pdftract/crates/pdftract-core
jedarden d84f8da3a4 feat(pdftract-5qj50): implement mojibake detection and repair via encoding_rs
Implements Phase 4.7 Correction Pipeline step 3: mojibake detection
and repair for Latin-1 bytes misinterpreted as UTF-8.

Changes:
- Add layout::correction module with detect_and_repair_mojibake function
- Implement CorrectableText trait for mutable text access
- Add trait implementations for hybrid::Span and schema::SpanJson
- Make encoding_rs a non-optional dependency (was cjk-gated)
- Detection heuristic: 2+ occurrences of telltale sequences (é, è, ’, etc.)
- Re-decode via encoding_rs::WINDOWS_1252 when detected
- Accept repair only if readability score improves by >0.05 epsilon
- Fast-path pass-through for ASCII-only and clean UTF-8 text

Closes: pdftract-5qj50

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-24 17:01:53 -04:00
..
benches feat(pdftract-9wevc): implement 20k English wordlist for readability scoring 2026-05-24 09:29:13 -04:00
build feat(pdftract-9wevc): implement 20k English wordlist for readability scoring 2026-05-24 09:29:13 -04:00
examples feat(pdftract-3s2i): implement Phase 5.5.2 validation filter 2026-05-24 04:57:17 -04:00
proptest-regressions/parser/lexer feat(pdftract-1jjn): implement PDF numeric literal lexer with full edge case support 2026-05-23 23:17:04 -04:00
src feat(pdftract-5qj50): implement mojibake detection and repair via encoding_rs 2026-05-24 17:01:53 -04:00
tests feat(bf-2ervu): implement mmap-backed PdfSource via memmap2 2026-05-24 08:40:11 -04:00
__test__.pdf feat(pdftract-15pz8): implement multi-process safe cache operations 2026-05-23 05:31:11 -04:00
build.rs feat(pdftract-9wevc): implement 20k English wordlist for readability scoring 2026-05-24 09:29:13 -04:00
Cargo.toml feat(pdftract-5qj50): implement mojibake detection and repair via encoding_rs 2026-05-24 17:01:53 -04:00
pdftract-core.cdx.json feat(pdftract-67tm8): implement MCP stdio transport with integration tests 2026-05-23 00:16:42 -04:00
README.md docs(pdftract-49f8): establish Cargo.lock policy and documentation 2026-05-20 18:13:14 -04:00

pdftract-core

The core Rust library for PDF text extraction. This crate provides the parsing, layout analysis, font encoding recovery, and text extraction primitives used by the CLI (pdftract-cli) and Python bindings (pdftract-py).

Cargo.lock Policy

This workspace checks in Cargo.lock at the repository root. This is unconventional for library crates—the Cargo Book historically suggested that only binary crates should check in lockfiles, allowing library consumers to resolve their own dependency versions.

pdftract departs from this convention for release reproducibility:

  1. SLSA Level 3 provenance requires that every milestone tag produces byte-identical artifacts across builds. Without a checked-in lockfile, two runs of cargo build on the same commit can resolve different transitive dependency versions, producing different binary hashes.

  2. Multi-output artifacts—this workspace produces Rust crates (pdftract-core, pdftract-cli), Python wheels (pdftract-py), and Docker images. All must be built from the same dependency tree.

  3. Supply-chain security—the lockfile pins checksums for all transitive dependencies, enabling cargo audit to detect yanked or compromised crates.

  4. Downstream consumers can still ignore the lockfile if needed. Cargo allows cargo build --frozen with a local lockfile override, or consumers can vendor the crate with their own dependency resolution.

The tradeoff—occasional merge conflicts when PRs update overlapping dependencies—is worth the guarantee of reproducible releases. See CONTRIBUTING.md for the lockfile-update workflow.

Modules

  • parser: PDF spec parsing (xref, trailer, object streams, indirect references)
  • font: Font encoding recovery, glyph name lookup, fingerprinting
  • layout: Page layout analysis, region segmentation, reading order
  • extract: Text extraction with provenance (bounding boxes, confidence scores)
  • ocr: Tesseract integration for raster pages

Usage

use pdftract_core::{extract_text, ExtractOptions};

let options = ExtractOptions::default();
let result = extract_text("document.pdf", &options)?;
println!("{}", result.text);