Move to ruff for style/lint/etc

.github/workflows/pythonpackage.yml

@@ -7,8 +7,16 @@ on:
   pull_request: []
 
 jobs:
+  lint:
+    runs-on: ubuntu-latest
+    env:
+      GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+    steps:
+      - uses: actions/checkout@v4
+      - uses: astral-sh/ruff-action@v3
+      - run: ruff format
   build:
-
+    needs: lint
     runs-on: ubuntu-latest
     strategy:
       max-parallel: 4
@@ -18,7 +26,6 @@ jobs:
           - "3.11"
           - "3.12"
           - "3.13"
-
     steps:
       - name: Install apt packages
         run: |
@@ -46,15 +53,6 @@ jobs:
           echo $poetryv
       - name: Install package for testing
         run: poetry install -v --all-extras
-      - name: Lint with flake8
-        run: |
-          pipx install flake8
-          # stop the build if there are Python syntax errors or undefined names
-          flake8 . --count --select=E9,F63,F7,F82 --show-source --statistics
-          # exit-zero treats all errors as warnings
-          # The GitHub editor is 127 chars wide
-          flake8 . --count --ignore=E302 --exit-zero --max-complexity=10 \
-            --max-line-length=127 --statistics
       - name: Test with pytest
         run: poetry run pytest --script-launch-mode=subprocess
       - name: Reinstall package for production

pyproject.toml

@@ -46,7 +46,6 @@ tact_add_config = "tact.cli_add_toml:main"
 tact_build_taxonomic_tree = "tact.cli_taxonomy:main"
 tact_check_results = "tact.cli_check_trees:main"
 
-
 [tool.poetry]
 include = ["examples"]
 classifiers = [
@@ -56,9 +55,36 @@ classifiers = [
     "Topic :: Scientific/Engineering :: Bio-Informatics"
 ]
 
-[tool.black]
+[tool.ruff]
 line-length = 118
-target-version = ['py38', 'py39', 'py310', 'py311']
+
+[tool.ruff.lint]
+fixable = ["ALL"]
+select = [
+    "NPY", # numpy warnings
+    "E",   # pycodestyle errors
+    "W",   # pycodestyle warnings
+    "F",   # pyflakes
+    "D",   # pydocstyle
+    "DOC", # pydoclint
+    "I",   # isort
+    "RUF", # ruff
+    "B",   # flake8-bugbear
+    "UP",  # pyupgrade
+    "ICN", # flake8-import-conventions
+    "FA",  # flake8-future-annotations
+    "C4",  # flake8-comprehensions
+    "A"    # flake8-builtins
+]
+
+[tool.ruff.lint.per-file-ignores]
+# Ignore all directories named `tests` and CLI modules.
+"tests/**" = ["D", "DOC"]
+"tact/cli_*" = ["D", "DOC"]
+"__init__.py" = ["D", "DOC"]
+
+[tool.ruff.lint.pydocstyle]
+convention = "google"
 
 [build-system]
 requires = ["poetry_core>=2.0"]

tact/cli_add_taxa.py

@@ -1,11 +1,8 @@
 #!/usr/bin/env python
-# -*- coding: utf-8 -*-
 
 # Try to assign tips to a pre-existing tree based on a taxonomy
 # Jonathan Chang, May 13, 2016
 
-from __future__ import division
-from __future__ import print_function
 
 import csv
 import logging
@@ -18,21 +15,20 @@
 import dendropy
 
 from . import fastmrca
-from .lib import crown_capture_probability
-from .lib import get_new_times
-from .tree_util import get_ages
-from .tree_util import get_birth_death_rates
-from .tree_util import get_min_age
-from .tree_util import get_short_branches
-from .tree_util import get_tip_labels
-from .tree_util import graft_node
-from .tree_util import is_binary
-from .tree_util import is_fully_locked
-from .tree_util import lock_clade
-from .tree_util import update_tree_view
-from .validation import validate_outgroups
-from .validation import validate_taxonomy_tree
-from .validation import BackboneCommand
+from .lib import crown_capture_probability, get_new_times
+from .tree_util import (
+    get_ages,
+    get_birth_death_rates,
+    get_min_age,
+    get_short_branches,
+    get_tip_labels,
+    graft_node,
+    is_binary,
+    is_fully_locked,
+    lock_clade,
+    update_tree_view,
+)
+from .validation import BackboneCommand, validate_outgroups, validate_taxonomy_tree
 
 logger = logging.getLogger(__name__)
 # Speed up logging for PyPy
@@ -87,9 +83,7 @@ def search_ancestors_for_valid_backbone_node(taxonomy_node, backbone_tips, ccp):
 
 
 def get_new_branching_times(backbone_node, taxonomy_node, told=None, tyoung=0, min_ccp=0.8, num_new_times=None):
-    """
-    Get `n_total` new branching times for a `node`.
-    """
+    """Get `n_total` new branching times for a `node`."""
     global mrca_rates
     taxon = taxonomy_node.label
     birth, death, ccp, _ = mrca_rates[taxon]
@@ -131,7 +125,7 @@ def get_new_branching_times(backbone_node, taxonomy_node, told=None, tyoung=0, m
 
 
 def fill_new_taxa(namespace, node, new_taxa, times, stem=False):
-    for new_species, new_age in zip(new_taxa, times):
+    for new_species, new_age in zip(new_taxa, times, strict=True):
         new_node = dendropy.Node()
         new_node.annotations.add_new("creation_method", "fill_new_taxa")
         new_node.age = new_age
@@ -184,7 +178,7 @@ def create_clade(namespace, species, ages):
     # Lock the child of the seed node so that things can still attach to the stem of this new clade
     lock_clade(tree.seed_node.child_nodes()[0])
     if list(get_short_branches(tree.seed_node)):
-        logger.info("{} short branches detected".format(len(list(get_short_branches(tree.seed_node)))))
+        logger.info(f"{len(list(get_short_branches(tree.seed_node)))} short branches detected")
     return tree
 
 
@@ -315,9 +309,7 @@ def run_precalcs(taxonomy_tree, backbone_tree, min_ccp=0.8, yule=False):
 )
 @click.option("-v", "--verbose", help="emit extra information (can be repeated)", count=True)
 def main(taxonomy, backbone, outgroups, output, min_ccp, verbose, yule, ultrametricity_precision):
-    """
-    Add tips onto a BACKBONE phylogeny using a TAXONOMY phylogeny.
-    """
+    """Add tips onto a BACKBONE phylogeny using a TAXONOMY phylogeny."""
     logger.addHandler(logging.FileHandler(output + ".log.txt"))
     if verbose >= 2:
         logger.setLevel(logging.DEBUG)

tact/cli_add_toml.py

@@ -1,35 +1,34 @@
-#!/usr/bin/env python
-# -*- coding: utf-8 -*-
-
-# Try to assign tips to a pre-existing tree based on a TOML configuration file
-# Jonathan Chang, Aug 14, 2021
+"""Command line interface to assign tips to a pre-existing tree based on a TOML configuration file."""
 
 from __future__ import annotations
 
-from concurrent.futures import ProcessPoolExecutor, as_completed
-from dataclasses import dataclass, field, InitVar
-from collections import defaultdict
 import copy
 import logging
+import operator
 import os
 import re
 import sys
-import typing
+from collections import defaultdict
+from concurrent.futures import ProcessPoolExecutor, as_completed
+from dataclasses import InitVar, dataclass, field
+from functools import reduce
 
 import click
 import dendropy
 import toml
 
 from .lib import get_new_times
-from .tree_util import get_ages
-from .tree_util import get_birth_death_rates
-from .tree_util import get_min_age
-from .tree_util import get_tip_labels
-from .tree_util import graft_node
-from .tree_util import is_binary
-from .tree_util import lock_clade
-from .tree_util import unlock_clade
-from .tree_util import update_tree_view
+from .tree_util import (
+    get_ages,
+    get_birth_death_rates,
+    get_min_age,
+    get_tip_labels,
+    graft_node,
+    is_binary,
+    lock_clade,
+    unlock_clade,
+    update_tree_view,
+)
 from .validation import BackboneCommand
 
 logger = logging.getLogger(__name__)
@@ -42,9 +41,9 @@
 
 @dataclass
 class TactConstraint:
-    """Class for keeping track of a constraint in TACT (positive or negative)"""
+    """Class for keeping track of a constraint in TACT (positive or negative)."""
 
-    mrca: typing.List[str] = field(default_factory=list)
+    mrca: list[str] = field(default_factory=list)
     stem: bool = False
 
     def __post_init__(self):
@@ -91,6 +90,7 @@ def __post_init__(self, include, exclude):
 
 
 def ensure_mrca(tree, tips, node=None):
+    """Perform initial checks to ensure we can do MRCA calculations."""
     try:
         node = node if node else tree.seed_node
         return tree.mrca(taxon_labels=tips, start_node=node)
@@ -112,7 +112,7 @@ def ensure_mrca(tree, tips, node=None):
 def do_tact(tree, item):
     # First, get the MRCA of _all_ `include` leafs. This is the basis of our rate computation,
     # and how we actually implement polyphyletic groups.
-    included_tips = sum([x.mrca for x in item.include], [])
+    included_tips = reduce(operator.iadd, [x.mrca for x in item.include], [])
     mrca_node = ensure_mrca(tree, included_tips)
 
     # Compute the rates on that (possibly expansive) MRCA node.
@@ -143,7 +143,7 @@ def do_tact(tree, item):
                 genera_map[genus].add(tip)
 
             if len(genera_map) > 1:
-                for genus, species in genera_map.items():
+                for species in genera_map.values():
                     node = tree.mrca(taxon_labels=species, start_node=inner_mrca_node)
                     if node and species == get_tip_labels(node):
                         if len(species) == 1:
@@ -178,6 +178,7 @@ def do_tact(tree, item):
 
 
 def do_replicate(backbone, to_tact, label):
+    """Perform a replicate of a TACT analysis."""
     logger.info(f"<<< Replicate {label} >>>")
     tree = copy.deepcopy(backbone)
     for item in to_tact:
@@ -208,9 +209,7 @@ def do_replicate(backbone, to_tact, label):
     default=os.cpu_count() or 1,
 )
 def main(config, backbone, output, verbose, ultrametricity_precision, replicates, cores):
-    """
-    Add tips onto a BACKBONE phylogeny using a CONFIG file
-    """
+    """Add tips onto a BACKBONE phylogeny using a CONFIG file."""
     logger.addHandler(logging.FileHandler(output + ".log.txt"))
     if verbose >= 2:
         logger.setLevel(logging.DEBUG)
@@ -226,7 +225,7 @@ def main(config, backbone, output, verbose, ultrametricity_precision, replicates
     to_tact = [TactItem(**x) for x in config["tact"]]
 
     # Ensure the proper ordering of TACT items based on divergence time of implied MRCA nodes
-    to_tact.sort(key=lambda item: ensure_mrca(backbone, sum([x.mrca for x in item.include], [])).age)
+    to_tact.sort(key=lambda ii: ensure_mrca(backbone, reduce(operator.iadd, [x.mrca for x in ii.include], [])).age)
 
     # Compute global birth/death rates. Not currently used (but could be?)
     backbone_tips = len(backbone.leaf_nodes())

tact/cli_check_trees.py

@@ -1,7 +1,4 @@
-#!/usr/bin/env python
-# -*- coding: utf-8 -*-
-
-from __future__ import division
+"""Command-line interface module to check TACT results."""
 
 import csv
 import functools
@@ -12,13 +9,11 @@
 import click
 import dendropy
 
-from .tree_util import get_birth_death_rates
-from .tree_util import get_monophyletic_node
-from .tree_util import get_tip_labels
-from .tree_util import get_tree
+from .tree_util import get_birth_death_rates, get_monophyletic_node, get_tip_labels, get_tree
 
 
 def analyze_taxon(bb_tips, st_tips, backbone, simtaxed, taxon_node):
+    """Perform various checks for a given taxon."""
     taxon = taxon_node.label
     if not taxon:
         return None
@@ -89,14 +84,13 @@ def analyze_taxon(bb_tips, st_tips, backbone, simtaxed, taxon_node):
 )
 @click.option("--chunksize", help="number of tree nodes to allocate to each core", type=click.IntRange(1))
 def main(simulated, backbone, taxonomy, output, cores, chunksize):
-    """
-    Check a SIMULATED phylogeny for consistency with its backbone source tree and a taxonomy.
+    """Check a SIMULATED phylogeny for consistency with its backbone source tree and a taxonomy.
 
     The SIMULATED phylogeny should have been generated by the tact_add_taxa script.
     All phylogenies should be in Newick format.
     """
     pool = multiprocessing.Pool(processes=cores)
-    click.echo("Using %d parallel cores" % cores, err=True)
+    click.echo(f"Using {cores} parallel cores", err=True)
     taxonomy = dendropy.Tree.get_from_path(taxonomy, schema="newick")
     tn = taxonomy.taxon_namespace
     click.echo("Taxonomy OK", err=True)