Merge branch 'master' of https://github.com/lazear/sage into feature/…

…mzMLb

README.md

@@ -57,7 +57,7 @@ Sage is well-integrated into the open-source proteomics ecosystem. The following
 - [SearchGUI](http://compomics.github.io/projects/searchgui): a graphical user interface for running searches
 - [PeptideShaker](http://compomics.github.io/projects/peptide-shaker): visualize peptide-spectrum matches
 - [MS2Rescore](http://compomics.github.io/projects/ms2rescore): AI-assisted rescoring of results
-- [Picked group FDR](github.com/kusterlab/picked_group_fdr): scalable protein group FDR for large-scale experiments
+- [Picked group FDR](https://github.com/kusterlab/picked_group_fdr): scalable protein group FDR for large-scale experiments
 - [sagepy](https://github.com/theGreatHerrLebert/sagepy): Python bindings to the sage-core library
 - [quantms](https://github.com/bigbio/quantms): nextflow pipeline for running searches with Sage
 - [OpenMS](https://github.com/OpenMS/OpenMS): Sage is included as a "TOPP" tool in OpenMS

crates/sage-cli/src/main.rs

@@ -9,7 +9,6 @@ use sage_core::mass::Tolerance;
 use sage_core::scoring::{Feature, Scorer};
 use sage_core::spectrum::{ProcessedSpectrum, SpectrumProcessor};
 use sage_core::tmt::TmtQuant;
-use std::path::PathBuf;
 use std::time::Instant;
 
 mod input;
@@ -192,32 +191,25 @@ impl Runner {
             self.parameters.deisotope,
         );
 
-        let bruker_extensions = ["d", "tdf", "tdf_bin"];
+        let bruker_extensions = [".d", ".tdf", ".tdf_bin"];
         let spectra = chunk
             .par_iter()
             .enumerate()
             .flat_map(|(idx, path)| {
-                let res = match path {
-                    path if bruker_extensions.contains(
-                        &PathBuf::from(path)
-                            .extension()
-                            .unwrap_or_default()
-                            .to_str()
-                            .unwrap_or_default(),
-                    ) =>
-                    {
-                        sage_cloudpath::util::read_tdf(path, chunk_idx * batch_size + idx)
-                    }
-                    #[cfg(feature = "mzmlb")]
-                    path if &PathBuf::from(path)
-                            .extension()
-                            .unwrap_or_default()
-                            .to_str()
-                            .unwrap_or_default().to_ascii_lowercase() == ".mzmlb" => {
-                                sage_cloudpath::util::read_mzmlb(path, chunk_idx * batch_size + idx)
-                            }
-                    _ => sage_cloudpath::util::read_mzml(path, chunk_idx * batch_size + idx, sn),
+                let file_id = chunk_idx * batch_size + idx;
+
+                let path_lower = path.to_lowercase();
+                let res = if path_lower.ends_with(".mgf.gz") || path_lower.ends_with(".mgf") {
+                    sage_cloudpath::util::read_mgf(path_lower, file_id)
+                } else if bruker_extensions
+                    .iter()
+                    .any(|ext| path_lower.ends_with(ext))
+                {
+                    sage_cloudpath::util::read_tdf(path, file_id)
+                } else {
+                    sage_cloudpath::util::read_mzml(path, file_id, sn)
                 };
+
                 match res {
                     Ok(s) => {
                         log::trace!("- {}: read {} spectra", path, s.len());

crates/sage-cloudpath/Cargo.toml

@@ -26,6 +26,7 @@ quick-xml = { version = "0.30.0", features = ["async-tokio"] }
 timsrust = "0.2.0"
 rayon = "1.5"
 reqwest = { version = "0.11", features = ["json", "rustls-tls"], default-features = false }
+regex = "1.6"
 
 serde = { version="1.0", features = ["derive"] }
 serde_json = "1.0"

crates/sage-cloudpath/src/lib.rs

@@ -6,6 +6,7 @@ use std::path::PathBuf;
 use std::str::FromStr;
 use tokio::io::{AsyncBufRead, AsyncRead, AsyncWriteExt, BufReader};
 
+pub mod mgf;
 pub mod mzml;
 pub mod tdf;
 pub mod util;
@@ -284,6 +285,8 @@ pub enum Error {
     MzML(#[from] mzml::MzMLError),
     #[error("TDF error: {0}")]
     TDF(#[from] timsrust::Error),
+    #[error("MGF error: {0}")]
+    MGF(#[from] mgf::MgfError),
 }
 
 #[cfg(test)]