use mzdata::mz_read

.github/workflows/test.yml

@@ -16,8 +16,19 @@ jobs:
  runs-on: ubuntu-latest
  steps:
  - uses: actions/checkout@v4
+
+ - name: Install stable toolchain
+ uses: actions-rs/toolchain@v1
+ with:
+ profile: minimal
+ toolchain: stable
+ override: true
+
  - name: Run Clippy
- run: cargo clippy --all-targets --all-features
+ uses: actions-rs/cargo@v1
+ with:
+ command: clippy
+ args: --all-targets --all-features
 
  pytest:
  runs-on: ${{ matrix.os }}

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "ms2rescore-rs"
-version = "0.3.0"
+version = "0.4.0"
 edition = "2021"
 
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
@@ -10,5 +10,6 @@ crate-type = ["cdylib"]
 
 [dependencies]
 pyo3 = "0.20.0"
-mzdata = "0.20.0"
+mzdata = "0.26.0"
 timsrust = "0.3.0"
+mzpeaks = "0.19.0"

src/file_types.rs

@@ -1,25 +1,34 @@
+use mzdata::io::MassSpectrometryFormat;
+
 pub enum SpectrumFileType {
  MascotGenericFormat,
  MzML,
+ MzMLb,
  BrukerRaw,
- // ThermoRaw,
+ ThermoRaw,
  Unknown,
 }
 
 pub fn match_file_type(spectrum_path: &str) -> SpectrumFileType {
- let extension = spectrum_path.split('.').last().unwrap_or("").to_lowercase();
- match extension.as_str() {
- "mgf" => SpectrumFileType::MascotGenericFormat,
- "mzml" => SpectrumFileType::MzML,
- "d" | "ms2" => SpectrumFileType::BrukerRaw,
- // "raw" => SpectrumFileType::ThermoRaw,
- _ => match (
- folder_contains_extension(spectrum_path, "bin"),
- folder_contains_extension(spectrum_path, "parquet"),
- ) {
- (true, true) => SpectrumFileType::BrukerRaw,
- _ => SpectrumFileType::Unknown,
- },
+ match mzdata::io::infer_from_path(spectrum_path).0 {
+ MassSpectrometryFormat::MGF => SpectrumFileType::MascotGenericFormat,
+ MassSpectrometryFormat::MzML => SpectrumFileType::MzML,
+ MassSpectrometryFormat::MzMLb => SpectrumFileType::MzMLb,
+ MassSpectrometryFormat::ThermoRaw => SpectrumFileType::ThermoRaw,
+ MassSpectrometryFormat::Unknown => {
+ let extension = spectrum_path.split('.').last().unwrap_or("").to_lowercase();
+ match extension.as_str() {
+ "d" | "ms2" => SpectrumFileType::BrukerRaw,
+ _ => match (
+ folder_contains_extension(spectrum_path, "bin"),
+ folder_contains_extension(spectrum_path, "parquet"),
+ ) {
+ (true, true) => SpectrumFileType::BrukerRaw,
+ _ => SpectrumFileType::Unknown,
+ },
+ }
+ }
+ _ => SpectrumFileType::Unknown
  }
 }
 

src/lib.rs

@@ -6,30 +6,37 @@ mod ms2_spectrum;
 
 use std::collections::HashMap;
 
-use pyo3::exceptions::PyOSError;
+use pyo3::exceptions::{PyException, PyValueError};
 use pyo3::prelude::*;
 
 use file_types::{match_file_type, SpectrumFileType};
 use precursor::Precursor;
 use ms2_spectrum::MS2Spectrum;
 
+/// Check if spectrum path matches a supported file type.
+#[pyfunction]
+pub fn is_supported_file_type(spectrum_path: String) -> bool {
+ let file_type = match_file_type(&spectrum_path);
+
+ !matches!(file_type, SpectrumFileType::Unknown)
+}
+
 /// Get mapping of spectrum identifiers to precursor information.
 #[pyfunction]
 pub fn get_precursor_info(spectrum_path: String) -> PyResult<HashMap<String, Precursor>> {
  let file_type = match_file_type(&spectrum_path);
 
  let precursors = match file_type {
- SpectrumFileType::MascotGenericFormat | SpectrumFileType::MzML => {
- parse_mzdata::parse_precursor_info(&spectrum_path, file_type)
+ SpectrumFileType::MascotGenericFormat | SpectrumFileType::MzML | SpectrumFileType::MzMLb | SpectrumFileType:: ThermoRaw => {
+ parse_mzdata::parse_precursor_info(&spectrum_path)
  }
  SpectrumFileType::BrukerRaw => parse_timsrust::parse_precursor_info(&spectrum_path),
- // SpectrumFileType::ThermoRaw => parse_with_mzdata_thermo(&spectrum_path, file_type),
- SpectrumFileType::Unknown => return Err(PyOSError::new_err("Unsupported file type")),
+ SpectrumFileType::Unknown => return Err(PyValueError::new_err("Unsupported file type")),
  };
 
  match precursors {
  Ok(precursors) => Ok(precursors),
- Err(e) => Err(PyOSError::new_err(e.to_string())),
+ Err(e) => Err(PyException::new_err(e.to_string())),
  }
 }
 
@@ -39,17 +46,16 @@ pub fn get_ms2_spectra(spectrum_path: String) -> PyResult<Vec<ms2_spectrum::MS2S
  let file_type = match_file_type(&spectrum_path);
 
  let spectra = match file_type {
- SpectrumFileType::MascotGenericFormat | SpectrumFileType::MzML => {
- parse_mzdata::read_ms2_spectra(&spectrum_path, file_type)
+ SpectrumFileType::MascotGenericFormat | SpectrumFileType::MzML | SpectrumFileType::MzMLb | SpectrumFileType:: ThermoRaw => {
+ parse_mzdata::read_ms2_spectra(&spectrum_path)
  }
  SpectrumFileType::BrukerRaw => parse_timsrust::read_ms2_spectra(&spectrum_path),
- // SpectrumFileType::ThermoRaw => parse_with_mzdata_thermo(&spectrum_path, file_type),
- SpectrumFileType::Unknown => return Err(PyOSError::new_err("Unsupported file type")),
+ SpectrumFileType::Unknown => return Err(PyValueError::new_err("Unsupported file type")),
  };
 
  match spectra {
  Ok(spectra) => Ok(spectra),
- Err(e) => Err(PyOSError::new_err(e.to_string())),
+ Err(e) => Err(PyException::new_err(e.to_string())),
  }
 }
 
@@ -59,6 +65,7 @@ pub fn get_ms2_spectra(spectrum_path: String) -> PyResult<Vec<ms2_spectrum::MS2S
 fn ms2rescore_rs(_py: Python, m: &PyModule) -> PyResult<()> {
  m.add_class::<Precursor>()?;
  m.add_class::<MS2Spectrum>()?;
+ m.add_function(wrap_pyfunction!(is_supported_file_type, m)?)?;
  m.add_function(wrap_pyfunction!(get_precursor_info, m)?)?;
  m.add_function(wrap_pyfunction!(get_ms2_spectra, m)?)?;
  Ok(())

src/parse_mzdata.rs

@@ -1,10 +1,8 @@
 use std::collections::HashMap;
-use std::fs::File;
 
-use mzdata::io::{MGFReader, MzMLReader};
 use mzdata::params::ParamValue;
+use mzdata::mz_read;
 
-use crate::file_types::SpectrumFileType;
 use crate::ms2_spectrum::MS2Spectrum;
 use crate::precursor::Precursor;
 
@@ -50,55 +48,26 @@ impl From<mzdata::spectrum::MultiLayerSpectrum> for MS2Spectrum {
 /// Parse precursor info from spectrum files with mzdata
 pub fn parse_precursor_info(
  spectrum_path: &str,
- file_type: SpectrumFileType,
 ) -> Result<HashMap<String, Precursor>, std::io::Error> {
- let file = File::open(spectrum_path)?;
- match file_type {
- SpectrumFileType::MascotGenericFormat => Ok(MGFReader::new(file)
+ mz_read!(spectrum_path.as_ref(), reader => {
+ reader.filter(|spectrum| spectrum.description.ms_level == 2)
  .filter_map(|spectrum| {
  spectrum.description.precursor.as_ref()?;
  Some((spectrum.description.id.clone(), Precursor::from(&spectrum)))
  })
- .collect::<HashMap<String, Precursor>>()),
-
- SpectrumFileType::MzML => Ok(MzMLReader::new(file)
- .filter_map(|spectrum| {
- if spectrum.description.ms_level != 2 {
- return None;
- }
- spectrum.description.precursor.as_ref()?;
- Some((spectrum.description.id.clone(), Precursor::from(&spectrum)))
- })
- .collect::<HashMap<String, Precursor>>()),
-
- _ => Err(std::io::Error::new(
- std::io::ErrorKind::InvalidInput,
- "Unsupported file type for mzdata",
- )),
- }
+ .collect::<HashMap<String, Precursor>>()
+ })
 }
 
 /// Read MS2 spectra from spectrum files with mzdata
 pub fn read_ms2_spectra(
  spectrum_path: &str,
- file_type: SpectrumFileType,
 ) -> Result<Vec<MS2Spectrum>, std::io::Error> {
- let file = File::open(spectrum_path)?;
- match file_type {
- SpectrumFileType::MascotGenericFormat => Ok(MGFReader::new(file)
- .map(MS2Spectrum::from)
- .collect::<Vec<MS2Spectrum>>()),
-
- SpectrumFileType::MzML => Ok(MzMLReader::new(file)
- .filter(|spectrum| spectrum.description.ms_level == 2)
+ mz_read!(spectrum_path.as_ref(), reader => {
+ reader.filter(|spectrum| spectrum.description.ms_level == 2)
  .map(MS2Spectrum::from)
- .collect::<Vec<MS2Spectrum>>()),
-
- _ => Err(std::io::Error::new(
- std::io::ErrorKind::InvalidInput,
- "Unsupported file type for mzdata",
- )),
- }
+ .collect::<Vec<MS2Spectrum>>()
+ })
 }
 
 // pub fn parse_precursor_info_thermo(