# fxread

[![MIT licensed](https://img.shields.io/badge/license-MIT-blue.svg)](./LICENSE.md)
![actions status](https://github.com/noamteyssier/fxread/workflows/CI/badge.svg)
[![codecov](https://codecov.io/gh/noamteyssier/fxread/branch/main/graph/badge.svg?token=N0TEG1JVJ8)](https://codecov.io/gh/noamteyssier/fxread)

A barebones fastx reader for rust.

## Summary

This crate attempts to be a faster and more lightweight alternative to [`bio-rs`](https://crates.io/crates/bio) and provides a standardized interface to working with fasta and fastq formats.
The goal of this crate is to be fast and flexible - it is about twice as fast as [`bio-rs`](https://crates.io/crates/bio) on average but about half as fast than [`fastq`](https://crates.io/crates/fastq) for standard fastx files. 
The difference between the different crates is reduced heavily though once gzip files are included (see [benchmark](https://github.com/noamteyssier/fxread_benchmark)). 

The speed up can be attributed to reducing the total number of vectors allocated for each record - but the limitation compared to [`fastq`](https://crates.io/crates/bio) is that each record has ownership over its data and is allocated once.
This creates extra overhead, but is very convenient as you can treat the reader directly as an iterator.

## Usage

Some benefits of this interface is that each `FastaReader` and `FastqReader` share the `FastxReader` trait and act as iterators over `Record`s.

### `initialize_reader` can determine the fastq format from the path name

```rust
use fxread::initialize_reader;

let path = "example/sequences.fq";
let reader = initialize_reader(path).unwrap();
assert_eq!(reader.count(), 10);
```

### `initialize_reader` can handle if the file is gzip or not without changing the downstream usage

```rust
use fxread::initialize_reader;

let path = "example/sequences.fq.gz";
let reader = initialize_reader(path).unwrap();
assert_eq!(reader.count(), 10);
```

Check out the [API Documentation](https://docs.rs/fxread) for usage