feat: add PDF test corpus (1,905 files, 10 source groups)

.gitignore

@@ -13,3 +13,11 @@ desktop.ini
 # Temporary files
 *.tmp
 *.bak
+
+# Local download/scraping scripts — NOT for committing
+scrape/
+
+# Python bytecode
+__pycache__/
+*.pyc
+*.pyo

ATTRIBUTION.json

@@ -230,8 +230,195 @@
     "author": "The Apache Software Foundation and ODF Toolkit contributors",
     "license": "Apache-2.0",
     "source": "https://github.com/tdf/odftoolkit",
-    "tags": ["odftoolkit", "odt", "odf", "test-fixtures"],
+    "tags": [
+      "odftoolkit",
+      "odt",
+      "odf",
+      "test-fixtures"
+    ],
     "donated": "2026-04-01",
     "notes": "ODF DOM, validation, and specification compliance test fixtures from the ODF Toolkit project"
+  },
+  {
+    "format": "pdf",
+    "path": "documents/pdf/pdf-js/*.pdf",
+    "title": "Mozilla pdf.js Test PDFs",
+    "author": "Mozilla Corporation and pdf.js contributors",
+    "license": "Apache-2.0",
+    "source": "https://github.com/mozilla/pdf.js",
+    "tags": [
+      "pdf-js",
+      "test-fixtures",
+      "annotations",
+      "forms",
+      "images",
+      "text-rendering",
+      "encryption"
+    ],
+    "donated": "2026-04-03",
+    "notes": "PDF test fixtures from Mozilla pdf.js viewer (Apache-2.0, pinned to v4.9.155)"
+  },
+  {
+    "format": "pdf",
+    "path": "documents/pdf/qpdf/*.pdf",
+    "title": "qpdf PDF Transformation Tool Test PDFs",
+    "author": "Jay Berkenbilt and qpdf contributors",
+    "license": "Apache-2.0",
+    "source": "https://github.com/qpdf/qpdf",
+    "tags": [
+      "qpdf",
+      "test-fixtures",
+      "encryption",
+      "page-operations",
+      "compression",
+      "edge-cases"
+    ],
+    "donated": "2026-04-03",
+    "notes": "PDF test fixtures from the qpdf PDF transformation tool (Apache-2.0, pinned to v12.3.2)"
+  },
+  {
+    "format": "pdf",
+    "path": "documents/pdf/apache-pdfbox/*.pdf",
+    "title": "Apache PDFBox Test PDFs",
+    "author": "The Apache Software Foundation and PDFBox contributors",
+    "license": "Apache-2.0",
+    "source": "https://github.com/apache/pdfbox",
+    "tags": [
+      "apache-pdfbox",
+      "test-fixtures",
+      "forms",
+      "text-extraction",
+      "images"
+    ],
+    "donated": "2026-04-03",
+    "notes": "PDF test fixtures from Apache PDFBox Java PDF library (Apache-2.0, pinned to 3.0.7)"
+  },
+  {
+    "format": "pdf",
+    "path": "documents/pdf/pikepdf/*.pdf",
+    "title": "pikepdf Test PDFs",
+    "author": "James R. Barlow and pikepdf contributors",
+    "license": "MPL-2.0",
+    "source": "https://github.com/pikepdf/pikepdf",
+    "tags": [
+      "pikepdf",
+      "test-fixtures",
+      "color-spaces",
+      "pdf-a"
+    ],
+    "donated": "2026-04-03",
+    "notes": "PDF test fixtures from the pikepdf Python PDF library (MPL-2.0)"
+  },
+  {
+    "format": "pdf",
+    "path": "documents/pdf/arxiv/*.pdf",
+    "title": "arXiv CC-BY-4.0 Open Access Papers",
+    "author": "Various authors (see individual papers at arxiv.org)",
+    "license": "CC-BY-4.0",
+    "source": "https://arxiv.org",
+    "tags": [
+      "arxiv",
+      "academic",
+      "text-heavy",
+      "equations",
+      "figures",
+      "multi-column"
+    ],
+    "donated": "2026-04-03",
+    "notes": "CC-BY-4.0 papers from arXiv (strictly filtered by license metadata) across CS, Math, Physics, Biology, Economics, Statistics"
+  },
+  {
+    "format": "pdf",
+    "path": "documents/pdf/openstax/*.pdf",
+    "title": "OpenStax Textbook PDFs",
+    "author": "OpenStax, Rice University",
+    "license": "CC-BY-4.0",
+    "source": "https://openstax.org",
+    "tags": [
+      "openstax",
+      "textbooks",
+      "educational",
+      "tables",
+      "images",
+      "equations",
+      "multi-column"
+    ],
+    "donated": "2026-04-03",
+    "notes": "Open educational textbook PDFs from OpenStax (CC-BY-4.0). Chapter excerpts extracted to meet size limits."
+  },
+  {
+    "format": "pdf",
+    "path": "documents/pdf/irs-forms/*.pdf",
+    "title": "IRS Tax Form PDFs",
+    "author": "Internal Revenue Service, U.S. Department of the Treasury",
+    "license": "CC0-1.0",
+    "source": "https://www.irs.gov/forms-instructions",
+    "tags": [
+      "irs",
+      "government",
+      "forms",
+      "fillable-forms",
+      "tables",
+      "public-domain"
+    ],
+    "donated": "2026-04-03",
+    "notes": "U.S. federal tax forms and instructions (public domain as U.S. government works; attributed as CC0-1.0)"
+  },
+  {
+    "format": "pdf",
+    "path": "documents/pdf/nasa-ntrs/*.pdf",
+    "title": "NASA Technical Reports",
+    "author": "National Aeronautics and Space Administration",
+    "license": "CC0-1.0",
+    "source": "https://ntrs.nasa.gov",
+    "tags": [
+      "nasa",
+      "government",
+      "technical-reports",
+      "text-heavy",
+      "images",
+      "tables",
+      "public-domain"
+    ],
+    "donated": "2026-04-03",
+    "notes": "NASA technical reports and publications (public domain as U.S. government works; attributed as CC0-1.0)"
+  },
+  {
+    "format": "pdf",
+    "path": "documents/pdf/gutenberg/*.pdf",
+    "title": "Project Gutenberg Public Domain Books",
+    "author": "Various (original works in public domain)",
+    "license": "CC0-1.0",
+    "source": "https://www.gutenberg.org",
+    "tags": [
+      "gutenberg",
+      "public-domain",
+      "books",
+      "text-heavy",
+      "multilingual",
+      "literature"
+    ],
+    "donated": "2026-04-03",
+    "notes": "Public domain books from Project Gutenberg in PDF format (multiple languages: English, German, French, Spanish, Italian)"
+  },
+  {
+    "format": "pdf",
+    "path": "documents/pdf/internet-archive/*.pdf",
+    "title": "Internet Archive Public Domain Scanned Books",
+    "author": "Various (original works in public domain; digitization by Internet Archive)",
+    "license": "CC0-1.0",
+    "source": "https://archive.org",
+    "tags": [
+      "internet-archive",
+      "scanned",
+      "ocr",
+      "public-domain",
+      "multilingual",
+      "rtl",
+      "cjk",
+      "books"
+    ],
+    "donated": "2026-04-03",
+    "notes": "Scanned public domain books from Internet Archive. Includes OCR text layers and multilingual content (Arabic, Chinese, French, German, Spanish, Russian, English) for testing scanned PDF handling and non-Latin script support."
   }
 ]

ATTRIBUTION.md

@@ -212,3 +212,133 @@ This file tracks all donated documents, their authors, licenses, and sources.
 - **Donated**: 2026-04-01
 - **Notes**: ODF DOM, validation, and specification compliance test fixtures from the ODF Toolkit project
 
+---
+
+## Mozilla pdf.js Test PDFs
+
+- **Format**: `pdf`
+- **Path**: `documents/pdf/pdf-js/*.pdf`
+- **Author**: Mozilla Corporation and pdf.js contributors
+- **License**: Apache-2.0
+- **Source**: https://github.com/mozilla/pdf.js
+- **Tags**: `pdf-js`, `test-fixtures`, `annotations`, `forms`, `images`, `text-rendering`, `encryption`
+- **Donated**: 2026-04-03
+- **Notes**: PDF test fixtures from Mozilla pdf.js viewer (Apache-2.0, pinned to v4.9.155)
+
+---
+
+## qpdf PDF Transformation Tool Test PDFs
+
+- **Format**: `pdf`
+- **Path**: `documents/pdf/qpdf/*.pdf`
+- **Author**: Jay Berkenbilt and qpdf contributors
+- **License**: Apache-2.0
+- **Source**: https://github.com/qpdf/qpdf
+- **Tags**: `qpdf`, `test-fixtures`, `encryption`, `page-operations`, `compression`, `edge-cases`
+- **Donated**: 2026-04-03
+- **Notes**: PDF test fixtures from the qpdf PDF transformation tool (Apache-2.0, pinned to v12.3.2)
+
+---
+
+## Apache PDFBox Test PDFs
+
+- **Format**: `pdf`
+- **Path**: `documents/pdf/apache-pdfbox/*.pdf`
+- **Author**: The Apache Software Foundation and PDFBox contributors
+- **License**: Apache-2.0
+- **Source**: https://github.com/apache/pdfbox
+- **Tags**: `apache-pdfbox`, `test-fixtures`, `forms`, `text-extraction`, `images`
+- **Donated**: 2026-04-03
+- **Notes**: PDF test fixtures from Apache PDFBox Java PDF library (Apache-2.0, pinned to 3.0.7)
+
+---
+
+## pikepdf Test PDFs
+
+- **Format**: `pdf`
+- **Path**: `documents/pdf/pikepdf/*.pdf`
+- **Author**: James R. Barlow and pikepdf contributors
+- **License**: MPL-2.0
+- **Source**: https://github.com/pikepdf/pikepdf
+- **Tags**: `pikepdf`, `test-fixtures`, `color-spaces`, `pdf-a`
+- **Donated**: 2026-04-03
+- **Notes**: PDF test fixtures from the pikepdf Python PDF library (MPL-2.0)
+
+---
+
+## arXiv CC-BY-4.0 Open Access Papers
+
+- **Format**: `pdf`
+- **Path**: `documents/pdf/arxiv/*.pdf`
+- **Author**: Various authors (see individual papers at arxiv.org)
+- **License**: CC-BY-4.0
+- **Source**: https://arxiv.org
+- **Tags**: `arxiv`, `academic`, `text-heavy`, `equations`, `figures`, `multi-column`
+- **Donated**: 2026-04-03
+- **Notes**: CC-BY-4.0 papers from arXiv (strictly filtered by license metadata) across CS, Math, Physics, Biology, Economics, Statistics
+
+---
+
+## OpenStax Textbook PDFs
+
+- **Format**: `pdf`
+- **Path**: `documents/pdf/openstax/*.pdf`
+- **Author**: OpenStax, Rice University
+- **License**: CC-BY-4.0
+- **Source**: https://openstax.org
+- **Tags**: `openstax`, `textbooks`, `educational`, `tables`, `images`, `equations`, `multi-column`
+- **Donated**: 2026-04-03
+- **Notes**: Open educational textbook PDFs from OpenStax (CC-BY-4.0). Chapter excerpts extracted to meet size limits.
+
+---
+
+## IRS Tax Form PDFs
+
+- **Format**: `pdf`
+- **Path**: `documents/pdf/irs-forms/*.pdf`
+- **Author**: Internal Revenue Service, U.S. Department of the Treasury
+- **License**: CC0-1.0
+- **Source**: https://www.irs.gov/forms-instructions
+- **Tags**: `irs`, `government`, `forms`, `fillable-forms`, `tables`, `public-domain`
+- **Donated**: 2026-04-03
+- **Notes**: U.S. federal tax forms and instructions (public domain as U.S. government works; attributed as CC0-1.0)
+
+---
+
+## NASA Technical Reports
+
+- **Format**: `pdf`
+- **Path**: `documents/pdf/nasa-ntrs/*.pdf`
+- **Author**: National Aeronautics and Space Administration
+- **License**: CC0-1.0
+- **Source**: https://ntrs.nasa.gov
+- **Tags**: `nasa`, `government`, `technical-reports`, `text-heavy`, `images`, `tables`, `public-domain`
+- **Donated**: 2026-04-03
+- **Notes**: NASA technical reports and publications (public domain as U.S. government works; attributed as CC0-1.0)
+
+---
+
+## Project Gutenberg Public Domain Books
+
+- **Format**: `pdf`
+- **Path**: `documents/pdf/gutenberg/*.pdf`
+- **Author**: Various (original works in public domain)
+- **License**: CC0-1.0
+- **Source**: https://www.gutenberg.org
+- **Tags**: `gutenberg`, `public-domain`, `books`, `text-heavy`, `multilingual`, `literature`
+- **Donated**: 2026-04-03
+- **Notes**: Public domain books from Project Gutenberg in PDF format (multiple languages: English, German, French, Spanish, Italian)
+
+---
+
+## Internet Archive Public Domain Scanned Books
+
+- **Format**: `pdf`
+- **Path**: `documents/pdf/internet-archive/*.pdf`
+- **Author**: Various (original works in public domain; digitization by Internet Archive)
+- **License**: CC0-1.0
+- **Source**: https://archive.org
+- **Tags**: `internet-archive`, `scanned`, `ocr`, `public-domain`, `multilingual`, `rtl`, `cjk`, `books`
+- **Donated**: 2026-04-03
+- **Notes**: Scanned public domain books from Internet Archive. Includes OCR text layers and multilingual content (Arabic, Chinese, French, German, Spanish, Russian, English) for testing scanned PDF handling and non-Latin script support.
+