RENAME: argweavers

1. Rename the module to resolve several issues
2. Drop the bedops sort-bed. It sort the BED file in alphabet order on iter number instead of numerical

.github/workflows/documentation.yml

@@ -4,11 +4,11 @@ on:
   push:
     paths:
       - 'docs/**'
-      - 'argweaver/**'
+      - 'argweavers/**'
   pull_request:
     paths:
       - 'docs/**'
-      - 'argweaver/**'
+      - 'argweavers/**'
   workflow_dispatch:
 
 permissions:

.gitignore

@@ -22,7 +22,7 @@ __pycache__/
 /requirements-dev.lock
 
 # maturin
-/argweaver/*.so
+/argweavers/*.so
 
 # test coverage
 /.coverage

README+s.md

@@ -5,8 +5,7 @@
 1. [rustup](https://rustup.rs/)
 2. [CMake](https://cmake.org/) with any modern C++ compiler
 3. [rye](https://rye-up.com) or [maturin](https://maturin.rs) for Python binding
-4. [samtools](http://www.htslib.org/) and
-   [bedops](https://bedops.readthedocs.io/en/latest/) for `smc2bed-all`
+4. [samtools](http://www.htslib.org/) for `smc2bed-all`
 
 ## Install
 

argweavers/__init__.py

@@ -0,0 +1,10 @@
+from importlib.metadata import version
+
+from . import argweavers  # make ruff happy
+from .argweavers import *  # noqa: F403
+
+__doc__ = argweavers.__doc__
+__version__ = version(__package__)
+
+if hasattr(argweavers, "__all__"):
+    __all__ = argweavers.__all__

argweaver/io/bed.py → argweavers/io/bed.py

@@ -3,7 +3,7 @@
 
 import pandas as pd
 
-from argweaver.utils import parse_region
+from argweavers.utils import parse_region
 
 __all__ = ["read_bed"]
 

argweaver/r.py → argweavers/r.py

@@ -5,7 +5,7 @@
 from rpy2.robjects import pandas2ri
 from rpy2.robjects.packages import importr
 
-from argweaver.utils import parse_region
+from argweavers.utils import parse_region
 
 if typing.TYPE_CHECKING:
     from os import PathLike

argweaver/scripts/smc2bed.py → argweavers/scripts/smc2bed.py

@@ -2,7 +2,7 @@
 import sys
 from contextlib import redirect_stdout
 
-from argweaver.s import smc2bed
+from argweavers import smc2bed
 
 
 def main(args=None, capture_output=False):

argweaver/scripts/smc2bed_all.py → argweavers/scripts/smc2bed_all.py

@@ -1,10 +1,13 @@
 import argparse
 import os
 import sys
+from io import BytesIO
 from itertools import count
 
-from argweaver.bin import require_executable
-from argweaver.scripts.smc2bed import main as smc2bed
+import pandas as pd
+
+from argweavers.bin import require_executable
+from argweavers.scripts.smc2bed import main as smc2bed
 
 
 class MetavarFormatter(argparse.HelpFormatter):
@@ -16,7 +19,6 @@ def _get_default_metavar_for_positional(self, action):
 
 
 def main(argv=None):
-    sort_bed = require_executable("sort-bed", "It is included in `bedops`")
     bgzip = require_executable("bgzip", "It is included in `samtools`")
     tabix = require_executable("tabix", "It is included in `samtools`")
     parser = argparse.ArgumentParser(
@@ -93,14 +95,16 @@ def main(argv=None):
             sys.stderr.write(f"ended at sample={num}\n")
             break
         sys.stderr.write(f"{num} {file}\n")
-        print(os.stat(file).st_size)
         out += smc2bed(
             ["--sample", str(num), *regionarg, file],
             capture_output=True,
         )
-        print(len(out))
         num += interval
-    out = sort_bed(["-"], input=out, capture_output=True).stdout
+    bed = pd.read_table(
+        BytesIO(out), header=None, names=["chrom", "start", "end", "iter", "tree"]
+    )
+    bed = bed.sort_values(["chrom", "start", "end", "iter"])
+    out = bed.to_csv(index=False, header=False, sep="\t").encode("utf-8")
     with open(f"{baseout}.bed.gz", "wb") as f:
         bgzip(["-"], input=out, stdout=f)
     tabix(["-p", "bed", f"{baseout}.bed.gz"])

docs/conf.py

@@ -21,10 +21,10 @@
 ]
 
 autodoc2_module_all_regexes = [
-    r"argweaver\..*",
+    r"argweavers\..*",
 ]
 autodoc2_packages = [
-    "../argweaver",
+    "../argweavers",
 ]
 autodoc2_render_plugin = "myst"
 graphviz_output_format = "svg"

docs/installation.md

@@ -3,19 +3,18 @@
 ## Pre-requisites
 
 1. [rustup](https://rustup.rs/)
-2. (Optional) [samtools](http://www.htslib.org/) and
-   [bedops](https://bedops.readthedocs.io/en/latest/) for executing `smc2bed-all`
+2. (Optional) [samtools](http://www.htslib.org/) for executing `smc2bed-all`
 3. (Optional) R for plotting
 
 ```bash
 # Install rustup
 curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
-# If you are using macOS, you can install samtools and bedops using homebrew
-brew install samtools bedops
-# Otherwise, install miniconda for samtools and bedops
+# If you are using macOS, you can install samtools using homebrew
+brew install samtools
+# Otherwise, install miniconda for samtools
 wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
 bash Miniconda3-latest-Linux-x86_64.sh
-conda install -c bioconda samtools bedops
+conda install -c bioconda samtools
 # Install R
 sudo apt-get install r-base
 # Or

pyproject.toml

@@ -38,11 +38,11 @@ dependencies = [
 ]
 
 [project.scripts]
-arg-likelihood = "argweaver.bin:arg_likelihood"
-arg-sample = "argweaver.bin:arg_sample"
-arg-summarize = "argweaver.bin:arg_summarize"
-smc2bed = "argweaver.scripts.smc2bed:main"
-smc2bed-all = "argweaver.scripts.smc2bed_all:main"
+arg-likelihood = "argweavers.bin:arg_likelihood"
+arg-sample = "argweavers.bin:arg_sample"
+arg-summarize = "argweavers.bin:arg_summarize"
+smc2bed = "argweavers.scripts.smc2bed:main"
+smc2bed-all = "argweavers.scripts.smc2bed_all:main"
 
 [tool.rye]
 dev-dependencies = [
@@ -55,6 +55,7 @@ dev-dependencies = [
     "myst-parser>=2.0.0",
     "sphinx-autodoc2>=0.5.0",
     "ipykernel>=6.29.2",
+    "cibuildwheel>=2.16.5",
 ]
 [tool.rye.scripts]
 cov-report = { chain = ["htmlcov", "htmlcov-serve"] }
@@ -63,7 +64,6 @@ htmlcov-serve = "python -m http.server --directory htmlcov"
 
 [tool.maturin]
 features = ["extension-module"]
-module-name = "argweaver.s"
 include = ["bin/*"]
 
 [[tool.mypy.overrides]]
@@ -78,11 +78,14 @@ target-version = "py38"
 
 [tool.pytest.ini_options]
 minversion = "7.0"
-addopts = ["--cov=argweaver"]
+addopts = ["--cov=argweavers"]
 
 [tool.coverage.report]
 exclude_also = [
     "if (typing\\.)?TYPE_CHECKING",
     "if __name__ == \"__main__\"",
     "class \\w+\\((argparse\\.)?HelpFormatter\\)",
 ]
+
+[tool.cibuildwheel]
+skip = "pp*"

src/lib.rs

@@ -54,7 +54,7 @@ fn smc2bed(args: Option<Vec<String>>) -> PyResult<()> {
 #[cfg(feature = "extension-module")]
 /// Ancestral recombination graph sampling method
 #[pymodule]
-fn s(_py: Python, m: &PyModule) -> PyResult<()> {
+fn argweavers(_py: Python, m: &PyModule) -> PyResult<()> {
     m.add_class::<sites::Sites>()?;
     m.add_function(wrap_pyfunction!(sites::read_sites, m)?)?;
     m.add_function(wrap_pyfunction!(smc2bed, m)?)?;

tests/conftest.py

@@ -3,7 +3,7 @@
 
 import pytest
 
-from argweaver.scripts.smc2bed_all import main as smc2bed_all
+from argweavers.scripts.smc2bed_all import main as smc2bed_all
 
 examples_dir = pathlib.Path(__file__).parent.parent / "examples"
 

tests/test_basic.py

@@ -1,5 +1,5 @@
-import argweaver
+import argweavers
 
 
 def test_basic():
-    assert argweaver.__version__ == "0.9.0"
+    assert argweavers.__version__ == "0.9.0"

tests/test_bin.py

@@ -1,6 +1,6 @@
 import pytest
 
-from argweaver.bin import require_executable
+from argweavers.bin import require_executable
 
 
 def test_require_executable():

tests/test_io.py

@@ -1,4 +1,4 @@
-from argweaver.io import read_bed, subset_bed
+from argweavers.io import read_bed, subset_bed
 
 
 def test_read_bed(bedfile):

tests/test_plot.py

@@ -1,7 +1,7 @@
 import matplotlib.pyplot as plt
 from matplotlib.testing.decorators import image_comparison
 
-from argweaver.plot import plot_tree
+from argweavers.plot import plot_tree
 
 
 @image_comparison(baseline_images=["test_plot_tree"], extensions=["png"])

tests/test_r.py

@@ -1,6 +1,6 @@
 import pytest
 
-from argweaver.r import plotTreesFromBed
+from argweavers.r import plotTreesFromBed
 
 
 def test_plot_trees(bedfile):

tests/test_scripts.py

@@ -1,6 +1,6 @@
 import hashlib
 
-from argweaver.scripts.smc2bed import main as smc2bed
+from argweavers.scripts.smc2bed import main as smc2bed
 
 
 def sha256sum(path):
@@ -18,5 +18,5 @@ def test_smc2bed(sim1_sample):
 
 
 def test_smc2bed_all(bedfile):
-    sha256 = "577951a303869a2ab3aaf99e500ff29ce5613c27ac1a7cbef7dbfa63524136c3"
+    sha256 = "94012cc3ecadddaf19f823ec361a72ac9b8073fe33dc75c1f731498c789d7023"
     assert sha256sum(bedfile) == sha256

tests/test_sites.py

@@ -1,4 +1,4 @@
-from argweaver import read_sites
+from argweavers import read_sites
 
 
 def test_read_sites():

tests/test_utils.py

@@ -1,6 +1,6 @@
 import pytest
 
-from argweaver.utils import parse_region
+from argweavers.utils import parse_region
 
 
 def test_parse_region():