natural-pdf 0.1.1__py3-none-any.whl → 0.1.3__py3-none-any.whl

{natural_pdf-0.1.1.dist-info → natural_pdf-0.1.3.dist-info}/RECORD

@@ -1,4 +1,4 @@
-natural_pdf/__init__.py,sha256=hsSosbPnvDRCfyYAL9bf1haVS6oBxLAl7cbKTWRTHkU,1784
+natural_pdf/__init__.py,sha256=hdqbTG3SHtu8jPIL7su6TpEhEbNsL89pgktCXPMKWCI,2825
 natural_pdf/analyzers/__init__.py,sha256=BkSmEqw5J76C2fvYHF86EXQJQWWFNIvjSwRMwfW-Ht0,140
 natural_pdf/analyzers/text_options.py,sha256=9IGRoem1O2mc1ZNGiM5-VPRZ3c8LLwEk1B3is9UxMoE,2777
 natural_pdf/analyzers/text_structure.py,sha256=e4G6v0bD7ZJCdo6DcuDD3iZt8KAwBfALMduwZHGh0wI,12415
@@ -6,25 +6,28 @@ natural_pdf/analyzers/utils.py,sha256=u5_FAUPmEG1ydPVuxpu7bVw507NB3WzisMNSUhsnuk
 natural_pdf/analyzers/layout/__init__.py,sha256=oq1uJ5UkGGMbBKGirV1aRKK3hxAUyjTLywYkPCQH1f0,33
 natural_pdf/analyzers/layout/base.py,sha256=D6KHDsbVKzZWCfW4vt0khPC3TA9JzQD3cF4VtTSyf28,6752
 natural_pdf/analyzers/layout/docling.py,sha256=iNeD10ZfolDVJNqayAUd0-Bs2tVr5INE7WK9c_Mll_8,11930
-natural_pdf/analyzers/layout/layout_analyzer.py,sha256=oQeqPDHL6vpj_3NHuzS5ja7KVAAL7PhQ7IOwustDBBo,8008
-natural_pdf/analyzers/layout/layout_manager.py,sha256=Qr5pxcv_Wk5IJRJr0IoYJJAz71RGJvQgqXONNBhNLOw,9221
-natural_pdf/analyzers/layout/layout_options.py,sha256=vZvTSg_M27OirZomcC5uWLSmPYXjvnnCEo5QKy9RjaQ,3503
+natural_pdf/analyzers/layout/layout_analyzer.py,sha256=JJasXl7QEiP4DgAvf-zu1w7Uakdf8ypvITkpQ-OQDgA,13340
+natural_pdf/analyzers/layout/layout_manager.py,sha256=6Zi9SBonpa0urWyeQBJnmxIL1hOn4xAx09ugkMrEhro,9555
+natural_pdf/analyzers/layout/layout_options.py,sha256=EmvPEnDsVGMJkDNfn6ORLnX545gbmlo3kVcz4anVm5Q,3325
 natural_pdf/analyzers/layout/paddle.py,sha256=QCasH_Z9UITX6wRGlE_HjmwkBuANz9Yyw5Yk7QvRVcw,12519
-natural_pdf/analyzers/layout/surya.py,sha256=hmPDfXzTkF2PQPgvg1xjTJSBKFuCmjZB3GTCHa-kpA4,6477
+natural_pdf/analyzers/layout/surya.py,sha256=Ibwo42TioJ-BZP3-2T13KCtH3kLSWQh7C9ZYuk1kUQo,12657
 natural_pdf/analyzers/layout/tatr.py,sha256=H0Xygk9jA46-vlPleoal94cuDyz-LHTSxVb3e6gpmV8,11956
 natural_pdf/analyzers/layout/yolo.py,sha256=NSQK3TcS1qN8D2MDxCvcwTpS_kvzGy3I2LepJDUceoQ,7699
+natural_pdf/collections/pdf_collection.py,sha256=Da8saWBTguxk16pNzMxCrFwatrWk_qrcG0RVPQybro8,12159
 natural_pdf/core/__init__.py,sha256=GUuFtj2Apc9biAdUOlnL8leL3BQncEzubvpiAUaU3ss,37
 natural_pdf/core/element_manager.py,sha256=H1896JSt48ASLSmG22xEXMY-xSKcpYsUlYmYMD48i6Q,17117
 natural_pdf/core/highlighting_service.py,sha256=a-40UMohOglYrw4klW1GuQ_p3jZOxnAfPOXPORThr4U,31476
-natural_pdf/core/page.py,sha256=tnxG-5OhFVuFHt0p-a9YSLU-nXjA8fftg5ViQdH5sOU,68512
-natural_pdf/core/pdf.py,sha256=UzxVfVeCnhSN7rxdJresUj_UNFkcFkeaEjLvwZMJS-c,28532
+natural_pdf/core/page.py,sha256=qhumZqmwHoBlGodiCvYE0z34Iu1WSs32V4_Iz_Sfaow,69350
+natural_pdf/core/pdf.py,sha256=MLN-asJ_d5spmCjLz7SDp74t__vioszfKEFooBul7nU,41167
 natural_pdf/elements/__init__.py,sha256=6FGHZm2oONd8zErahMEawuB4AvJR5jOZPt4KtEwbj80,40
 natural_pdf/elements/base.py,sha256=9SQ-O2qbQe9Avbf9JI-p6vWlyThZVch-p1yqXWSrBHw,35750
 natural_pdf/elements/collections.py,sha256=RJf4cBZeLfCtfS0-SjzYFRCtbzYjWsgk3LrcTwJAYMs,62392
 natural_pdf/elements/line.py,sha256=QvVdhf_K6rwJkq3q67JmgdZpDhrBgWuSMF-Q25malP4,4783
 natural_pdf/elements/rect.py,sha256=dls9g-R213O78HvfAJMak3_eV14Zh654Zw7hqTTXxDQ,3949
-natural_pdf/elements/region.py,sha256=MXQK00LLMvwuq94NigeeCVFoGov_RWFe9ZylnIMpzB0,72453
+natural_pdf/elements/region.py,sha256=5dXHYbbdO1QNgkD6b6I34ezHt-SHKx_aH1ubzbfMHQs,74370
 natural_pdf/elements/text.py,sha256=OAuy0ozaemj6yjMwhXPsJ76VZtRPeJbmrFTzpDJA2_U,11017
+natural_pdf/exporters/__init__.py,sha256=Nqnn8clbgv-5l0PgxcTOldg8mkMKrFn4TvPL-rYUUGg,1
+natural_pdf/exporters/searchable_pdf.py,sha256=PPkF64hFNNhPlZPuyJRvC_scAg3WCOiIvwgIP8nlZ9E,10225
 natural_pdf/ocr/__init__.py,sha256=mbUUsCfeU6yRsEqNn3I4Len-XY6FfjfKhTAoWDLA1f4,1943
 natural_pdf/ocr/engine.py,sha256=xDnvhnm4Lr7d83ezglDqOtl9xfx74zOOTyYW-fZHQEQ,4183
 natural_pdf/ocr/engine_easyocr.py,sha256=6srZhXqlH3UpNWw5iFq7u4TS5HQsMSTWYuuWo3oYZp8,8273
@@ -34,6 +37,12 @@ natural_pdf/ocr/ocr_manager.py,sha256=mAyCntdAnrNv8TIvGYlGs40G2tDAdMQ_Jqb3owiPWW
 natural_pdf/ocr/ocr_options.py,sha256=A2CQV172id-90zMpPZWb8CD09ZP0BuQnnCZGEFP4SaQ,3787
 natural_pdf/qa/__init__.py,sha256=kagdfqNMpTnyzjC2EFy_PBX5us38NnJL548ESSQVzfI,107
 natural_pdf/qa/document_qa.py,sha256=QYKKor0RqUQcEdFEBEUdq7L0ktq1WSMfQ-ynTc64cPU,15926
+natural_pdf/search/__init__.py,sha256=sYv7-XrSohUgE2UH8sFpGfl66SG092jZoNokZaDdxsY,4125
+natural_pdf/search/haystack_search_service.py,sha256=qhvqVJMxz4-KTnQF0MPO7YLQxTlYe27PCgKJgYeAels,27580
+natural_pdf/search/haystack_utils.py,sha256=BXU5yIEcFIWliSX44slMYLlUMfwCXEfve-ZYmVcEt3k,18773
+natural_pdf/search/search_options.py,sha256=PrIGkvM9A9wpqaz6tDB-9hWiSp9fqhi8mf7FQl1qoGI,3510
+natural_pdf/search/search_service_protocol.py,sha256=5EYzHFUoFvaYw3khnQNz1dsOHqTvBChekvk_qf2mu5w,6811
+natural_pdf/search/searchable_mixin.py,sha256=QPtPSJHCP5n0Twp4uHKSns8J6HuvGjyipTNbB66JFLg,24896
 natural_pdf/selectors/__init__.py,sha256=Jfk-JBZEpQ7V5FWVGuLJQLH-qOfqNLC2AdicncMhrmY,121
 natural_pdf/selectors/parser.py,sha256=JK1zDVISACkUhzmzWfQMMW8hvsV422lRBFKgDBWOWC4,24108
 natural_pdf/templates/__init__.py,sha256=i7N8epDxZoDDsK4p2iUiMwzKVs97i_KtNk8ATArqlC4,19
@@ -45,8 +54,8 @@ natural_pdf/utils/visualization.py,sha256=14BM-K4ovDqHniNbxbP_y9KaEYNlkbpELGAv9_
 natural_pdf/widgets/__init__.py,sha256=qckw3DjdVTsASPLJ8uUrGKg3MFhvzHndUpeNGlqwg6A,215
 natural_pdf/widgets/viewer.py,sha256=h_amj_uvf-vRqEsFg4P00fgKxawLAd9jjC1ohUza4BY,37479
 natural_pdf/widgets/frontend/viewer.js,sha256=w8ywfz_IOAAv2nP_qaf2VBUkF1KhjT3zorhJxM1-CfU,4371
-natural_pdf-0.1.1.dist-info/licenses/LICENSE,sha256=9zfwINwJlarbDmdh6iJV4QUG54QSJlSAUcnC1YiC_Ns,1074
-natural_pdf-0.1.1.dist-info/METADATA,sha256=8o22GEPtEqlSqexFQxy6tVoHTB35LmT63sjbjbjORRE,10009
-natural_pdf-0.1.1.dist-info/WHEEL,sha256=CmyFI0kx5cdEMTLiONQRbGQwjIoR1aIYB7eCAQ4KPJ0,91
-natural_pdf-0.1.1.dist-info/top_level.txt,sha256=XtfS3IiR1fTjaQG9TjGDjZsB1Ih2GXQteDbJ2dXlLvQ,12
-natural_pdf-0.1.1.dist-info/RECORD,,
+natural_pdf-0.1.3.dist-info/licenses/LICENSE,sha256=9zfwINwJlarbDmdh6iJV4QUG54QSJlSAUcnC1YiC_Ns,1074
+natural_pdf-0.1.3.dist-info/METADATA,sha256=kBSb1SueOGQFw97pvHBxlJYcuNwxAB-lInLKows0BEs,5069
+natural_pdf-0.1.3.dist-info/WHEEL,sha256=CmyFI0kx5cdEMTLiONQRbGQwjIoR1aIYB7eCAQ4KPJ0,91
+natural_pdf-0.1.3.dist-info/top_level.txt,sha256=XtfS3IiR1fTjaQG9TjGDjZsB1Ih2GXQteDbJ2dXlLvQ,12
+natural_pdf-0.1.3.dist-info/RECORD,,

natural_pdf-0.1.1.dist-info/METADATA

@@ -1,295 +0,0 @@
-Metadata-Version: 2.4
-Name: natural-pdf
-Version: 0.1.1
-Summary: A more intuitive interface for working with PDFs
-Author-email: Jonathan Soma <jonathan.soma@gmail.com>
-License-Expression: MIT
-Project-URL: Homepage, https://github.com/jsoma/natural-pdf
-Project-URL: Repository, https://github.com/jsoma/natural-pdf
-Classifier: Programming Language :: Python :: 3
-Classifier: Operating System :: OS Independent
-Requires-Python: >=3.7
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: pdfplumber>=0.7.0
-Requires-Dist: Pillow>=8.0.0
-Requires-Dist: colour>=0.1.5
-Requires-Dist: numpy>=1.20.0
-Requires-Dist: urllib3>=1.26.0
-Requires-Dist: torch>=2.0.0
-Requires-Dist: torchvision>=0.15.0
-Requires-Dist: transformers>=4.30.0
-Requires-Dist: huggingface_hub>=0.19.0
-Provides-Extra: interactive
-Requires-Dist: ipywidgets<9.0.0,>=7.0.0; extra == "interactive"
-Provides-Extra: easyocr
-Requires-Dist: easyocr; extra == "easyocr"
-Provides-Extra: paddle
-Requires-Dist: paddlepaddle; extra == "paddle"
-Requires-Dist: paddleocr; extra == "paddle"
-Provides-Extra: layout-yolo
-Requires-Dist: doclayout_yolo; extra == "layout-yolo"
-Provides-Extra: surya
-Requires-Dist: surya-ocr; extra == "surya"
-Provides-Extra: qa
-Provides-Extra: all
-Requires-Dist: ipywidgets<9.0.0,>=7.0.0; extra == "all"
-Requires-Dist: easyocr; extra == "all"
-Requires-Dist: paddlepaddle; extra == "all"
-Requires-Dist: paddleocr; extra == "all"
-Requires-Dist: doclayout_yolo; extra == "all"
-Requires-Dist: surya-ocr; extra == "all"
-Dynamic: license-file
-
-# Natural PDF
-
-A friendly library for working with PDFs, built on top of [pdfplumber](https://github.com/jsvine/pdfplumber).
-
-Natural PDF lets you find and extract content from PDFs using simple code that makes sense.
-
-- [Complete documentation here](https://jsoma.github.io/natural-pdf)
-- [Live demos here](https://colab.research.google.com/github/jsoma/natural-pdf/)
-
-## Features
-
-- **Fluent API** for chaining operations
-- **CSS-like selectors** for finding elements
-- **Spatial navigation** with intuitive methods like `above()`, `below()`, and `select_until()`
-- **Element collections** for batch operations
-- **Visual highlighting** for debugging (persistent highlights)
-- **Interactive element viewer** for Jupyter environments (`.viewer()`)
-- **Region visualization** with direct image extraction of specific regions
-- **Text style analysis** for document structure
-- **Exclusion zones** for headers, footers, and other areas to ignore
-- **OCR integration** with multiple engines (EasyOCR, PaddleOCR, Surya)
-- **Document layout analysis** for detecting document structure with ML models
-- **Table extraction** with multiple detection methods
-- **Structured logging** with configurable levels and handlers
-
-## Installation
-
-```bash
-pip install natural-pdf
-```
-
-# Installs the core library along with required AI dependencies (PyTorch, Transformers)
-```bash
-# Install with support for specific OCR and layout engines
-pip install natural-pdf[easyocr]
-pip install natural-pdf[paddle]
-pip install natural-pdf[surya]
-pip install natural-pdf[layout_yolo]
-
-# Install with support for the interactive Jupyter widget
-pip install natural-pdf[interactive]
-
-# Just install everything
-pip install natural-pdf[all]
-```
-
-## Quick Start
-
-```python
-from natural_pdf import PDF
-
-# Open a local PDF
-pdf = PDF('document.pdf')
-
-# Or open a PDF from a URL
-pdf = PDF('https://example.com/document.pdf')
-
-# Get the first page
-page = pdf.pages[0]
-
-# Find elements using CSS-like selectors
-heading = page.find('text:contains("Summary"):bold')
-
-# Extract content below the heading
-content = heading.below().extract_text()
-print(content)
-
-# Exclude headers and footers
-page.add_exclusion(page.find('text:contains("CONFIDENTIAL")').above())
-page.add_exclusion(page.find_all('line')[-1].below())
-
-# Extract clean text
-clean_text = page.extract_text()
-print(clean_text)
-```
-
-- [Complete documentation here](https://jsoma.github.io/natural-pdf)
-
-## Selectors
-
-The library supports CSS-like selectors for finding elements:
-
-```python
-# Find text containing a specific string
-element = page.find('text:contains("Revenue")')
-
-# Find bold text with a specific font size
-headings = page.find_all('text[size>=12]:bold')
-
-# Find thick red lines
-lines = page.find_all('line[width>=2][color~=(1,0,0)]')
-```
-
-## Spatial Navigation
-
-Navigate through the document with intuitive spatial methods:
-
-```python
-# Get content below a heading
-heading = page.find('text:contains("Introduction")')
-content = heading.below().extract_text()
-
-# Get content from one element to another
-start = page.find('text:contains("Start")')
-end = page.find('text:contains("End")')
-region = start.select_until(end)
-content = region.extract_text()
-```
-
-## Exclusion Zones
-
-Exclude headers, footers, or other areas from extraction:
-
-```python
-# Page-level exclusion
-page.add_exclusion(page.find('text:contains("Page")').above())
-page.add_exclusion(page.find_all('line')[-1].below())
-
-# PDF-level exclusion with lambdas
-pdf.add_exclusion(
-    lambda page: page.find('text:contains("Header")').above(),
-    label="headers"
-)
-
-# Extract text with exclusions applied
-text = pdf.extract_text()
-
-# Extract from a specific region with exclusions
-summary = page.find('text:contains("Summary")')
-conclusion = page.find('text:contains("Conclusion")')
-region = page.create_region(summary.x0, summary.top, conclusion.x1, conclusion.bottom)
-region_text = region.extract_text(apply_exclusions=True)  # Excludes headers/footers
-
-# Disable exclusions for a specific extraction
-full_text = page.extract_text(apply_exclusions=False)
-```
-
-Exclusions work efficiently with different region types:
-- Regions without intersection with exclusion zones → exclusions ignored entirely
-- Rectangular regions with header/footer exclusions → optimized cropping
-- Complex regions with partial exclusions → advanced filtering with warning
-
-## OCR Integration
-
-Extract text from scanned documents using OCR, with support for multiple engines ([EasyOCR](https://www.jaided.ai/easyocr/), [PaddleOCR](https://paddlepaddle.github.io/PaddleOCR/latest/en/index.html), [Surya](https://github.com/VikParuchuri/surya)):
-
-```python
-# Apply OCR using a specific engine (e.g., PaddleOCR)
-ocr_elements = page.apply_ocr(engine='paddle', languages=['en', 'zh-cn'])
-
-# Extract text (will use previously applied OCR results if available)
-text = page.extract_text()
-
-# Configure advanced engine options using Options classes
-from natural_pdf.ocr import PaddleOCROptions
-paddle_opts = PaddleOCROptions(languages=['en'], use_angle_cls=False, rec_batch_num=8)
-ocr_elements = page.apply_ocr(engine='paddle', options=paddle_opts)
-
-# Force OCR regardless of existing text
-ocr_text = page.extract_text(ocr=True)
-
-# Find OCR-detected text with high confidence
-high_confidence = page.find_all('text[source=ocr][confidence>=0.8]')
-
-# Visualize OCR results with color-coded confidence levels
-for elem in page.find_all('text[source=ocr]'):
-    if elem.confidence >= 0.8:
-        color = (0, 1, 0, 0.3)  # Green for high confidence
-    elif elem.confidence >= 0.5:
-        color = (1, 1, 0, 0.3)  # Yellow for medium confidence
-    else:
-        color = (1, 0, 0, 0.3)  # Red for low confidence
-        
-    elem.highlight(color=color, label=f"OCR ({elem.confidence:.2f})")
-page.save_image('ocr_results.png', labels=True)
-```
-
-## Logging
-
-The library includes a structured logging system to provide visibility into its operations:
-
-```python
-import logging
-from natural_pdf import PDF, configure_logging
-
-# Configure logging with INFO level to console
-configure_logging(level=logging.INFO)
-
-# Or log to a file with DEBUG level
-file_handler = logging.FileHandler("natural_pdf.log")
-file_handler.setFormatter(logging.Formatter('%(asctime)s - %(name)s - %(levelname)s - %(message)s'))
-configure_logging(level=logging.DEBUG, handler=file_handler)
-
-# Now operations will generate logs
-pdf = PDF("document.pdf")
-# Log: natural_pdf.core.pdf - INFO - Initializing PDF from document.pdf
-
-# Run layout detection with verbose logging
-regions = pdf.pages[0].analyze_layout(
-    model="paddle",
-    model_params={"verbose": True}
-)
-# Log: natural_pdf.analyzers.layout.paddle - INFO - Starting PaddleLayout detection...
-# Log: natural_pdf.analyzers.layout.paddle - DEBUG - Parameters: confidence=0.2...
-```
-
-Logs follow a hierarchical structure matching the library's module organization:
-- `natural_pdf.core` - Core PDF operations
-- `natural_pdf.analyzers` - Layout analysis operations
-- `natural_pdf.ocr` - OCR engine operations
-
-## Document QA
-
-Ask questions directly to your documents:
-
-```python
-# Ask questions about the document content
-result = pdf.ask("What was the company's revenue in 2022?")
-print(f"Answer: {result['answer']}")
-print(f"Confidence: {result['confidence']:.2f}")
-
-# Access more details in the result dictionary
-result = pdf.ask("Who is the CEO?")
-print(f"Answer: {result['answer']}")
-print(f"Found on page: {result['page_num']}")
-print(f"Source text: {result.get('source_text', 'N/A')}")
-```
-
-## More details
-
-[Complete documentation here](https://jsoma.github.io/natural-pdf)
-
-## Visual Debugging & Interactive Viewer
-
-Use highlighting to understand element selection and analysis results. Add persistent highlights using `.highlight()` and view them with the interactive `.viewer()` or static `.save_image()`. You can also generate temporary previews of selected elements using `ElementCollection.show()`.
-
-```python
-# Highlight selected elements persistently
-page.find_all('text:bold').highlight(label="Bold Text")
-
-# Launch the interactive widget in Jupyter (shows persistent highlights)
-# Requires: pip install natural-pdf[interactive]
-page.viewer()
-
-# Save a static image file with highlights and legend
-page.save_image("highlighted_page.png", labels=True)
-
-# Show a temporary preview image of specific elements, grouped by attribute
-preview_image = page.find_all('region[type*=table]').show(group_by='type')
-# In Jupyter, this image will display automatically
-preview_image
-```