Merge pull request #2 from nleroy917/main

New updates

bindings/.github/workflows/CI.yml → .github/workflows/CI.yml

@@ -6,13 +6,8 @@
 name: CI
 
 on:
-  push:
-    branches:
-      - main
-      - master
-    tags:
-      - '*'
-  pull_request:
+  release:
+    types: [created]
   workflow_dispatch:
 
 permissions:
@@ -23,7 +18,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        target: [x86_64, x86, aarch64, armv7, s390x, ppc64le]
+        target: [x86_64, x86, aarch64]
     steps:
       - uses: actions/checkout@v3
       - uses: actions/setup-python@v4
@@ -36,11 +31,12 @@ jobs:
           args: --release --out dist --find-interpreter
           sccache: 'true'
           manylinux: auto
+          working-directory: ./bindings
       - name: Upload wheels
         uses: actions/upload-artifact@v3
         with:
           name: wheels
-          path: dist
+          path: ./bindings/dist
 
   windows:
     runs-on: windows-latest
@@ -59,11 +55,12 @@ jobs:
           target: ${{ matrix.target }}
           args: --release --out dist --find-interpreter
           sccache: 'true'
+          working-directory: ./bindings
       - name: Upload wheels
         uses: actions/upload-artifact@v3
         with:
           name: wheels
-          path: dist
+          path: ./bindings/dist
 
   macos:
     runs-on: macos-latest
@@ -81,11 +78,12 @@ jobs:
           target: ${{ matrix.target }}
           args: --release --out dist --find-interpreter
           sccache: 'true'
+          working-directory: ./bindings
       - name: Upload wheels
         uses: actions/upload-artifact@v3
         with:
           name: wheels
-          path: dist
+          path: ./bindings/dist
 
   sdist:
     runs-on: ubuntu-latest
@@ -96,25 +94,28 @@ jobs:
         with:
           command: sdist
           args: --out dist
+          working-directory: ./bindings
       - name: Upload sdist
         uses: actions/upload-artifact@v3
         with:
           name: wheels
-          path: dist
+          path: ./bindings/dist
 
   release:
     name: Release
     runs-on: ubuntu-latest
+    environment: release
+    permissions:
+      id-token: write
     if: "startsWith(github.ref, 'refs/tags/')"
     needs: [linux, windows, macos, sdist]
     steps:
       - uses: actions/download-artifact@v3
         with:
           name: wheels
+          path: ./bindings/dist
       - name: Publish to PyPI
         uses: PyO3/maturin-action@v1
-        env:
-          MATURIN_PYPI_TOKEN: ${{ secrets.PYPI_API_TOKEN }}
         with:
           command: upload
-          args: --non-interactive --skip-existing *
+          args: --non-interactive --skip-existing ./bindings/dist/*

NOTICE

@@ -3,3 +3,4 @@ Copyright [2023] EMBL-European Bioinformatics Institute
 
 This product includes software developed at:
 - EMBL-European Bioinformatics Institute
+- University of Virginia, Center for Public Health Genomics

bindings/Cargo.toml

@@ -1,5 +1,5 @@
 [package]
-name = "gc_count"
+name = "rust_gc_count_py"
 version = "0.1.0"
 edition = "2021"
 

bindings/gc_count.pyi

@@ -0,0 +1,57 @@
+from typing import List
+
+class ChecksumResult:
+    def __init__(self, id: str, length: int, sha512: str, md5: str):
+        self.id = id
+        self.length = length
+        self.sha512 = sha512
+        self.md5 = md5
+
+    def __repr__(self):
+        return f"ChecksumResult(id={self.id}, length={self.length}, sha512={self.sha512}, md5={self.md5})"
+
+    def __str__(self):
+        return f"ChecksumResult(id={self.id}, length={self.length}, sha512={self.sha512}, md5={self.md5})"
+
+def checksum(file: str, verbose: bool) -> List[ChecksumResult]:
+    """
+    Calculate the sequence lengths and checksums from a fasta file. It will
+    produce a list of ChecksumResult objects, each containing the following
+    
+    Sequence ID as it appears in the FASTA file
+    Sequence length
+    Refget ga4gh identifier (SQ.sha512t24u)
+    MD5 checksum hex encoded
+
+    :param file: The file to checksum
+    :param verbose: Whether to print out the progress
+    """
+
+def write_gc_count_to_file(
+    input: str,
+    output: str,
+    compression_level: int,
+    window_size: int,
+    omit_tail: bool,
+    chrom_sizes_path: str,
+    write_chrom_sizes: bool,
+    verbose: bool,
+) -> None:
+    """
+    Calculate the GC content and write it to a file. The file will be a
+    tab-separated file with the following columns:
+    
+    Chromosome name
+    Start position
+    End position
+    GC content
+
+    :param input: The input file to calculate the GC content from
+    :param output: The output file to write the GC content to
+    :param compression_level: The compression level to use for the output file
+    :param window_size: The window size to use for calculating the GC content
+    :param omit_tail: Whether to omit the tail of the sequence
+    :param chrom_sizes_path: The path to the chromosome sizes file
+    :param write_chrom_sizes: Whether to write the chromosome sizes to the output file
+    :param verbose: Whether to print out the progress
+    """       

bindings/src/checksumseq.rs

@@ -14,14 +14,9 @@ pub fn checksum(file: String, verbose: Option<bool>) -> Vec<PyChecksumResult> {
 
     while let Some(record) = reader.next() {
         let record = record.unwrap();
-        let (id, length, sha512, md5) = process_sequence(record, verbose);
-
-        results.push(PyChecksumResult {
-            id,
-            length,
-            sha512,
-            md5
-        });
+        let result = process_sequence(record, verbose);
+
+        results.push(PyChecksumResult::from(result));
     }
 
     results

bindings/src/gc_count_utils.rs

@@ -0,0 +1,21 @@
+use pyo3::prelude::*;
+use std::path::PathBuf;
+
+#[pyfunction]
+pub fn write_gc_count_to_file(
+    input: String,
+    output: String,
+    compression_level: u32,
+    window_size: i32,
+    omit_tail: bool,
+    chrom_sizes_path: String,
+    write_chrom_sizes: bool,
+    verbose: bool,
+) {
+
+    let input = PathBuf::from(input);
+    let output = PathBuf::from(output);
+    let chrom_sizes_path = PathBuf::from(chrom_sizes_path);
+
+    rust_gc_count::gc_count::write_gc_to_file(input, output, compression_level, window_size, omit_tail, chrom_sizes_path, write_chrom_sizes, verbose)
+}

bindings/src/lib.rs

@@ -1,13 +1,16 @@
 use pyo3::prelude::*;
 
+pub mod gc_count_utils;
 pub mod checksumseq;
 pub mod models;
 
 use crate::checksumseq::checksum;
+use crate::gc_count_utils::write_gc_count_to_file;
 
 /// A Python module implemented in Rust.
 #[pymodule]
 fn gc_count(_py: Python, m: &PyModule) -> PyResult<()> {
     m.add_function(wrap_pyfunction!(checksum, m)?)?;
+    m.add_function(wrap_pyfunction!(write_gc_count_to_file, m)?)?;
     Ok(())
 }

bindings/src/models.rs

@@ -1,4 +1,5 @@
 use pyo3::prelude::*;
+use rust_gc_count::checksum::ChecksumResult;
 
 #[pyclass]
 #[pyo3(name="ChecksumResult")]
@@ -11,4 +12,22 @@ pub struct PyChecksumResult {
     pub sha512: String,
     #[pyo3(get,set)]
     pub md5: String
+}
+
+#[pymethods]
+impl PyChecksumResult {
+    fn __repr__(&self) -> String {
+        format!("<ChecksumResult for {}>", self.id)
+    }
+}
+
+impl From<ChecksumResult> for PyChecksumResult {
+    fn from(value: ChecksumResult) -> Self {
+        PyChecksumResult {
+            id: value.id,
+            length: value.length,
+            sha512: value.sha512,
+            md5: value.md5
+        }
+    }
 }

src/checksumseq.rs

@@ -59,10 +59,13 @@ fn main() {
     while let Some(record) = reader.next() {
         let record = record.expect("Error reading record");
         let result = process_sequence(record, args.verbose);
-        let line = format!(
-            "{0:#}\t{1:#}\tSQ.{2:#}\t{3:#}\n",
-            result.0, result.1, result.2, result.3
-        );
+
+        let id = result.id;
+        let length = result.length;
+        let sha512 = result.sha512;
+        let md5 = result.md5;
+
+        let line = format!("{0:#}\t{1:#}\tSQ.{2:#}\t{3:#}\n", id, length, sha512, md5);
         writer
             .write_all(line.as_bytes())
             .expect("Could not write to file");
@@ -89,9 +92,9 @@ acgT\n
     while let Some(record) = reader.next() {
         let record = record.expect("Error reading record");
         let result = process_sequence(record, false);
-        assert_eq!(result.0, "id");
-        assert_eq!(result.1, 4);
-        assert_eq!(result.2, "aKF498dAxcJAqme6QYQ7EZ07-fiw8Kw2");
-        assert_eq!(result.3, "f1f8f4bf413b16ad135722aa4591043e");
+        assert_eq!(result.id, "id");
+        assert_eq!(result.length, 4);
+        assert_eq!(result.sha512, "aKF498dAxcJAqme6QYQ7EZ07-fiw8Kw2");
+        assert_eq!(result.md5, "f1f8f4bf413b16ad135722aa4591043e");
     }
 }