pandora_lib_promethion/src/io/writers.rs

use std::{
    fs::{self, File, OpenOptions},
    io::{BufWriter, Write},
    path::Path,
};

use anyhow::Context;
// use bgzip::{BGZFWriter, Compression};
use log::info;

use crate::io::readers::get_reader;

// pub fn get_gz_writer(path: &str, force: bool) -> anyhow::Result<BGZFWriter<File>> {
//     if !path.ends_with(".gz") {
//         anyhow::bail!("The file should end with gz");
//     }
//
//     if force && Path::new(path).exists() {
//         fs::remove_file(path).with_context(|| anyhow::anyhow!("Failed to remove file: {path}"))?;
//     }
//
//     let file = OpenOptions::new()
//         .write(true) // Open the file for writing
//         .create_new(true)
//         .truncate(true)
//         .open(path)
//         .with_context(|| anyhow::anyhow!("failed to open the file: {path}"))?;
//
//     info!("Writing into {path}");
//     Ok(BGZFWriter::new(file, Compression::default()))
// }

use noodles_bgzf as bgzf;

pub fn get_gz_writer(
    path: &str,
    force: bool,
) -> anyhow::Result<bgzf::io::Writer<BufWriter<std::fs::File>>> {
    if !path.ends_with(".gz") {
        anyhow::bail!("file should end with .gz: {path}");
    }

    if force && std::path::Path::new(path).exists() {
        fs::remove_file(path).with_context(|| format!("failed to remove file: {path}"))?;
    }

    let file = OpenOptions::new()
        .write(true)
        .create_new(true)
        .open(path)
        .with_context(|| format!("failed to create BGZF file: {path}"))?;

    Ok(bgzf::io::Writer::new(BufWriter::new(file)))
}

pub fn get_writer(path: &str) -> anyhow::Result<BufWriter<File>> {
    let file = OpenOptions::new()
        .write(true) // Open the file for writing
        .create_new(true)
        .truncate(true)
        .open(path)
        .with_context(|| anyhow::anyhow!("failed to open the file: {path}"))?;

    info!("Writing into {path}");
    Ok(BufWriter::new(file))
}

pub fn convert_bgz(input: impl AsRef<Path>, force: bool) -> anyhow::Result<()> {
    let input = input.as_ref();

    let mut reader = get_reader(&input.to_string_lossy())?;
    let mut writer = get_gz_writer(&format!("{}.gz", input.display()), force)?;

    std::io::copy(&mut reader, &mut writer)?;

    writer
        .try_finish()
        .with_context(|| format!("failed finishing BGZF writer: {}", input.display()))?;

    let mut inner = writer
        .finish()
        .with_context(|| format!("failed returning inner BGZF writer: {}", input.display()))?;

    inner
        .flush()
        .with_context(|| format!("failed flushing inner writer: {}", input.display()))?;

    inner
        .into_inner()
        .with_context(|| format!("failed unwrapping BufWriter: {}", input.display()))?
        .sync_all()
        .with_context(|| format!("failed syncing file: {}", input.display()))?;

    Ok(())
}

pub fn finalize_bgzf_file(
    writer: bgzf::io::Writer<BufWriter<File>>,
    path: &str,
) -> anyhow::Result<()> {
    let mut inner = writer
        .finish()
        .with_context(|| format!("failed finishing BGZF writer: {path}"))?;

    inner
        .flush()
        .with_context(|| format!("failed flushing BufWriter: {path}"))?;

    inner
        .into_inner()
        .with_context(|| format!("failed unwrapping BufWriter: {path}"))?
        .sync_all()
        .with_context(|| format!("failed syncing file: {path}"))?;

    Ok(())
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::helpers::test_init;

    #[test]
    fn bgz_convert() -> anyhow::Result<()> {
        test_init();

        let a = "/mnt/beegfs02/scratch/t_steimle/data/wgs/DUMCO/diag/DUMCO_diag_modkit_pileup.bed";
        convert_bgz(a, true)?;

        let a = "/mnt/beegfs02/scratch/t_steimle/data/wgs/DUMCO/norm/DUMCO_norm_modkit_pileup.bed";
        convert_bgz(a, true)?;

        Ok(())
    }
}