PyPI - natural-pdf - Versions diffs - 0.1.4__py3-none-any.whl → 0.1.6__py3-none-any.whl - Mend

natural-pdf 0.1.4py3-none-any.whl → 0.1.6py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (141) hide show

docs/api/index.md +386 -0
docs/assets/favicon.png +3 -0
docs/assets/favicon.svg +3 -0
docs/assets/javascripts/custom.js +17 -0
docs/assets/logo.svg +3 -0
docs/assets/sample-screen.png +0 -0
docs/assets/social-preview.png +17 -0
docs/assets/social-preview.svg +17 -0
docs/assets/stylesheets/custom.css +65 -0
docs/document-qa/index.ipynb +435 -0
docs/document-qa/index.md +79 -0
docs/element-selection/index.ipynb +915 -0
docs/element-selection/index.md +229 -0
docs/index.md +170 -0
docs/installation/index.md +69 -0
docs/interactive-widget/index.ipynb +962 -0
docs/interactive-widget/index.md +12 -0
docs/layout-analysis/index.ipynb +818 -0
docs/layout-analysis/index.md +185 -0
docs/ocr/index.md +209 -0
docs/pdf-navigation/index.ipynb +314 -0
docs/pdf-navigation/index.md +97 -0
docs/regions/index.ipynb +816 -0
docs/regions/index.md +294 -0
docs/tables/index.ipynb +658 -0
docs/tables/index.md +144 -0
docs/text-analysis/index.ipynb +370 -0
docs/text-analysis/index.md +105 -0
docs/text-extraction/index.ipynb +1478 -0
docs/text-extraction/index.md +292 -0
docs/tutorials/01-loading-and-extraction.ipynb +1710 -0
docs/tutorials/01-loading-and-extraction.md +95 -0
docs/tutorials/02-finding-elements.ipynb +340 -0
docs/tutorials/02-finding-elements.md +149 -0
docs/tutorials/03-extracting-blocks.ipynb +147 -0
docs/tutorials/03-extracting-blocks.md +48 -0
docs/tutorials/04-table-extraction.ipynb +114 -0
docs/tutorials/04-table-extraction.md +50 -0
docs/tutorials/05-excluding-content.ipynb +270 -0
docs/tutorials/05-excluding-content.md +109 -0
docs/tutorials/06-document-qa.ipynb +332 -0
docs/tutorials/06-document-qa.md +91 -0
docs/tutorials/07-layout-analysis.ipynb +288 -0
docs/tutorials/07-layout-analysis.md +66 -0
docs/tutorials/07-working-with-regions.ipynb +413 -0
docs/tutorials/07-working-with-regions.md +151 -0
docs/tutorials/08-spatial-navigation.ipynb +508 -0
docs/tutorials/08-spatial-navigation.md +190 -0
docs/tutorials/09-section-extraction.ipynb +2434 -0
docs/tutorials/09-section-extraction.md +256 -0
docs/tutorials/10-form-field-extraction.ipynb +512 -0
docs/tutorials/10-form-field-extraction.md +201 -0
docs/tutorials/11-enhanced-table-processing.ipynb +54 -0
docs/tutorials/11-enhanced-table-processing.md +9 -0
docs/tutorials/12-ocr-integration.ipynb +604 -0
docs/tutorials/12-ocr-integration.md +175 -0
docs/tutorials/13-semantic-search.ipynb +1328 -0
docs/tutorials/13-semantic-search.md +77 -0
docs/visual-debugging/index.ipynb +2970 -0
docs/visual-debugging/index.md +157 -0
docs/visual-debugging/region.png +0 -0
natural_pdf/__init__.py +50 -33
natural_pdf/analyzers/__init__.py +2 -1
natural_pdf/analyzers/layout/base.py +32 -24
natural_pdf/analyzers/layout/docling.py +131 -72
natural_pdf/analyzers/layout/gemini.py +264 -0
natural_pdf/analyzers/layout/layout_analyzer.py +156 -113
natural_pdf/analyzers/layout/layout_manager.py +125 -58
natural_pdf/analyzers/layout/layout_options.py +43 -17
natural_pdf/analyzers/layout/paddle.py +152 -95
natural_pdf/analyzers/layout/surya.py +164 -92
natural_pdf/analyzers/layout/tatr.py +149 -84
natural_pdf/analyzers/layout/yolo.py +89 -45
natural_pdf/analyzers/text_options.py +22 -15
natural_pdf/analyzers/text_structure.py +131 -85
natural_pdf/analyzers/utils.py +30 -23
natural_pdf/collections/pdf_collection.py +146 -97
natural_pdf/core/__init__.py +1 -1
natural_pdf/core/element_manager.py +419 -337
natural_pdf/core/highlighting_service.py +268 -196
natural_pdf/core/page.py +1044 -521
natural_pdf/core/pdf.py +516 -313
natural_pdf/elements/__init__.py +1 -1
natural_pdf/elements/base.py +307 -225
natural_pdf/elements/collections.py +805 -543
natural_pdf/elements/line.py +39 -36
natural_pdf/elements/rect.py +32 -30
natural_pdf/elements/region.py +889 -879
natural_pdf/elements/text.py +127 -99
natural_pdf/exporters/__init__.py +0 -1
natural_pdf/exporters/searchable_pdf.py +261 -102
natural_pdf/ocr/__init__.py +57 -35
natural_pdf/ocr/engine.py +150 -46
natural_pdf/ocr/engine_easyocr.py +146 -150
natural_pdf/ocr/engine_paddle.py +118 -175
natural_pdf/ocr/engine_surya.py +78 -141
natural_pdf/ocr/ocr_factory.py +114 -0
natural_pdf/ocr/ocr_manager.py +122 -124
natural_pdf/ocr/ocr_options.py +16 -20
natural_pdf/ocr/utils.py +98 -0
natural_pdf/qa/__init__.py +1 -1
natural_pdf/qa/document_qa.py +119 -111
natural_pdf/search/__init__.py +37 -31
natural_pdf/search/haystack_search_service.py +312 -189
natural_pdf/search/haystack_utils.py +186 -122
natural_pdf/search/search_options.py +25 -14
natural_pdf/search/search_service_protocol.py +12 -6
natural_pdf/search/searchable_mixin.py +261 -176
natural_pdf/selectors/__init__.py +2 -1
natural_pdf/selectors/parser.py +159 -316
natural_pdf/templates/__init__.py +1 -1
natural_pdf/templates/spa/css/style.css +334 -0
natural_pdf/templates/spa/index.html +31 -0
natural_pdf/templates/spa/js/app.js +472 -0
natural_pdf/templates/spa/words.txt +235976 -0
natural_pdf/utils/debug.py +32 -0
natural_pdf/utils/highlighting.py +8 -2
natural_pdf/utils/identifiers.py +29 -0
natural_pdf/utils/packaging.py +418 -0
natural_pdf/utils/reading_order.py +65 -63
natural_pdf/utils/text_extraction.py +195 -0
natural_pdf/utils/visualization.py +70 -61
natural_pdf/widgets/__init__.py +2 -3
natural_pdf/widgets/viewer.py +749 -718
{natural_pdf-0.1.4.dist-info → natural_pdf-0.1.6.dist-info}/METADATA +53 -17
natural_pdf-0.1.6.dist-info/RECORD +141 -0
{natural_pdf-0.1.4.dist-info → natural_pdf-0.1.6.dist-info}/WHEEL +1 -1
natural_pdf-0.1.6.dist-info/top_level.txt +4 -0
notebooks/Examples.ipynb +1293 -0
pdfs/.gitkeep +0 -0
pdfs/01-practice.pdf +543 -0
pdfs/0500000US42001.pdf +0 -0
pdfs/0500000US42007.pdf +0 -0
pdfs/2014 Statistics.pdf +0 -0
pdfs/2019 Statistics.pdf +0 -0
pdfs/Atlanta_Public_Schools_GA_sample.pdf +0 -0
pdfs/needs-ocr.pdf +0 -0
natural_pdf/templates/ocr_debug.html +0 -517
natural_pdf-0.1.4.dist-info/RECORD +0 -61
natural_pdf-0.1.4.dist-info/top_level.txt +0 -1
{natural_pdf-0.1.4.dist-info → natural_pdf-0.1.6.dist-info}/licenses/LICENSE +0 -0

pdfs/.gitkeep ADDED Viewed

File without changes

pdfs/01-practice.pdf ADDED Viewed

@@ -0,0 +1,543 @@
+%PDF-1.7
+%µ¶
+1 0 obj
+<</Type/Pages/Count 1/Kids[2 0 R]>>
+endobj
+2 0 obj
+<</Type/Page/MediaBox[0 0 612 792]/Resources 3 0 R/Parent 1 0 R/Contents 6 0 R>>
+endobj
+3 0 obj
+<</Font<</F0<</Type/Font/BaseFont/Helvetica/Subtype/Type1>>/F1<</Type/Font/BaseFont/Helvetica-Bold/Subtype/Type1>>>>>>
+endobj
+4 0 obj
+<</Length 4482>>
+stream
+q
+    1 0 0 1 50 700 cm
+    1 g
+    325 65 180 -35 re
+    S
+Q
+q
+    1 0 0 1 50 700 cm
+    BT
+        335 50 Td
+        /F0 8 Tf
+        (Jungle Health and Safety Inspection Service) Tj
+        0 -10 Td
+        1 0 0 rg
+        (INS-UP70N51NCL41R) Tj
+    ET
+Q
+q
+    1 0 0 1 50 700 cm
+    BT
+        /F1 10 Tf
+        (Site: ) Tj
+        /F0 10 Tf
+        (Durham's Meatpacking  ) Tj
+        0.5 g
+        (Chicago, Ill.) Tj
+        0 g
+        0 -20 Td
+        /F1 10 Tf
+        (Date:  ) Tj
+        /F0 10 Tf
+        (February 3, 1905) Tj
+        0 -20 Td
+        /F1 10 Tf
+        (Violation Count: ) Tj
+        /F0 10 Tf
+        (7) Tj
+        0 -20 Td
+        /F1 10 Tf
+        (Summary: ) Tj
+        % 0 -20 Td
+        /F0 10 Tf
+        (Worst of any, however, were the fertilizer men, and those who served in the cooking rooms.) Tj
+        0 -16 Td
+        (These people could not be shown to the visitor - for the odor of a fertilizer man would scare any ordinary ) Tj
+        0 -16 Td
+        (visitor at a hundred yards, and as for the other men, who worked in tank rooms full of steam, and in ) Tj
+        0 -16 Td
+        (some of which there were open vats near the level of the floor, their peculiar trouble was that they fell) Tj
+        0 -16 Td
+        (into the vats; and when they were fished out, there was never enough of them left to be worth ) Tj
+        0 -16 Td
+        (exhibiting - sometimes they would be overlooked for days, till all but the bones of them had gone out) Tj
+        0 -16 Td
+        (to the world as Durham's Pure Leaf Lard!) Tj
+    ET
+Q
+q
+    1 0 0 1 50 440 cm
+    0 G
+    2 w
+    0 0 m
+    500 0 l
+    S
+    BT
+    0 -30 Td
+    /F1 12 Tf
+    (Violations) Tj
+    ET
+Q
+q
+    1 0 0 1 50 400 cm
+    0.5 G
+    0 0 m
+    500 0 l
+    S
+    0 -20 m
+    500 -20 l
+    S
+    0 -40 m
+    500 -40 l
+    S
+    0 -60 m
+    500 -60 l
+    S
+    0 -80 m
+    500 -80 l
+    S
+    0 -100 m
+    500 -100 l
+    S
+    0 -120 m
+    500 -120 l
+    S
+    0 -140 m
+    500 -140 l
+    S
+    0 -160 m
+    500 -160 l
+    S
+    0 0 m
+    0 -160 l
+    S
+    50 0 m
+    50 -160 l
+    S
+    400 0 m
+    400 -160 l
+    S
+    450 0 m
+    450 -160 l
+    S
+    500 0 m
+    500 -160 l
+    S
+Q
+q
+    1 0 0 1 55 386 cm
+    BT
+        q
+            /F1 10 Tf
+            (Statute) Tj
+            50 0 Td
+            (Description) Tj
+            350 0 Td
+            (Level) Tj
+            50 0 Td
+            (Repeat?) Tj
+        Q
+        /F0 10 Tf
+        q
+            0 -20 Td
+            (4.12.7) Tj
+            50 0 Td
+            (Unsanitary Working Conditions.) Tj
+            350 0 Td
+            (Critical) Tj
+        Q
+        q
+            0 -40 Td
+            (5.8.3) Tj
+            50 0 Td
+            (Inadequate Protective Equipment.) Tj
+            350 0 Td
+            (Serious) Tj
+        Q
+        q
+            0 -60 Td
+            (6.3.9) Tj
+            50 0 Td
+            (Ineffective Injury Prevention.) Tj
+            350 0 Td
+            (Serious) Tj
+        Q
+        q
+            0 -80 Td
+            (7.1.5) Tj
+            50 0 Td
+            (Failure to Properly Store Hazardous Materials.) Tj
+            350 0 Td
+            (Critical) Tj
+        Q
+        q
+            0 -100 Td
+            (8.9.2) Tj
+            50 0 Td
+            (Lack of Adequate Fire Safety Measures.) Tj
+            350 0 Td
+            (Serious) Tj
+        Q
+        q
+            0 -120 Td
+            (9.6.4) Tj
+            50 0 Td
+            (Inadequate Ventilation Systems.) Tj
+            350 0 Td
+            (Serious) Tj
+        Q
+        q
+            0 -140 Td
+            (10.2.7) Tj
+            50 0 Td
+            (Insufficient Employee Training for Safe Work Practices.) Tj
+            350 0 Td
+            (Serious) Tj
+        Q
+    ET
+    q
+        0 G
+        0.5 w
+        465 -20 8 8 re
+        S
+        465 -20 m
+        473 -12 l
+        s
+        465 -12 m
+        473 -20 l
+        s
+        465 -40 8 8 re
+        S
+        465 -40 m
+        473 -32 l
+        s
+        465 -32 m
+        473 -40 l
+        s
+        465 -60 8 8 re
+        S
+        465 -80 8 8 re
+        S
+        465 -100 8 8 re
+        S
+        465 -120 8 8 re
+        S
+        465 -120 m
+        473 -112 l
+        s
+        465 -112 m
+        473 -120 l
+        s
+        465 -140 8 8 re
+        S
+    Q
+Q
+q
+    1 0 0 1 230 20 cm
+    BT
+        /F0 8 Tf
+        (Jungle Health and Safety Inspection Service) Tj
+    ET
+Q
+endstream
+endobj
+5 0 obj
+<</Type/Catalog/Pages 1 0 R>>
+endobj
+6 0 obj
+<</Length 2774>>
+stream
+q
+1 0 0 1 50 700 cm
+1 g
+325 65 180 -35 re
+S
+Q
+q
+1 0 0 1 50 700 cm
+BT
+335 50 Td
+/F0 8 Tf
+(Jungle Health and Safety Inspection Service) Tj
+0 -10 Td
+1 0 0 rg
+(INS-UP70N51NCL41R) Tj
+ET
+Q
+q
+1 0 0 1 50 700 cm
+BT
+/F1 10 Tf
+(Site: ) Tj
+/F0 10 Tf
+(Durham's Meatpacking  ) Tj
+.5 g
+(Chicago, Ill.) Tj
+0 g
+0 -20 Td
+/F1 10 Tf
+(Date:  ) Tj
+/F0 10 Tf
+(February 3, 1905) Tj
+0 -20 Td
+/F1 10 Tf
+(Violation Count: ) Tj
+/F0 10 Tf
+(7) Tj
+0 -20 Td
+/F1 10 Tf
+(Summary: ) Tj
+/F0 10 Tf
+(Worst of any, however, were the fertilizer men, and those who served in the cooking rooms.) Tj
+0 -16 Td
+(These people could not be shown to the visitor - for the odor of a fertilizer man would scare any ordinary ) Tj
+0 -16 Td
+(visitor at a hundred yards, and as for the other men, who worked in tank rooms full of steam, and in ) Tj
+0 -16 Td
+(some of which there were open vats near the level of the floor, their peculiar trouble was that they fell) Tj
+0 -16 Td
+(into the vats; and when they were fished out, there was never enough of them left to be worth ) Tj
+0 -16 Td
+(exhibiting - sometimes they would be overlooked for days, till all but the bones of them had gone out) Tj
+0 -16 Td
+(to the world as Durham's Pure Leaf Lard!) Tj
+ET
+Q
+q
+1 0 0 1 50 440 cm
+0 G
+2 w
+0 0 m
+500 0 l
+S
+BT
+0 -30 Td
+/F1 12 Tf
+(Violations) Tj
+ET
+Q
+q
+1 0 0 1 50 400 cm
+.5 G
+0 0 m
+500 0 l
+S
+0 -20 m
+500 -20 l
+S
+0 -40 m
+500 -40 l
+S
+0 -60 m
+500 -60 l
+S
+0 -80 m
+500 -80 l
+S
+0 -100 m
+500 -100 l
+S
+0 -120 m
+500 -120 l
+S
+0 -140 m
+500 -140 l
+S
+0 -160 m
+500 -160 l
+S
+0 0 m
+0 -160 l
+S
+50 0 m
+50 -160 l
+S
+400 0 m
+400 -160 l
+S
+450 0 m
+450 -160 l
+S
+500 0 m
+500 -160 l
+S
+Q
+q
+1 0 0 1 55 386 cm
+BT
+q
+/F1 10 Tf
+(Statute) Tj
+50 0 Td
+(Description) Tj
+350 0 Td
+(Level) Tj
+50 0 Td
+(Repeat?) Tj
+Q
+/F0 10 Tf
+q
+0 -20 Td
+(4.12.7) Tj
+50 0 Td
+(Unsanitary Working Conditions.) Tj
+350 0 Td
+(Critical) Tj
+Q
+q
+0 -40 Td
+(5.8.3) Tj
+50 0 Td
+(Inadequate Protective Equipment.) Tj
+350 0 Td
+(Serious) Tj
+Q
+q
+0 -60 Td
+(6.3.9) Tj
+50 0 Td
+(Ineffective Injury Prevention.) Tj
+350 0 Td
+(Serious) Tj
+Q
+q
+0 -80 Td
+(7.1.5) Tj
+50 0 Td
+(Failure to Properly Store Hazardous Materials.) Tj
+350 0 Td
+(Critical) Tj
+Q
+q
+0 -100 Td
+(8.9.2) Tj
+50 0 Td
+(Lack of Adequate Fire Safety Measures.) Tj
+350 0 Td
+(Serious) Tj
+Q
+q
+0 -120 Td
+(9.6.4) Tj
+50 0 Td
+(Inadequate Ventilation Systems.) Tj
+350 0 Td
+(Serious) Tj
+Q
+q
+0 -140 Td
+(10.2.7) Tj
+50 0 Td
+(Insufficient Employee Training for Safe Work Practices.) Tj
+350 0 Td
+(Serious) Tj
+Q
+ET
+q
+0 G
+.5 w
+465 -20 8 8 re
+S
+465 -20 m
+473 -12 l
+s
+465 -12 m
+473 -20 l
+s
+465 -40 8 8 re
+S
+465 -40 m
+473 -32 l
+s
+465 -32 m
+473 -40 l
+s
+465 -60 8 8 re
+S
+465 -80 8 8 re
+S
+465 -100 8 8 re
+S
+465 -120 8 8 re
+S
+465 -120 m
+473 -112 l
+s
+465 -112 m
+473 -120 l
+s
+465 -140 8 8 re
+S
+Q
+Q
+q
+1 0 0 1 230 20 cm
+BT
+/F0 8 Tf
+(Jungle Health and Safety Inspection Service) Tj
+ET
+Q
+endstream
+endobj
+xref
+0 7
+0000000000 65536 f
+0000000016 00000 n
+0000000068 00000 n
+0000000165 00000 n
+0000000300 00000 n
+0000004833 00000 n
+0000004879 00000 n
+trailer
+<</Size 7/Root 5 0 R>>
+startxref
+7704
+%%EOF

pdfs/0500000US42001.pdf ADDED Viewed

Binary file

pdfs/0500000US42007.pdf ADDED Viewed

Binary file

pdfs/2014 Statistics.pdf ADDED Viewed

Binary file

pdfs/2019 Statistics.pdf ADDED Viewed

Binary file

pdfs/Atlanta_Public_Schools_GA_sample.pdf ADDED Viewed

Binary file

pdfs/needs-ocr.pdf ADDED Viewed

Binary file

natural-pdf 0.1.4__py3-none-any.whl → 0.1.6__py3-none-any.whl

natural-pdf 0.1.4py3-none-any.whl → 0.1.6py3-none-any.whl