use std::{str::FromStr, sync::Arc};
use arrow::{error::ArrowError, error::Result as ArrowResult, record_batch::RecordBatch};
use futures::Stream;
use tokio::io::{AsyncBufRead, AsyncBufReadExt};
use super::{array_builder::GFFArrayBuilder, GFFConfig};
pub struct BatchReader<R> {
reader: R,
config: Arc<GFFConfig>,
region: Option<Arc<noodles::core::Region>>,
}
impl<R> BatchReader<R>
where
R: AsyncBufRead + Unpin + Send,
{
pub fn new(reader: R, config: Arc<GFFConfig>) -> Self {
Self {
reader,
config,
region: None,
}
}
pub fn with_region(mut self, region: Arc<noodles::core::Region>) -> Self {
self.region = Some(region);
self
}
pub fn into_stream(self) -> impl Stream<Item = Result<RecordBatch, ArrowError>> {
futures::stream::unfold(self, |mut reader| async move {
match reader.read_batch().await {
Ok(Some(batch)) => Some((Ok(batch), reader)),
Ok(None) => None,
Err(e) => Some((Err(ArrowError::ExternalError(Box::new(e))), reader)),
}
})
}
async fn read_line(&mut self) -> std::io::Result<Option<noodles::gff::Line>> {
loop {
let mut buf = String::new();
match self.reader.read_line(&mut buf).await {
Ok(0) => return Ok(None),
Ok(_) => {
buf.pop();
#[cfg(target_os = "windows")]
if buf.ends_with('\r') {
buf.pop();
}
let line = match noodles::gff::Line::from_str(&buf) {
Ok(line) => line,
Err(e) => match e {
noodles::gff::line::ParseError::InvalidDirective(_) => {
continue;
}
noodles::gff::line::ParseError::InvalidRecord(e) => {
return Err(std::io::Error::new(
std::io::ErrorKind::InvalidData,
format!("invalid record: {buf} error: {e}"),
))
}
},
};
buf.clear();
return Ok(Some(line));
}
Err(e) => return Err(e),
};
}
}
fn filter(&self, record: &noodles::gff::Record) -> Result<bool, ArrowError> {
let chrom = record.reference_sequence_name();
match &self.region {
Some(region) => {
let region_name = std::str::from_utf8(region.name())?;
if chrom != region_name {
return Ok(false);
}
if !region.interval().contains(record.start()) {
return Ok(false);
}
Ok(true)
}
None => Ok(true),
}
}
async fn read_batch(&mut self) -> ArrowResult<Option<RecordBatch>> {
let mut gff_array_builder = GFFArrayBuilder::new();
let mut batch_size = 0;
loop {
match self.read_line().await? {
None => break,
Some(line) => match line {
noodles::gff::Line::Comment(_) => {}
noodles::gff::Line::Directive(_) => {}
noodles::gff::Line::Record(record) => {
if !self.filter(&record)? {
continue;
}
gff_array_builder.append(&record)?;
batch_size += 1;
if batch_size == self.config.batch_size {
break;
}
}
},
}
}
if gff_array_builder.is_empty() {
return Ok(None);
}
let batch =
RecordBatch::try_new(self.config.file_schema.clone(), gff_array_builder.finish())?;
match &self.config.projection {
Some(projection) => Ok(Some(batch.project(projection)?)),
None => Ok(Some(batch)),
}
}
}