### ✍️ Intelligent Generation (Builder)
* **Strict EPUB 2 / 3 Generation**: Conditional compilation isolating legacy `NCX` from modern `NAV`, generating compliant `content.opf`.
* **Streaming Large Files**: Stream massive assets (videos/images) directly into the EPUB ZIP pipe without loading them into memory (`add_resource_stream`).
* **Rich Metadata & Layouts**: Full Dublin Core property refinements (authors vs. translators), Pre-Paginated Fixed-Layout (FXL), and Page Spreads (Comics/Manga).
* **Automatic Property Inference**: Automatically detects ``, `<svg>`, and `<math>` to inject EPUB 3 required properties.
* **Landmarks & Page-Lists**: Build comprehensive guide mappings for academic and textbook parity.

---

## Installation

Add this to your `Cargo.toml`:

```toml
[dependencies]
epub-rs = "0.1.0"
```

## Quick Start

### 1. Read an EPUB & Extract Text
```rust
use epub_rs::parser::EpubArchive;
use std::fs::File;

fn main() -> Result<(), Box<dyn std::error::Error>> {
let file = File::open("book.epub")?;
let mut archive = EpubArchive::new(file)?;

// Parse OPF and metadata
let book = archive.parse()?;
println!("Title: {:?}", book.metadata.title);

// Read the first chapter from the spine
let first_chapter_id = &book.spine[0].idref;
let html_bytes = archive.get_resource_by_id(&book, first_chapter_id)?;

// Extract plain text for search indexing
let plain_text = epub_rs::processor::extract_text(&html_bytes)?;
println!("Content: {}", plain_text);

Ok(())
}
```

### 2. Generate a Compliant EPUB 3 Book
```rust
use epub_rs::generator::{EpubBuilder, TocEntry};
use epub_rs::model::{Creator, EpubVersion, Metadata};
use std::fs::File;

fn main() {
let metadata = Metadata {
title: Some("My Awesome Book".to_string()),
creators: vec![Creator::new("Rustacean")],
language: Some("en".to_string()),
..Default::default()
};

let builder = EpubBuilder::new()
.version(EpubVersion::V30)
.metadata(metadata)
// Auto-inject built-in typography and dark mode CSS
.theme(epub_rs::generator::Theme::Modern)
// Add nested table of contents
.set_toc(vec![TocEntry::new("Chapter 1", "text/ch1.xhtml")])
// Generate book and HTML
.add_chapter("ch1", "text/ch1.xhtml", b"<h1>Hello</h1><p>World!</p>".to_vec());

let mut file = File::create("output.epub").unwrap();
builder.generate(&mut file).expect("Failed to generate EPUB");
}
```

### 3. Build a Web Reader (CFI Injection)
Pass HTML directly to the browser with exact book-location identifiers, removing the need for complex frontend calculation.

```rust
let chapter_html_with_cfi = archive.get_chapter_with_cfi(&book, "chapter_1_id")?;
// output: <p data-cfi="epubcfi(/6/4!/4/2)">...</p>
```

### 4. Semantic TTS Extraction
Ideal for Text-To-Speech, extracts language-tagged block structures instead of flat strings.

```rust
let elements = archive.get_semantic_content(&book, "chapter_1_id")?;
for el in elements {
println!("Read {} (in {:?}): {}", el.tag_name, el.language, el.text);
// e.g., Read p (in Some("fr")): Bonjour!
}
```

## Performance
Built on Cloudflare's `lol_html` and `zip-rs`, `epub-rs` processes DOMs in a single pass without allocating heavy AST trees.

* **~20 µs**: Open ZIP, parse OPF, setup Domain Models (10 chapters).
* **~140 µs**: Build, assemble, and compress a full EPUB to memory.
* **~30 µs**: Find 50 regex text matches and reverse-map them to exact CFI ranges.

*(Benchmarks executed on Apple Silicon M-series via `cargo bench`)*

## License
MIT License
s**: Build, assemble, and compress a full EPUB to memory.
* **~30 µs**: Find 50 regex text matches and reverse-map them to exact CFI ranges.

*(Benchmarks executed on Apple Silicon M-series via `cargo bench`)*