# FB2 for Rust

This library is a set of models that enable [quick-xml](https://crates.io/crates/quick-xml) to deserialize an FB2 file in a structured way.
The library tries to parse an invalid FB2, fix it, and output a valid one.
That's because plenty of real FB2 books don't follow the standard...
We can either ignore invalid elements, fail deserialization, or accept them and try to map them to the standard schema.
This model tries to follow the third approach where it makes sense.
The library model almost conforms to the [standard XSD schema](https://github.com/gribuser/fb2)
with a few exceptions:

- the first section content element can be an image
- sections can have both content and nested sections
- most fields are optional contrary to the XSD schema
- validation of strings is not performed besides structured dates, versions, sequence number, and language tags
- binaries are not associated with the image references, e.g. `Rc<Binary>`
- xs:ID uniqueness is not verified
- minOccurs/maxOccurs are not verified, i.e. the parser doesn't verify whether there is at most 2 "output" elements
- xs:gYear with a timezone fails deserialization
- xs:date with a timezone fails deserialization
- XML sequence is not always enforced

The current version of the model enables quick-xml to deserialize 95% of FB2 files if not more.

# Example

```rust
use fb2::FictionBook;
use std::fs::File;
use std::io::BufReader;

fn main() {
    let file = File::open("examples/books/churchill_trial.fb2").unwrap();
    let reader = BufReader::new(file);
    let book: FictionBook = quick_xml::de::from_reader(reader).unwrap();
    println!("{:#?}", book);
}
```

Try with:

```shell
cargo run --example parse_sample
```

# How to deserialize windows-1251 (or other encoding)

Enable the `encoding` feature of quick-xml:

```toml
quick-xml = { version = "<version>", features = ["encoding", "serialize"] }
```

Then, deserialize as usual.