Allow fasta_digest to accept Path or bytes

[nsheff]

Right now digest_fasta strictly requires a string.

works:

from gtars.digests import digest_fasta
path_to_fasta_file = "../gtars/gtars/tests/data/base.fa"
df = digest_fasta(path_to_fasta_file)

Doesn't work:

from pathlib import Path
p = Path(path_to_fasta_file)
digest_fasta(p)
# TypeError: argument 'fasta': 'PosixPath' object cannot be converted to 'PyString'

Also fails:

digest_fasta(path_to_fasta_file.encode('utf-8'))
TypeError: argument 'fasta': 'bytes' object cannot be converted to 'PyString'

Would it be easy/possible/advisable to make this friendly enough to convert string-like inputs to string automatically?

Raised by @stolarczyk

[nleroy917]

PyO3 has a lot of native mappings from Python types to Rust types. If you want to accept Path objects and str objects from Python into Rust, I think the function argument needs to be of type PathBuf. Although, if I am reading their chart correctly, PyString should be sufficient. You currently have it at &str so maybe that's the issue.

that said... I wonder if we could leverage the .extract() API from pyo3 and do a nested if let Ok(...) pattern to accept multiple types.

Something like:

#[pyfunction]
pub fn digest_fasta(fasta: PyObject) -> PyResult<Vec<PyDigestResult>> {
  
  let fasta_file;
  if let Ok(fasta) = fasta.extract::<PyString>() {
    fasta_file = fasta;
  } else if let Ok(fasta) = fasta.extract::<PathBuf>() {
     fasta_file = fasta;
  } else {
    return Err(PyErr::new::<pyo3::exceptions::PyIOError, _>(format!("Must be of type String or Path"))
  }

    match gtars::digests::digest_fasta(fasta) {
        Ok(digest_results) => {
            let py_digest_results: Vec<PyDigestResult> = digest_results.into_iter().map(PyDigestResult::from).collect();
            Ok(py_digest_results)
        },
        Err(e) => Err(PyErr::new::<pyo3::exceptions::PyIOError, _>(format!("Error processing FASTA file: {}", e))),
    }
}

[nsheff]

I went a different direction in the end.

The .extract works, but this copies the python string into a rust object. so you're duplicating data. this is very bad news for something when you're copying a huge string you're trying to digest (like a chromosome).

Instead I can use .downsample with PyAny. There's also a new way in PyO3 recommend for using PyAny, where you use a Bound type. ANyway, I got this all working and you can see this in my PR.

[donaldcampbelljr]

This appears to be solved, so I'm going to close.