wip

.github/workflows/readthedocs.yaml

@@ -0,0 +1,20 @@
+name: readthedocs/actions
+on:
+  pull_request_target:
+    types:
+      - opened
+    # Execute this action only on PRs that touch
+    # documentation files.
+    paths:
+      - "docs/**"
+
+permissions:
+  pull-requests: write
+
+jobs:
+  documentation-links:
+    runs-on: ubuntu-24.04
+    steps:
+      - uses: readthedocs/actions/preview@v1
+        with:
+          project-slug: "fgpyo"

build.py

@@ -2,55 +2,87 @@
 import os
 import glob
 import shutil
-from distutils.command.build_ext import build_ext
-from distutils.core import Distribution, Extension
+from setuptools import Extension, Distribution
+from typing import List
 
 from Cython.Build import cythonize
+from Cython.Distutils.build_ext import new_build_ext as cython_build_ext
+import multiprocessing
+from pathlib import Path
 
-
+SOURCE_DIR = Path("bwapy")
+BUILD_DIR = Path("cython_build")
 compile_args = []
 link_args = []
-include_dirs = ['.']
+# Relative to project root directory
+include_dirs = ["bwa", "bwapy"]
 libraries = ['m', 'z', 'pthread']
-library_dirs=['.']
-extra_objects = glob.glob(os.path.join('bwa', '*.o'))
+library_dirs=['bwa']
+extra_objects = []  #glob.glob(os.path.join('bwa', '*.o'))
 
+c_files = [str(x) for x in Path("bwa").rglob("*.c") if x.name not in ['example.c', 'main.c']]
+extension_module = Extension(
+    name='bwapy.libbwapy',
+    sources=['bwapy/libbwapy.pyx'] + c_files,
+    depends=['bwa/bwtaln.h'],
+    extra_compile_args=compile_args,
+    extra_link_args=link_args,
+    extra_objects=extra_objects,
+    include_dirs=include_dirs,
+    language='c',
+    libraries=libraries,
+    library_dirs=library_dirs,
+)
 
-def build():
-    extensions = [
-        Extension(
-            'bwapy',
-            sources=['bwapy/bwapy.pyx'],
-            depends=['bwa/bwtaln.h'],
-            extra_compile_args=compile_args,
-            extra_link_args=link_args,
-            extra_objects=extra_objects,
-            include_dirs=include_dirs,
-            language='c',
-            libraries=libraries,
-            library_dirs=library_dirs,
-        )
-    ]
-    ext_modules = cythonize(
-        extensions,
-        include_path=include_dirs,
-        compiler_directives={"binding": True, "language_level": 3},
+
+
+def cythonize_helper(extension_modules: List[Extension]) -> List[Extension]:
+    """Cythonize all Python extensions"""
+
+    return cythonize(
+        module_list=extension_modules,
+
+        # Don't build in source tree (this leaves behind .c files)
+        build_dir=BUILD_DIR,
+
+        # Don't generate an .html output file. Would contain source.
+        annotate=False,
+
+        # Parallelize our build
+        nthreads=multiprocessing.cpu_count() * 2,
+
+        # Tell Cython we're using Python 3. Becomes default in Cython 3
+        compiler_directives={"language_level": "3"},
+
+        # (Optional) Always rebuild, even if files untouched
+        force=True,
     )
 
-    distribution = Distribution({"name": "extended", "ext_modules": ext_modules})
-    distribution.package_dir = "extended"
 
-    cmd = build_ext(distribution)
-    cmd.ensure_finalized()
-    cmd.run()
 
-    # Copy built extensions back to the project
-    for output in cmd.get_outputs():
-        relative_extension = os.path.relpath(output, cmd.build_lib)
-        shutil.copyfile(output, relative_extension)
-        mode = os.stat(relative_extension).st_mode
-        mode |= (mode & 0o444) >> 2
-        os.chmod(relative_extension, mode)
+def build():
+    # Collect and cythonize all files
+    extension_modules = cythonize_helper([extension_module])
+
+    # Use Setuptools to collect files
+    distribution = Distribution({
+        "ext_modules": extension_modules,
+        "cmdclass": {
+            "build_ext": cython_build_ext,
+        },
+    })
+
+    # Grab the build_ext command and copy all files back to source dir.
+    # Done so Poetry grabs the files during the next step in its build.
+    build_ext_cmd = distribution.get_command_obj("build_ext")
+    build_ext_cmd.ensure_finalized()
+    # Set the value to 1 for "inplace", with the goal to build extensions
+    # in build directory, and then copy all files back to the source dir
+    # (under the hood, "copy_extensions_to_source" will be called after
+    # building the extensions). This is done so Poetry grabs the files
+    # during the next step in its build.
+    build_ext_cmd.inplace = 1
+    build_ext_cmd.run()
 
 
 if __name__ == "__main__":

bwapy/__init__.py

@@ -0,0 +1,3 @@
+import bwapy.libbwapy as libbwapy
+
+__all__ = (libbwapy.__all__)

bwapy/bwapy.pyx → bwapy/libbwapy.pyx

@@ -7,21 +7,29 @@ from libc.stdint cimport uint8_t
 from pysam import FastqProxy
 cimport cython
 
-cdef extern from "bwa/bntseq.h":
+
+__all__ = [
+    "BwaIndex",
+    "BwaOptions",
+    "Bwa",
+]
+
+
+cdef extern from "bntseq.h":
     unsigned char nst_nt4_table[256]
 
-cdef extern from "bwa/bwa.h":
+cdef extern from "bwa.h":
     char * bwa_idx_infer_prefix(const char * hint)
 
-cdef extern from "bwa/bwt.h":
+cdef extern from "bwt.h":
     ctypedef struct bwt_t:
         pass
 
     bwt_t *bwt_restore_bwt(const char *fn);
     void bwt_destroy(bwt_t *bwt);
 
 
-cdef extern from "bwa/bwtaln.h":
+cdef extern from "bwtaln.h":
     ctypedef struct gap_opt_t:
         int trim_qual
         int s_mm
@@ -50,7 +58,7 @@ cdef extern from "bwa/bwtaln.h":
 
 
 
-cdef extern from "bwa/bwtaln.h":
+cdef extern from "bwtaln.h":
     ctypedef struct bwa_seq_t:
         int tid
         int len
@@ -124,42 +132,44 @@ cdef class BwaIndex:
         self._bwt = NULL
 
 
-#cdef class Bwa:
-#
-#    def __init__(self, prefix: str):
-#        self._index = BwaIndex(prefix=prefix)
-#
-#
-#    def align(self, query: *FastqProxy) -> None:
-#        self._calign(query)
-#
-#    cdef _calign(self, query):
-#        cdef bwa_seq_t* seqs
-#        cdef bwa_seq_t* s
-#        cdef char* s_char
-#        num_seqs = len(query)
-#        seqs = <bwa_seq_t*>calloc(sizeof(bwa_seq_t), num_seqs)
-#        for i in range(num_seqs):
-#            q = query[i]
-#            s = &seqs[i]
-#            s.tid = -1
-#            seq_len = len(q.sequence)
-#            s.len = seq_len
-#            s.clip_len = seq_len
-#            s.seq = <uint8_t*>calloc(sizeof(uint8_t), seq_len)
-#            for i in range(len
-#            for base in q.sequence:
-#                base = nst_nt4_table[ord(base)]
-#                s.seq[i] = base
-#            #if q.qual is not None:
-#            #    # TODO
-#            #    pass
-#            # TODO: reverse the seq
-#            #s.name = <char*>calloc(sizeof(char), len(q.name))
-#            #strncpy(s.name, force_bytes(q.name), len(q.name))
-#
-#        #bwa_free_read_seq(seqs)
-#        #bwa_cal_sa_reg_gap
-#
-#
-##void bwa_aln_core(const char *prefix, const char *fn_fa, const gap_opt_t *opt)
+cdef class Bwa:
+
+    cdef BwaIndex _index
+
+    def __init__(self, prefix: str):
+        self._index = BwaIndex(prefix=prefix)
+
+
+    def align(self, query: *FastqProxy) -> None:
+        self._calign(query)
+
+    cdef _calign(self, query):
+        cdef bwa_seq_t* seqs
+        cdef bwa_seq_t* s
+        cdef char* s_char
+        num_seqs = len(query)
+        seqs = <bwa_seq_t*>calloc(sizeof(bwa_seq_t), num_seqs)
+        for i in range(num_seqs):
+            q = query[i]
+            s = &seqs[i]
+            s.tid = -1
+            seq_len = len(q.sequence)
+            s.len = seq_len
+            s.clip_len = seq_len
+            s.seq = <uint8_t*>calloc(sizeof(uint8_t), seq_len)
+            #for i in range(len
+            for base in q.sequence:
+                base = nst_nt4_table[ord(base)]
+                s.seq[i] = base
+            #if q.qual is not None:
+            #    # TODO
+            #    pass
+            # TODO: reverse the seq
+            #s.name = <char*>calloc(sizeof(char), len(q.name))
+            #strncpy(s.name, force_bytes(q.name), len(q.name))
+
+        #bwa_free_read_seq(seqs)
+        #bwa_cal_sa_reg_gap
+
+
+#void bwa_aln_core(const char *prefix, const char *fn_fa, const gap_opt_t *opt)

pyproject.toml

@@ -8,6 +8,7 @@ readme = "README.md"
 homepage = "https://github.com/fulcrumgenomics/bwapy"
 repository = "https://github.com/fulcrumgenomics/bwapy"
 keywords = ["bioinformatics"]
+packages = [{ include = "bwapy" }]
 classifiers = [
 	"Development Status :: 3 - Alpha",
 	"Environment :: Console",
@@ -20,7 +21,13 @@ classifiers = [
 	"Topic :: Software Development :: Documentation",
 	"Topic :: Software Development :: Libraries :: Python Modules",
 ]
-include = ["LICENSE"]
+include = [
+	"LICENSE",
+	{ path = "bwapy/**/*.so", format = "wheel" },
+	{ path = "**/*.so", format = "wheel" },
+	{ path = "*.so", format = "wheel" },
+	#{ path = "*.so", format = "wheel" }
+]
 
 [tool.poetry.build]
 generate-setup-file = false
@@ -53,7 +60,7 @@ mkdocstrings-python = { version = ">=1.6.2" }
 mkdocstrings = { version = ">=0.23.0" }
 
 [build-system]
-requires = ["poetry-core>=1.0.0", "cython>=3.0.11"]
+requires = ["poetry-core>=1.0.0", "cython>=3.0.11", "setuptools>=75.1.0"]
 build-backend = "poetry.core.masonry.api"
 
 [tool.ruff]