easypdf-reader

PDF reading layer: parsing, text extraction, page manipulation (merge/split/rotate/reorder/watermark), with three adaptive read strategies.

Role

easypdf-reader handles all PDF input operations in the easypdf-rust workspace. Built on the lopdf backend, it provides text extraction, metadata reading, page manipulation (merge, split, rotate, reorder, watermark, layer), and PDF/A validation. It automatically selects the optimal read strategy based on file size to balance speed for small files against memory efficiency for large ones.

Core Capabilities

• Three read strategies (Full / Lazy / Streaming) -- auto-selected by file size: 0-5 MB = Full, 5-100 MB = Lazy, >100 MB = Streaming (crates/easypdf-reader/src/strategy.rs:56-68)
• Text extraction (extract_text()) -- plain text from PDF, with CMap/ToUnicode support for CJK fonts (crates/easypdf-reader/src/reader/extract.rs)
• Page manipulation (PdfManipulator) -- merge, split, rotate, reorder, extract pages, add text watermark, add optional content groups (layers) (crates/easypdf-reader/src/manipulate.rs)
• PDF repair (open_with_repair()) -- auto-detect and fix corrupted PDF files (crates/easypdf-core/src/io/repair.rs)
• Resource guards -- decompression bomb and element explosion protection via ResourceLimits (crates/easypdf-core/src/io/guards.rs)
• Streaming scanner (StreamScanner) -- byte-stream scanning without building a full Document object for very large files (crates/easypdf-reader/src/streaming/)
• PDF/A validation (validate_pdfa()) -- check PDF/A-1b compliance (crates/easypdf-reader/src/manipulate.rs:260)
• Benchmark suite -- reader session benchmarks (crates/easypdf-reader/benches/reader_session.rs)

Dependencies

Internal

External

Main API

PdfReader

use easypdf_reader::{PdfReader, ReadStrategy};

// Auto strategy (selected by file size)
let reader = PdfReader::open("document.pdf")?;
let text = reader.extract_text()?;

// Specify strategy
let reader = PdfReader::open_with_strategy("large.pdf", ReadStrategy::Lazy)?;
let text = reader.pages(0..5).extract_text()?;

// From memory bytes
let reader = PdfReader::from_bytes(pdf_bytes)?;

// With auto-repair
let reader = PdfReader::open_with_repair("corrupted.pdf", true, ReadStrategy::Full)?;

// With custom resource limits
let reader = PdfReader::open_with_limits(input, ResourceLimits::default())?;

ReadStrategy

use easypdf_reader::ReadStrategy;

// Auto-select by file size
let strategy = ReadStrategy::auto(50_000_000); // 50 MB -> Lazy

// Manual selection
let s = ReadStrategy::Full;     // <5 MB, full object tree
let s = ReadStrategy::Lazy;     // 5-100 MB, lazy page loading
let s = ReadStrategy::Streaming; // >100 MB, byte-stream scan

PdfManipulator

use easypdf_reader::PdfManipulator;

// Merge multiple PDFs
PdfManipulator::merge_files(&["a.pdf", "b.pdf"], "merged.pdf")?;

// Open and manipulate
let mut m = PdfManipulator::open("input.pdf")?;
m.rotate_page(0, Rotation::Clockwise90)?;
m.reorder_pages(&[2, 0, 1])?;
m.extract_pages(&(0..5))?;
m.add_text_watermark("CONFIDENTIAL", 48.0, 0.3)?;
m.add_layer("Annotations")?;
m.validate_pdfa()?;

Known Limitations

• ReadStrategy::Streaming does not build a complete object tree -- precision is lower than Full/Lazy, especially for CJK text boundaries (crates/easypdf-reader/src/strategy.rs:47-51)
• Streaming mode skips cross-reference parsing and font encoding (CMap/ToUnicode) for speed

License

Apache-2.0

Project: https://github.com/easy-4-rust/easypdf-rust crates.io: https://crates.io/crates/easypdf-reader docs.rs: https://docs.rs/easypdf-reader