jedarden
31e715633d
Add four research documents on text quality and document-type handling
...
- text-readability-validation: character/word/entropy/perplexity checks,
symbol font detection, remediation decision tree, span quality metadata
- post-ocr-text-correction: error taxonomy, confusable tables, noisy channel
n-gram model, regex patterns, hyphenation, layout-based correction pipeline
- presentation-and-spreadsheet-pdfs: detection heuristics, slide structure,
bullet hierarchy, speaker notes, hairline grid detection, sheet boundaries,
cell type inference, Rust output schema
- semantic-text-reconstruction: beam search n-gram reconstruction, NER
validation, domain lexicons, cross-span consistency, abbreviation expansion,
citation repair, coherence scoring, ReconstructedSpan output schema
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-05-16 15:07:30 -04:00
jedarden
a7673c906f
Add 12 research documents covering full PDF extraction surface
...
Infrastructure and parsing:
- raster-ocr-pipeline: trigger detection, preprocessing, Tesseract integration,
assisted OCR, HOCR alignment, multi-language, performance
- image-and-figure-extraction: XObjects, inline images, filter decoding,
color spaces, geometry, form XObjects, transparency, figure detection
- form-fields-and-annotations: AcroForm types, XFA, widget appearance
streams, rich text, annotation text, output schema
- pdf-encryption-and-security: R2-R6 key derivation, object-level
decryption, permission flags, RustCrypto implementation approach
- page-geometry-and-document-structure: page tree, all five page boxes,
rotation, coordinate inversion, page labels, outlines, named destinations
- optional-content-groups: OCG/OCMD visibility, usage dictionary, default
state resolution, content stream marking, multilingual layer patterns
- invisible-and-hidden-text: all 8 Tr modes, PDF/A invisible layer pattern,
white-on-white, zero-opacity, clipped text, color tracking
- malformed-pdf-repair-and-recovery: xref recovery, stream length repair,
syntax tolerance, partial extraction, structured warnings
Quality and metadata:
- xmp-and-document-metadata: /Info vs XMP, all namespaces, RDF/XML
parsing, conflict resolution, encrypted metadata, thumbnails
- embedded-files-and-portfolios: EmbeddedFile streams, Filespec,
AF relationships, Portfolio detection, ZUGFeRD/Factur-X, security
- performance-and-streaming-architecture: mmap, lazy loading, NDJSON
streaming, rayon parallelism, font caching, axum HTTP server
- benchmark-and-test-methodology: CER/WER/TEDS metrics, corpus
categories, reading order scoring, regression CI, public datasets
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-05-16 15:05:42 -04:00
jedarden
b805593973
Add six research documents covering output-side extraction topics
...
- table-structure-reconstruction: line detection, gap analysis, Hough
transform, graph-based cell reconstruction, merged cells, multi-page tables
- mathematical-expression-handling: five encoding cases, OpenType MATH table,
symbol font recovery, spatial heuristics, LaTeX reconstruction, fallback tiers
- language-detection-and-script-handling: UAX #24/#9, Arabic/Hebrew bidi,
CJK vertical text, ligature normalization, whatlang/lingua integration
- document-classification-and-zone-labeling: margin heuristics, font
clustering, cross-page recurrence, footnote/caption/sidebar detection
- post-extraction-normalization: hyphen handling, ligature expansion,
paragraph reconstruction, Unicode normalization, pipeline ordering
- chunking-for-llm-consumption: semantic snapping, heading hierarchy,
sliding window overlap, table chunking strategies, token budget, late chunking
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-05-16 14:56:25 -04:00
jedarden
c2870e6640
Add research docs and SDK invocation notes
...
Four research documents covering PDF spec fundamentals, font types and
encoding, glyph Unicode recovery, and tagged PDF structure/reading order.
SDK invocation notes with subprocess and HTTP examples for Python, Node.js,
Go, Ruby, Java, Rust, and Bash.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-05-16 14:33:34 -04:00