionStrategy:
     """
     Get the current global inclusion strategy.
-    
+
     This could be made configurable via environment variable or settings.
     For now, returns the default strategy.
-    
+
     Returns:
         The current inclusion strategy (default: "center")
     """
@@ -132,38 +138,35 @@ def get_inclusion_strategy() -> InclusionStrategy:
     return "center"
 
 
-def calculate_element_overlap_percentage(
-    element: "Element",
-    region: "Region"
-) -> float:
+def calculate_element_overlap_percentage(element: "Element", region: "Region") -> float:
     """
     Calculate what percentage of an element overlaps with a region.
-    
+
     Args:
         element: The element to check
         region: The region to check against
-        
+
     Returns:
         float: Percentage of element area that overlaps with region (0.0 to 1.0)
     """
     if not hasattr(element, "bbox") or not hasattr(region, "bbox"):
         return 0.0
-    
+
     # Calculate intersection bounds
     intersect_x0 = max(element.x0, region.x0)
-    intersect_y0 = max(element.top, region.top) 
+    intersect_y0 = max(element.top, region.top)
     intersect_x1 = min(element.x1, region.x1)
     intersect_y1 = min(element.bottom, region.bottom)
-    
+
     # Check if there's an intersection
     if intersect_x1 <= intersect_x0 or intersect_y1 <= intersect_y0:
         return 0.0
-    
+
     # Calculate areas
     element_area = (element.x1 - element.x0) * (element.bottom - element.top)
     if element_area == 0:
         return 0.0
-        
+
     intersect_area = (intersect_x1 - intersect_x0) * (intersect_y1 - intersect_y0)
-    
-    return intersect_area / element_area
+
+    return intersect_area / element_area

{natural_pdf-0.2.17.dist-info → natural_pdf-0.2.19.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: natural-pdf
-Version: 0.2.17
+Version: 0.2.19
 Summary: A more intuitive interface for working with PDFs
 Author-email: Jonathan Soma <jonathan.soma@gmail.com>
 License-Expression: MIT

{natural_pdf-0.2.17.dist-info → natural_pdf-0.2.19.dist-info}/RECORD

@@ -1,5 +1,6 @@
-natural_pdf/__init__.py,sha256=N9ubwsFpmPj7WHA6Uewgn6IbmU2r0BeUGIdIhmTl6nw,4701
+natural_pdf/__init__.py,sha256=JPuQBMN0mZPnPB4z-RAHm8jPSVLKbgw4gxfSXyEgdX4,4957
 natural_pdf/cli.py,sha256=0zO9ZoRiP8JmyGBaVavrMATnvbARWTl7WD2PEefu9BM,4061
+natural_pdf/judge.py,sha256=mRPJfdIkkL_Y6uQXnb3Wtrna04XlhPrDvxPrDiVevH4,58838
 natural_pdf/text_mixin.py,sha256=eFCiHj6Okcw3aum4955BepcI2NPRalkf9UFFVTc_H30,4012
 natural_pdf/analyzers/__init__.py,sha256=3XGoNq3OgiVkZP7tOdeP5XVUl7fDgyztdA8DlOcMLXg,1138
 natural_pdf/analyzers/guides.py,sha256=BqFgt-bRSOkEoFCvNsYyY8j__00X-8DJ_TLb2Hx9qsQ,202430
@@ -7,6 +8,13 @@ natural_pdf/analyzers/shape_detection_mixin.py,sha256=mgpyJ4jIulz9l9HCqThabJIsLS
 natural_pdf/analyzers/text_options.py,sha256=qEkDaYWla0rIM_gszEOsu52q7C_dAfV81P2HLJZM2sw,3333
 natural_pdf/analyzers/text_structure.py,sha256=3WWusi-BI0krUnJxB05DD6XmKj5qRNvQBqH7zOQGm1M,28451
 natural_pdf/analyzers/utils.py,sha256=PYbzJzSAHZ7JsMes84WIrSbA0zkjJGs0CLvIeINsf_k,2100
+natural_pdf/analyzers/checkbox/__init__.py,sha256=2ZWAIUoRqgGlwVeEU0JNMkQ-mO4nxWNFQ6fLOx0jfRQ,243
+natural_pdf/analyzers/checkbox/base.py,sha256=Hu2WrlaG2gNbTFa8fYSzjyUFmOZvbdTdonnMd9lwl44,9610
+natural_pdf/analyzers/checkbox/checkbox_analyzer.py,sha256=rDO7YIT_fAd5BmpXMOUnZaSHUmFuXqVXZK-HNyS3Ezw,13647
+natural_pdf/analyzers/checkbox/checkbox_manager.py,sha256=ZR8yfhWiykxBe6h4smsDuY-So47j0tcGEXhF0FEIorE,5959
+natural_pdf/analyzers/checkbox/checkbox_options.py,sha256=-2V3_yduBhD4iVjn-EhgK7D6qA2xH9NJorfgDcar6PU,2094
+natural_pdf/analyzers/checkbox/mixin.py,sha256=KYnr_Xx4U2bp6c35GG2hk6yX_z4NgX7ZW9zT1xmEKEw,3710
+natural_pdf/analyzers/checkbox/rtdetr.py,sha256=Oxz4XVJKDuVWzBQDqM_hqslCH66n1HJg4_hdXS4aAs4,6944
 natural_pdf/analyzers/layout/__init__.py,sha256=oq1uJ5UkGGMbBKGirV1aRKK3hxAUyjTLywYkPCQH1f0,33
 natural_pdf/analyzers/layout/base.py,sha256=F5xPOJcI65N4nxwm0szvhtbDD6lVMqWDut8PSkTCobU,8349
 natural_pdf/analyzers/layout/docling.py,sha256=4BJYyNVR6VegZGxyisvNIBBRvVk6YKPyDVs7ZdVfzEU,12676
@@ -23,28 +31,28 @@ natural_pdf/analyzers/layout/yolo.py,sha256=2Iz2-WsMy--ftkZQ8j5PGqp_1fTD7Mskl2kN
 natural_pdf/classification/manager.py,sha256=BaqBL9GeMvYgoJsiQeI2J8aUKQ5Qxu_ELRvmCWquld8,22172
 natural_pdf/classification/mixin.py,sha256=CXygXXhe_qx1563SmIjiu4uSnZkxCkuRR4fGvLokS2w,9416
 natural_pdf/classification/results.py,sha256=5ha77CxK0GYwkBMJbvUBZkBjsL5GpOveIZDK9nO4j8I,3239
-natural_pdf/collections/mixins.py,sha256=Se2C5AcpP9B5E0d0pIrey6-f_P32tAXTK4M7666MNj0,5688
+natural_pdf/collections/mixins.py,sha256=ZsS61WFu6Ipree4O_zFECKWoKHC3pYVwZU7tUP6OTOQ,6145
 natural_pdf/core/__init__.py,sha256=QC8H4M3KbXwMFiQORZ0pdPlzx1Ix6oKKQSS7Ib2KEaA,38
-natural_pdf/core/element_manager.py,sha256=619R97OtMd7uhaax7fZNJmhy9GxSs9HCNP4OzGgP828,55882
+natural_pdf/core/element_manager.py,sha256=7fy65zzD42LvDJKj8X1pbJAQYL5lk9wGdTtgE0rsPpA,56057
 natural_pdf/core/highlighting_service.py,sha256=wEV-koqHoHf7S3wZ3j8D2L-ucGp3Nd0YhhStz9yqeLc,70406
-natural_pdf/core/page.py,sha256=-0OaIoXz0zjT_jnPjjI2jpb8vvNKh-1W56auA5UBhTA,158791
-natural_pdf/core/page_collection.py,sha256=bLZ3TqTQbmP3oYrbfEi7HUoPMbcGplEtUMZ3Z1y7fuw,66728
+natural_pdf/core/page.py,sha256=NiJxBHLx4Otwr7iMza1gsEAfSqTMvTu_6zex4aocZOw,162710
+natural_pdf/core/page_collection.py,sha256=OjIS9iEtFrHw0liJHGI-CFwZbHHA4Lt7vK69wN76Igg,68255
 natural_pdf/core/page_groupby.py,sha256=V2e_RNlHaasUzYm2h2vNJI7_aV_fl3_pg7kU3F2j0z8,8218
-natural_pdf/core/pdf.py,sha256=i8dYCimL_k5FV6BmPI1a2Dk7XZfwLP8TziXr2n3O_fI,105639
+natural_pdf/core/pdf.py,sha256=Cc4A6b49apGfxk7DFcN4oCfoiYmpnH2-jFf_Gb6B5mg,106345
 natural_pdf/core/pdf_collection.py,sha256=s3ogu4CEHrHMTRqQMJUKJZ-9Ii8b_B9dWbVLTFj0s7g,34992
 natural_pdf/core/render_spec.py,sha256=y9QkMiIvWaKiEBlV0TjyldADIEUY3YfWLQXxStHu1S4,15480
 natural_pdf/describe/__init__.py,sha256=kIV7ORmWWB1SAur7nK2aAwR-wHqSedhKfUsaUl4hG0A,586
-natural_pdf/describe/base.py,sha256=M4TGXR8ppTvznTnA1ZDgMQMkDpgu1pwGMNaOcgHf2iY,20154
+natural_pdf/describe/base.py,sha256=pU_fDkWG_hQlne2nNIdOC1xXyTrPc-kmTwd685nZiSk,21024
 natural_pdf/describe/elements.py,sha256=3Y541z5TQ2obrfZFiFi1YQMsCt3oYrhMHpD5j1tuppw,12639
 natural_pdf/describe/mixin.py,sha256=rkX14aGrSz7Jvxx8Rbxv3eSfbO-_29DipwpstrV2pDQ,3109
 natural_pdf/describe/summary.py,sha256=cfT4ZQkeatCDAOwWPwhtEVXisNgk6E57fAXAnoRysSU,7645
 natural_pdf/elements/__init__.py,sha256=ICNikmLeIEuSYypz-KnkBn8xR1hR7rge4hsa1KLkyWY,42
-natural_pdf/elements/base.py,sha256=YYdoss63yv3IzQeuHbNypo7VLz2UJDFK5b6lqQe5tR8,76090
-natural_pdf/elements/element_collection.py,sha256=dlKoIaqmK_pC_cEcTX9LA2bNbZmc8iXcTTDfpHDlyUM,139812
+natural_pdf/elements/base.py,sha256=NunXdrZW53iG-Q4Pe9DHmWpzigHg-JrkjOLZ016I_b0,82679
+natural_pdf/elements/element_collection.py,sha256=z3gRONShw6MrdTJYXVjBi9uNr3dNQtRXgyYKm-VPB7A,141371
 natural_pdf/elements/image.py,sha256=zu-P2Y8fRoEXf6IeZU0EYRWsgZ6I_a5vy1FA3VXTGkQ,1424
 natural_pdf/elements/line.py,sha256=TFn7KXjPT_jUQyQyabU0F7XYU4dC-qadwodJMZF4DCU,3844
 natural_pdf/elements/rect.py,sha256=kmUmhwnihd-aTweAO-LsngRDo5Iqmx7lcSa8ZBlE_2E,4544
-natural_pdf/elements/region.py,sha256=qJ86iToSjrCUjVrEbO0M0S1nTuZDW9tpI4jF9T5xJKs,168777
+natural_pdf/elements/region.py,sha256=ql_pZvfjbT0j2zekqMrGBWDzNVo4erNiQ9aK67J7KTw,173382
 natural_pdf/elements/text.py,sha256=Jo4gnrsJe1PStdoWF2Bt8RSeSmOcfA9DxvMJl7EoAmI,21344
 natural_pdf/export/mixin.py,sha256=L1q3MIEFWuvie4j4_EmW7GT3NerbZ1as0XMUoqTS7gM,5083
 natural_pdf/exporters/__init__.py,sha256=QffoARekR6WzXEd05oxOytly4qPdBizuIF-SUkeFpig,643
@@ -85,7 +93,7 @@ natural_pdf/search/search_options.py,sha256=sq_e8_jSROicD94b_xtDtLnjEr_Zsy4icjzP
 natural_pdf/search/search_service_protocol.py,sha256=u8pbuWP96fnQEe6mnreY9DrdiDAHP6ZCY7phvSbFlP8,6697
 natural_pdf/search/searchable_mixin.py,sha256=hqQ_AuID5eTGRCtKYdFLZ1zF35y73uk3x1M1VW9Il8U,23514
 natural_pdf/selectors/__init__.py,sha256=oZGeqSv53EqmIZOhcnawuaGGlRg1h79vArXuZCWKm4A,123
-natural_pdf/selectors/parser.py,sha256=HbPgmtXXA4lRSAVkCzw6vpCi3oh66e-53yUEPhYLGX8,46909
+natural_pdf/selectors/parser.py,sha256=wXlTL2t05xj47sMoG-vhjQFyEVou8NZie7wKKm60iMA,49063
 natural_pdf/tables/__init__.py,sha256=sCvCGbGsL6BiqlNxAYfVv003bIDLI11FmjHhaWfcU6w,104
 natural_pdf/tables/result.py,sha256=-8ctA-jCJYSHtlfAoqTvhUwO5zSP2BQxxetAjqEsNyg,8665
 natural_pdf/templates/__init__.py,sha256=jYBxzfi73vew0f6yhIh1MlRxw4F_TVN2hKQR0YXOFe0,20
@@ -101,7 +109,7 @@ natural_pdf/utils/packaging.py,sha256=TM0jafwS5yVbTGC-RMi4TyWunf9cUUo9h5J6rMzkT-
 natural_pdf/utils/pdfminer_patches.py,sha256=Ob81OMoNUGMUIy9nMw3deSQ_Z6cQmhbRlHUC3EHw2jk,4201
 natural_pdf/utils/reading_order.py,sha256=u7XyVZdKMPMK0CL1C7xFogKnZ92b0JKT068KFjQWe18,7437
 natural_pdf/utils/sections.py,sha256=HZX7829-fquKgIF7vUN2tL10-aXckEaM25g_2VcgWU4,12941
-natural_pdf/utils/spatial.py,sha256=JOH2LHnF5WBDcjNQsHQdj458zwUgKtSWW7Tj0motn70,5968
+natural_pdf/utils/spatial.py,sha256=YjzGO4A013ZDYGYbs7hl4RbJOaKrg8-x__Dl_BamwUA,5908
 natural_pdf/utils/text_extraction.py,sha256=CCwPTmMoTgtQt2P00X_ADIf6ZGNfxvjCO9FO0_HqG40,13900
 natural_pdf/utils/visualization.py,sha256=zhZEHgYnZFuX7YxTHXF8Y3D97uHp2beTKMaC-JkCFwk,22364
 natural_pdf/vision/__init__.py,sha256=TkoQtdODlh0n_99dsjLIWKE9dgK0m4jfrui_cQ3gTwU,221
@@ -111,15 +119,30 @@ natural_pdf/vision/similarity.py,sha256=HWmXDBNLSOlRWH-_1K3FVR7tSsRuMFqXZwrVhhg2
 natural_pdf/vision/template_matching.py,sha256=91XQt5tp-vmcMX_4b2Bz-YwIAlb-hc8E5ih_qAHQuCk,7145
 natural_pdf/widgets/__init__.py,sha256=QTVaUmsw__FCweFYZebwPssQxxUFUMd0wpm_cUbGZJY,181
 natural_pdf/widgets/viewer.py,sha256=KW3JogdR2TMg2ECUMYp8hwd060hfg8EsYBWxb5IEzBY,24942
-natural_pdf-0.2.17.dist-info/licenses/LICENSE,sha256=9zfwINwJlarbDmdh6iJV4QUG54QSJlSAUcnC1YiC_Ns,1074
+natural_pdf-0.2.19.dist-info/licenses/LICENSE,sha256=9zfwINwJlarbDmdh6iJV4QUG54QSJlSAUcnC1YiC_Ns,1074
 optimization/memory_comparison.py,sha256=0i_foFSRmppj-fY069qjwH36s_zkx-1L2ASAAlepWzA,6541
 optimization/pdf_analyzer.py,sha256=HjrmTgu2qchxPeDckc5kjgxppGwd40UESrYS9Myj7pY,19352
 optimization/performance_analysis.py,sha256=JBXnR9hc7Ix7YCnt3EJPSpsyqIUgKsc7GEffQ_TDCBk,13033
 optimization/test_cleanup_methods.py,sha256=PmLOL4MRgvV0j_DW9W1TS8MsGGgu57QCuq6_5y7zK3s,6209
 optimization/test_memory_fix.py,sha256=A3knK74fNhvHknDbLhbTmA276x1ifl-3ivJ_7BhVSTI,6170
+temp/check_model.py,sha256=rhnqTRUaq2VyyqXHuLBxM7ZEoJwf0ExlSJnvkMDYPRU,1710
+temp/check_pdf_content.py,sha256=adFIVMI6m36l0R3112ESt9oqX_zM-mhDvTusBcjqBy8,233
+temp/checkbox_checks.py,sha256=XsR6bmaVNiSH-HsDzthtJcz4vcKOYJ5IbAi6vtfo7P0,20293
+temp/checkbox_simple.py,sha256=d1NiE1IbGSG2nMtvFPgBgxF6OSZLm7TIC2nkrDSG8fE,3975
+temp/checkbox_ux_ideas.py,sha256=Pa1NXi-wmtEGAPb1RW9fiQ4mcKf1G88OMm7zIABqGoI,15302
+temp/context_manager_prototype.py,sha256=uMRO7xrWsbxBUCUaY7xGtEFcIj-QT9j2DQ2JMkinW2M,6150
+temp/convert_to_hf.py,sha256=DMqZAWvOA_StujfSmkD-hJCnCy4dlvyjIOl2_1l_mOg,1881
+temp/demo_text_closest.py,sha256=qRnAynLhF-P_q9t_WFaxE_5QLbZMiMp4v9llFipfqZA,2721
 temp/fix_page_exclusions.py,sha256=YIj62zF38TdoBARAuSIvEbetl_JfXG-mp4v9p355qmo,1358
+temp/inspect_model.py,sha256=AaRDqhRH9kqcUcfmrSUsNw1xWkxargNA3BWIvzxwHGM,1692
+temp/rtdetr_dinov2_test.py,sha256=9FUL3hiHweYJIbEeH0AZTrLJSnWatxwymNG9CZEXrGA,1553
+temp/test_closest_debug.py,sha256=QP53iAEwy2KRSZlwH2eQ07JILxRgfYwBrvro9i2ITXQ,809
+temp/test_closest_debug2.py,sha256=Hbh0nkG7xS0NfayH2Qg_IzLkeKh6mH-OWo0o2i9777I,740
+temp/test_context_exploration.py,sha256=DlFXDuKavvUskLjHMwqPVGGrPpYT8zBHErX1uzHnWxw,2611
 temp/test_draw_guides.py,sha256=_eSSBElGHQkd2QD_KA_Okw70v0dlY5m-1-C5SQwKAJw,642
 temp/test_draw_guides_interactive.py,sha256=FsH-2ZQGsGx_8QfVCWUAkLbOcJz-VfiwROzQD4AD7kQ,926
+temp/test_durham.py,sha256=A0J78TiVXCLHP4xy67G6GlOtrE2sgWP7FsLMH6fjBaA,916
+temp/test_empty_string.py,sha256=FovOW7hDwkShVT7nYVH_UMv3IwQjX0pHhxC9WHAfo2U,470
 temp/test_exclusion_with_debug.py,sha256=CScxHvb43KrB5dzXuTOhuzjcBXZBdfYB5ygiKkEW26g,1393
 temp/test_find_exclusions_fix.py,sha256=1l5aEqnElcl3kiykdtmJFlVxQ1xMKGm1UckGYEQg--c,2103
 temp/test_find_exclusions_fix_no_recursion.py,sha256=qZspTBwxunRM93N_-fZ2fR5Lodj0ArQX3h10HlTXhfc,3592
@@ -133,6 +156,7 @@ temp/test_marker_order.py,sha256=TFZkMxRiNoZGVcdDivYnkIDNvwHaiyKUdYoy2rTTIiI,141
 temp/test_original_exclusions_now_work.py,sha256=G6LmaF-P9Qhj0j4lT_4ncfCddllfP6L8F_x2prUBr9w,1904
 temp/test_pdf_exclusions_with_guides.py,sha256=QaMl0frgKC8kCPQ2BUI8kqyvqsIjQPXKV_St1rK3zxg,2754
 temp/test_region_exclusions_detailed.py,sha256=EftdW3JY3JH_LX5QlWKt-4drM-joPggK2fKUZRXVTMA,814
+temp/test_similarity.py,sha256=2Nv8QbSwjaBwMwJsvpZgwOiMIRxPMux5QeZE_rgQ63A,441
 temp/test_stripes_real_pdf.py,sha256=FIvDoJrnuioOMw1A0aTCCfZLeg99lusfe0Fb0MiqnhQ,2618
 temp/test_vertical_stripes.py,sha256=Yf3TJfb_faqAFzlgb7i5u6dDHjF4UMSHIGM99vangRk,1877
 temp/test_widget_functionality.py,sha256=jsEGHYK1dWWa8uEcfGRRj1ReHRMzNoIaMZU4d-o-Djs,2448
@@ -148,8 +172,8 @@ 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-0.2.17.dist-info/METADATA,sha256=8K5PCwh_OuI8vkWRLChHeT-LuEd0sRmigkRm55ZNeDo,6960
-natural_pdf-0.2.17.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-natural_pdf-0.2.17.dist-info/entry_points.txt,sha256=1R_KMv7g60UBBpRqGfw7bppsMNGdayR-iJlb9ohEk_8,81
-natural_pdf-0.2.17.dist-info/top_level.txt,sha256=ZDKhxE_tg508o9BpagsjCGcI8GY4cF_8bg0e0IaLsPI,41
-natural_pdf-0.2.17.dist-info/RECORD,,
+natural_pdf-0.2.19.dist-info/METADATA,sha256=vtMsWwMW9cR2LdQhdDFhDG4WWIkctrT7_3P7klvyJ-8,6960
+natural_pdf-0.2.19.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+natural_pdf-0.2.19.dist-info/entry_points.txt,sha256=1R_KMv7g60UBBpRqGfw7bppsMNGdayR-iJlb9ohEk_8,81
+natural_pdf-0.2.19.dist-info/top_level.txt,sha256=ZDKhxE_tg508o9BpagsjCGcI8GY4cF_8bg0e0IaLsPI,41
+natural_pdf-0.2.19.dist-info/RECORD,,

temp/check_model.py

@@ -0,0 +1,49 @@
+from ultralytics import RTDETR
+import os
+
+model_path = "/Users/soma/Development/natural-pdf/model-weights/checkbox-nano.pt"
+print(f"Model exists: {os.path.exists(model_path)}")
+
+try:
+    model = RTDETR(model_path)
+    print(f"Model loaded successfully")
+    print(f"Model names: {model.names}")
+    print(f"Model task: {model.task}")
+    
+    # Try to get architecture info
+    if hasattr(model.model, 'yaml'):
+        print(f"Model yaml: {model.model.yaml}")
+    
+    # Check the model structure
+    if hasattr(model.model, 'model'):
+        for i, module in enumerate(model.model.model):
+            print(f"Layer {i}: {module}")
+            if i > 5:  # Just show first few layers
+                break
+                
+except Exception as e:
+    print(f"Error: {e}")
+    
+    # Try loading as generic model to inspect
+    import torch
+    try:
+        checkpoint = torch.load(model_path, map_location='cpu', weights_only=False)
+        print(f"\nCheckpoint keys: {list(checkpoint.keys())}")
+        
+        # Check for model configuration
+        if 'model' in checkpoint and hasattr(checkpoint['model'], 'yaml'):
+            print(f"Model yaml: {checkpoint['model'].yaml}")
+        
+        # Check train args for model info
+        if 'train_args' in checkpoint:
+            args = checkpoint['train_args']
+            print(f"\nTraining args:")
+            print(f"  Model: {getattr(args, 'model', 'Unknown')}")
+            print(f"  Task: {getattr(args, 'task', 'Unknown')}")
+            
+        # Check epoch info
+        if 'epoch' in checkpoint:
+            print(f"  Epochs trained: {checkpoint['epoch']}")
+            
+    except Exception as e2:
+        print(f"Error loading checkpoint: {e2}")

temp/check_pdf_content.py

@@ -0,0 +1,9 @@
+from natural_pdf import PDF
+
+pdf = PDF('pdfs/01-practice.pdf')
+page = pdf.pages[0]
+texts = page.find_all('text')
+print(f'Total text elements: {len(texts)}')
+print('Sample texts:')
+for t in texts[:20]:
+    print(f'  - {repr(t.text)}')