Skip to content

README.md

Latest commit

 

History

History
168 lines (113 loc) · 4.1 KB

README.md

File metadata and controls

168 lines (113 loc) · 4.1 KB
Raw

PDF Text Extractor

A lightweight Rust crate for extracting plain text from PDF documents. Designed for zero-knowledge environments with minimal dependencies.

🎯 Overview

The extractor crate provides a simple, dependency-light solution for reading textual content from PDF files. It implements a minimal subset of PDF text extraction rules without relying on heavy external libraries like lopdf or flate2.

🚀 Quick Start

use extractor::extract_text;

fn main() -> Result<(), Box<dyn std::error::Error>> {
    // Read PDF file
    let pdf_bytes = std::fs::read("document.pdf")?;

    // Extract text from all pages
    let pages = extract_text(pdf_bytes)?;

    // Print text from each page
    for (i, page_text) in pages.iter().enumerate() {
        println!("Page {}: {}", i + 1, page_text);
    }

    Ok(())
}

📋 Main Interface

pub fn extract_text(pdf_bytes: Vec<u8>) -> Result<Vec<String>, PdfError>

Parameters:

  • pdf_bytes: Raw PDF file bytes

Returns:

  • Ok(Vec<String>): Vector of strings, one per page
  • Err(PdfError): Error if extraction fails

🔤 Encoding & Font Support

The extractor supports a comprehensive set of font encodings commonly used in PDF documents:

Supported Encodings

  • StandardEncoding – Adobe's base Latin encoding (ISO 8859-1 subset)
  • WinAnsiEncoding – Windows-1252 encoding
  • MacRomanEncoding – Classic Mac OS encoding
  • MacExpertEncoding – Extended Mac encoding for typography
  • PDFDocEncoding – Default encoding for PDF strings

Font Features

  • Glyph name mapping – Converts font glyph names to Unicode characters
  • CID font support – Handles Composite Font (CID) structures
  • ToUnicode mapping – Supports Unicode character mapping tables
  • Font differences – Handles custom character substitutions

Character Coverage

The extractor includes a comprehensive lookup table for common glyph names, supporting:

  • Latin characters (A-Z, a-z)
  • Accented characters (é, ñ, ü, etc.)
  • Special symbols (©, ®, ™, etc.)
  • Mathematical symbols (×, ÷, ±, etc.)
  • Currency symbols (€, £, ¥, etc.)

🏗️ Architecture

Core Components

  • PDF Parser – Lightweight PDF structure parser
  • Font Decoder – Font encoding and glyph mapping
  • Text Stream Processor – Extracts text from PDF streams
  • Decompression – Handles zlib/deflate compression

Dependencies

  • miniz_oxide – Pure Rust zlib/deflate decompression
  • No external PDF libraries
  • Zero-knowledge VM compatible

📝 Usage Examples

Basic Text Extraction

use extractor::extract_text;

let pdf_bytes = include_bytes!("sample.pdf");
let pages = extract_text(pdf_bytes.to_vec())?;
println!("Extracted {} pages", pages.len());

Working with Specific Pages

let pages = extract_text(pdf_bytes)?;

// Get first page
if let Some(first_page) = pages.first() {
    println!("First page: {}", first_page);
}

// Get last page
if let Some(last_page) = pages.last() {
    println!("Last page: {}", last_page);
}

// Check if specific text exists
let search_text = "Important Document";
let found = pages.iter().any(|page| page.contains(search_text));
println!("Text found: {}", found);

🧪 Testing

Public Tests

Run the basic test suite:

cargo test -p extractor

Private Tests

Run tests with sample PDF files:

cargo test -p extractor --features private_tests

⚠️ Limitations

Supported PDF Features

  • ✅ Simple text extraction
  • ✅ Common font encodings
  • ✅ Basic compression (zlib/deflate)
  • ✅ Standard PDF structure

Unsupported Features

  • ❌ Image or graphics extraction
  • ❌ Form field extraction
  • ❌ Advanced font features (ligatures, kerning)
  • ❌ PDF/A or PDF/X specific features

🤝 Contributing

When contributing to the extractor:

  • Keep dependencies minimal
  • Ensure ZK-VM compatibility
  • Add tests for new encodings
  • Document any breaking changes
  • Maintain performance for simple PDFs

📄 License

This crate is licensed under the same terms as the parent repository.