New mzio version based on initial commit of mzcore

.github/workflows/python-builds.yml

@@ -11,48 +11,76 @@ jobs:
   linux:
     runs-on: ubuntu-latest
     steps:
-    - uses: actions/checkout@v3
+    - name: Checkout current repository
+      uses: actions/checkout@v3
+
+    - name: Clone remote repository
+      run: |
+        git clone https://github.com/david-bouyssie/mzcore.git ../mzcore
+        ls /home/runner/work/mzio/mzcore
+
     - uses: PyO3/maturin-action@v1
       with:
-        manylinux: auto
+        manylinux: off # auto
         command: build
-        args: --release --sdist -o dist --find-interpreter
+        args: -o dist --strip --find-interpreter
         working-directory: mzio-py
+
     - name: Upload wheels to artifact
       uses: actions/upload-artifact@v3
       with:
         name: wheels
-        path: dist
+        path: mzio-py/dist
+
+    - name: Setup tmate session
+      if: ${{ failure() }}
+      uses: mxschmitt/action-tmate@v3
 
   windows:
     runs-on: windows-latest
     steps:
-    - uses: actions/checkout@v3
+    - name: Checkout current repository
+      uses: actions/checkout@v3
+
+    - name: Clone remote repository
+      run: |
+        git clone https://github.com/david-bouyssie/mzcore.git ../mzcore
+        ls ../mzcore
+
     - uses: PyO3/maturin-action@v1
       with:
         command: build
-        args: --release -o dist --find-interpreter
+        args: -o dist --strip --find-interpreter
         working-directory: mzio-py
-    - name: Upload wheels
+
+    - name: Upload wheels to artifact
       uses: actions/upload-artifact@v3
       with:
         name: wheels
-        path: dist
+        path: mzio-py/dist
 
   macos:
     runs-on: macos-latest
     steps:
-    - uses: actions/checkout@v3
+    - name: Checkout current repository
+      uses: actions/checkout@v3
+
+    - name: Clone remote repository
+      run: |
+        git clone https://github.com/david-bouyssie/mzcore.git ../mzcore
+        ls ../mzcore
+
     - uses: PyO3/maturin-action@v1
       with:
         command: build
-        args: --release -o dist --universal2 --find-interpreter
+        args: -o dist --strip --find-interpreter --target universal2-apple-darwin
         working-directory: mzio-py
-    - name: Upload wheels
+
+    - name: Upload wheels to artifact
       uses: actions/upload-artifact@v3
       with:
         name: wheels
-        path: dist
+        path: mzio-py/dist
 
   # release:
   #   name: Release

Cargo.toml

@@ -0,0 +1,6 @@
+[workspace]
+
+members = [
+    "mzio-rs",
+    "mzio-py"
+]

mzio-py/Cargo.toml

@@ -10,7 +10,7 @@ name = "mzio_py"
 crate-type = ["cdylib"]
 
 [dependencies]
-anyhow = "1.0.68"
+anyhow = "1.0.75"
 fallible-iterator = "0.2.0"
-mzio = { path = "../mzio-rs" }
-pyo3 = { version = "0.17.3", features = ["extension-module", "anyhow"] }
+pyo3 = { version = "0.20.0", features = ["extension-module", "anyhow"] }
+mzio = { path = "../mzio-rs", features = ["fasta", "mgf"] }

mzio-py/pyproject.toml

@@ -35,5 +35,5 @@ doc = [
 ]
 
 [build-system]
-requires = ["maturin>=0.14,<0.15"]
+requires = ["maturin>=1.0.0,<1.3.0"]
 build-backend = "maturin"

mzio-py/src/fasta/entry.rs

@@ -2,8 +2,9 @@
 use std::collections::HashMap;
 
 // 3rd party imports
+use anyhow::Result;
 use pyo3::prelude::*;
-use mzio::fasta::entry::Entry as BaseEntry;
+use mzio::fasta::entry::FastaEntry as BaseEntry;
 
 /// Wrapper for the rust implementation entry
 /// 
@@ -35,37 +36,38 @@ impl Entry {
                 entry_name,
                 protein_name,
                 keyword_attributes,
-                sequence
+                sequence,
+                None
             )
         }
     }
 
     /// Returns the database type
     ///
     #[getter]
-    pub fn database(&self) -> PyResult<&str> {
-        Ok(&self.base_entry.get_database())
+    pub fn database(&self) -> Result<&String> {
+        Ok(self.base_entry.get_database())
     }
 
     /// Returns the accession
     ///
     #[getter]
-    pub fn accession(&self) -> PyResult<&str> {
-        Ok(&self.base_entry.get_accession())
+    pub fn accession(&self) -> Result<&String> {
+        Ok(self.base_entry.get_accession())
     }
 
     /// Entry name
     ///
     #[getter]
-    pub fn entry_name(&self) -> PyResult<&str> {
-        Ok(&self.base_entry.get_entry_name())
+    pub fn entry_name(&self) -> Result<&String> {
+        Ok(self.base_entry.get_entry_name())
     }
 
     /// Returns the protein name
     ///
     #[getter]
-    pub fn protein_name(&self) -> PyResult<&str> {
-        Ok(&self.base_entry.get_protein_name())
+    pub fn protein_name(&self) -> Result<&String> {
+        Ok(self.base_entry.get_protein_name())
     }
 
     /// Returns additional keyword attributes, e.g
@@ -74,16 +76,23 @@ impl Entry {
     ///
     // !!! TODO Reference to HasMap is no convertible to PyResult by default.
     #[getter]
-    pub fn keyword_attributes(&self) -> PyResult<HashMap<String, String>> {
+    pub fn keyword_attributes(&self) -> Result<HashMap<String, String>> {
         // TODO: avoid clone?
         Ok(self.base_entry.get_keyword_attributes().clone())
     }
 
     /// Returns the amino acid sequence
     ///
     #[getter]
-    pub fn sequence(&self) -> PyResult<&str> {
-        Ok(&self.base_entry.get_sequence())
+    pub fn sequence(&self) -> Result<&String> {
+        Ok(self.base_entry.get_sequence())
+    }
+
+    /// Returns the plain header (before parsing)
+    ///
+    #[getter]
+    pub fn plain_header(&self) -> Result<&String> {
+        Ok(self.base_entry.get_sequence())
     }
 }
 

mzio-py/src/fasta/reader.rs

@@ -1,10 +1,10 @@
 // std imports
 use std::path::PathBuf;
 
-// 3rd party modules
-use pyo3::prelude::*;
+// 3rd party imports
 use anyhow::Result;
-use mzio::fasta::reader::Reader as BaseReader;
+use pyo3::prelude::*;
+use mzio::fasta::reader::FastaReader as BaseReader;
 
 // internal imports
 use crate::fasta::entry::Entry;
@@ -17,8 +17,9 @@ pub struct Reader {
 #[pymethods]
 impl Reader {
     #[new]
-    fn new(fasta_file_path: PathBuf, buffer_size: usize) -> Result<Self> {
-        match BaseReader::new(&fasta_file_path, buffer_size) {
+    fn new(fasta_file_path: PathBuf, buffer_size: usize, keep_plain_header: bool) -> Result<Self> {
+        match BaseReader::new(&fasta_file_path, buffer_size, keep_plain_header
+        ) {
             Ok(base_reader) => Ok(Self{base_reader}),
             Err(err) => Err(err)
         }

mzio-py/src/fasta/writer.rs

@@ -1,15 +1,14 @@
 // std imports
 use std::path::PathBuf;
 
-// 3rd party modules
+// 3rd party imports
 use anyhow::Result;
 use pyo3::prelude::*;
-use mzio::fasta::writer::Writer as BaseWriter;
+use mzio::fasta::writer::FastaWriter as BaseWriter;
 
 // internal imports
 use crate::fasta::entry::Entry;
 
-
 #[pyclass]
 pub struct Writer {
     base_writer: BaseWriter
@@ -24,15 +23,15 @@ impl Writer {
     /// * `fasta_file_path` - Path to FASTA file
     /// 
     #[new]
-    pub fn new(fasta_file_path: PathBuf) -> PyResult<Self> {
-        match BaseWriter::new(&fasta_file_path) {
+    pub fn new(fasta_file_path: PathBuf, sort_keyword_attributes: bool) -> PyResult<Self> {
+        match BaseWriter::new_with_default_seq_formatting(&fasta_file_path, sort_keyword_attributes) {
             Ok(base_writer) => Ok(Self{base_writer}),
             Err(err) => Err(err.into())
         }
     }
 
-    pub fn write_entry(&mut self, entry: &Entry, sort_keyword_attributes: bool) -> Result<usize> {
-        match self.base_writer.write_entry(entry.into(), sort_keyword_attributes) {
+    pub fn write_entry(&mut self, entry: &Entry) -> Result<usize> {
+        match self.base_writer.write_entry(entry.into()) {
             Ok(written_bytes) => Ok(written_bytes),
             Err(err) => Err(err)
         }

mzio-py/src/lib.rs

@@ -34,9 +34,9 @@ fn register_fasta_module(py: Python, parent_module: &PyModule) -> PyResult<()> {
 /// `parent_module` - Parent module of the mgf module
 fn register_mgf_module(py: Python, parent_module: &PyModule) -> PyResult<()> {
     let child_module = PyModule::new(py, "mgf")?;
-    child_module.add_class::<mgf::spectrum::Spectrum>()?;
-    child_module.add_class::<mgf::reader::Reader>()?;
-    child_module.add_class::<mgf::writer::Writer>()?;
+    child_module.add_class::<mgf::spectrum::MgfSpectrum>()?;
+    child_module.add_class::<mgf::reader::MgfReader>()?;
+    child_module.add_class::<mgf::writer::MgfWriter>()?;
     parent_module.add_submodule(child_module)?;
     Ok(())
 }

mzio-py/src/mgf/reader.rs

@@ -1,39 +1,40 @@
 // std imports
 use std::path::PathBuf;
 
-// 3rd party modules
+// 3rd party imports
 use pyo3::prelude::*;
 use anyhow::Result;
-use mzio::mgf::reader::Reader as BaseReader;
-
 use fallible_iterator::FallibleIterator;
+use mzio::mgf::reader::MgfReader as BaseMgfReader;
 
 // internal imports
-use crate::mgf::spectrum::Spectrum;
+use crate::mgf::spectrum::MgfSpectrum;
 
 #[pyclass]
-pub struct Reader {
-    base_reader: BaseReader
+pub struct MgfReader {
+    base_reader: BaseMgfReader
 }
 
 #[pymethods]
-impl Reader {
+impl MgfReader {
     #[new]
-    #[args(buffer_size=4096)]
+    #[pyo3(signature = (mgf_file_path, buffer_size=4096))]
     fn new(mgf_file_path: PathBuf, buffer_size: usize) -> Result<Self> {
-        match BaseReader::new(&mgf_file_path, buffer_size) {
+        let base_reader = BaseMgfReader::new(&mgf_file_path, buffer_size)?;
+        Ok(Self{base_reader})
+        /*match BaseReader::new(&mgf_file_path, buffer_size) {
             Ok(base_reader) => Ok(Self{base_reader}),
             Err(err) => Err(err)
-        }
+        }*/
     }
 
     fn __iter__(slf: PyRef<'_, Self>) -> PyRef<'_, Self> {
         slf
     }
 
-    fn __next__(mut slf: PyRefMut<'_, Self>) -> Option<Spectrum> {
+    fn __next__(mut slf: PyRefMut<'_, Self>) -> Option<MgfSpectrum> {
         match slf.base_reader.next().ok()? {
-            Some(base_spectrum) => Some(Spectrum::from(base_spectrum)),
+            Some(base_spectrum) => Some(MgfSpectrum::from(base_spectrum)),
             None => None
         }
     }