pdf_text

Positioned PDF text extraction primitives for semantic PDF comparison.

pdf_text consumes interpreted content stream operations and emits positioned text runs with glyph evidence, normalized text, bounding boxes, provenance, and marked-content references. It is the text-extraction stage used by semantic-pdf-diff before layout clustering and semantic diffing.

What This Crate Provides

• Extraction of text from Tj and TJ operations.
• Basic text matrix, text line movement, leading, character spacing, word spacing, and horizontal scaling support.
• Glyph token evidence, raw text bytes, normalized text, and approximate bounding boxes.
• Marked-content tag and /MCID preservation for tagged-PDF mapping.
• A public font resource model with resource name, object ID, subtype, base-font, encoding, descendant-font, and /ToUnicode references.
• Whitespace normalization for deterministic downstream matching.
• MISSING_TOUNICODE diagnostics when extraction falls back to literal or hex string bytes instead of a Unicode map.

Pipeline Context

pdf_text sits between content interpretation and semantic layout:

pdf_content ContentProgram -> pdf_text TextRun values -> pdf_semantic nodes

The current CLI uses this crate's font resource model, then applies a narrow /ToUnicode CMap mapping before text-run extraction when page resources expose a decoded CMap stream. Broader font decoding and shaping remain planned work.

Current Compatibility Boundary

This crate is designed for digitally generated PDFs with extractable text. It does not perform OCR, full font shaping, real glyph outline measurement, or renderer-grade text positioning. Image-only PDFs should use the CLI's external OCR adapter path, not this crate as a substitute for OCR.

Bounding boxes are approximate extraction evidence for diff reports, not a pixel-accurate rendering contract.