Release 1.11.0

Author: kfuku52

.gitignore

@@ -3,7 +3,7 @@
 !/tests/test_*.py
 /data/
 build/*
-csubst/__pycache__
+__pycache__/
 csubst/*-darwin.so
 csubst/*.c
 csubst/.DS_Store

README.md

@@ -55,14 +55,15 @@ pip install git+https://github.com/kfuku52/csubst
 csubst dataset --name PGK
 
 # Run csubst analyze
-csubst analyze --alignment_file alignment.fa --rooted_tree_file tree.nwk --foreground foreground.txt
+csubst analyze --alignment_file alignment.fa.gz --rooted_tree_file tree.nwk --foreground foreground.txt
 ```
 
 ## Usage
-CSUBST provides four main subcommands:
+CSUBST provides five main subcommands:
 
 - `csubst dataset`: generate bundled example datasets (e.g., `PGK`, `PEPC`).
 - `csubst analyze`: run convergence analysis and output metrics such as `omegaC`, `dNC`, and `dSC`.
+- `csubst inspect`: summarize branch mappings and inspect ancestral states.
 - `csubst site`: compute site-wise combinatorial substitutions for selected branch combinations, generate tree + site summary plots, and optionally map sites to protein structures.
 - `csubst simulate`: simulate codon sequence evolution under user-defined convergent scenarios.
 
@@ -81,13 +82,13 @@ csubst dataset --name PGK
 
 # 2) Run convergence analysis
 csubst analyze \
-  --alignment_file alignment.fa \
+  --alignment_file alignment.fa.gz \
   --rooted_tree_file tree.nwk \
   --foreground foreground.txt
 
 # 3) Inspect site-wise convergence for a branch pair (example)
 csubst site \
-  --alignment_file alignment.fa \
+  --alignment_file alignment.fa.gz \
   --rooted_tree_file tree.nwk \
   --branch_id 23,51
 ```

csubst/__init__.py

@@ -1 +1 @@
-__version__ = '1.10.2'
+__version__ = '1.11.0'

csubst/csubst

@@ -65,6 +65,11 @@ def _resolve_output_namespace_defaults(argv):
                 'outdir': 'csubst_inspect',
                 'output_prefix': 'csubst',
             }
+        if token == 'site':
+            return {
+                'outdir': 'csubst_site',
+                'output_prefix': 'csubst',
+            }
         break
     return {
         'outdir': '.',
@@ -143,6 +148,8 @@ def command_simulate(args):
 def command_site(args):
     print('csubst site start:', datetime.datetime.now(datetime.timezone.utc), flush=True)
     start = time.time()
+    if str(getattr(args, 'log_file', '')).strip() == '':
+        args.log_file = runtime.default_site_log_path(base_dir=os.getcwd(), create_dir=False)
     g = get_global_parameters_or_exit(args)
     from csubst.main_site import main_site
     main_site(g)

csubst/main_analyze.py

@@ -10,7 +10,6 @@
 from csubst import omega
 from csubst import param
 from csubst import parser_misc
-from csubst import sequence
 from csubst import substitution
 from csubst import table
 from csubst import ete
@@ -598,29 +597,6 @@ def main_analyze(g):
         prepare_state=False,
     )
     g = parser_misc.prep_state(g, apply_site_filtering=False)
-    loaded_branch_ids = g.get('state_loaded_branch_ids', None)
-    if loaded_branch_ids is not None:
-        txt = 'Selective state loading active: writing alignments only for loaded nodes ({:,}).'
-        print(txt.format(loaded_branch_ids.shape[0]), flush=True)
-    sequence.write_alignment(
-        runtime.output_path(g, 'alignment_codon.fa'),
-        mode='codon',
-        g=g,
-        branch_ids=loaded_branch_ids,
-    )
-    sequence.write_alignment(
-        runtime.output_path(g, 'alignment_aa.fa'),
-        mode='aa',
-        g=g,
-        branch_ids=loaded_branch_ids,
-    )
-    if str(g.get('nonsyn_recode', 'no')).strip().lower() == '3di20':
-        sequence.write_alignment(
-            runtime.output_path(g, 'alignment_3di.fa'),
-            mode='nsy',
-            g=g,
-            branch_ids=loaded_branch_ids,
-        )
     g = parser_misc.apply_site_filters(g)
     g = combination.get_dep_ids(g)
     ON_tensor = substitution.get_substitution_tensor(state_tensor=g['state_nsy'], mode='asis', g=g, mmap_attr='N')

csubst/main_site.py

@@ -3231,6 +3231,30 @@ def _build_site_outdir(mode, branch_txt, lineage_input_branch_txt=None, mode_exp
     return './csubst_site.mode' + mode + '.branch_id' + branch_txt
 
 
+def _maybe_relocate_site_log_file(g):
+    site_jobs = g.get('site_jobs', [])
+    if len(site_jobs) != 1:
+        return g
+    log_file = str(g.get('log_file', '')).strip()
+    if log_file == '':
+        return g
+    current_log_path = os.path.abspath(log_file)
+    default_log_path = runtime.default_site_log_path(base_dir=os.getcwd(), create_dir=False)
+    if current_log_path != default_log_path:
+        return g
+    target_dir = os.path.abspath(site_jobs[0]['site_outdir'])
+    os.makedirs(target_dir, exist_ok=True)
+    target_log_path = os.path.join(target_dir, os.path.basename(default_log_path))
+    if current_log_path != target_log_path:
+        if os.path.exists(current_log_path):
+            os.replace(current_log_path, target_log_path)
+        parent_dir = os.path.dirname(current_log_path)
+        if (parent_dir != '') and os.path.isdir(parent_dir) and (len(os.listdir(parent_dir)) == 0):
+            os.rmdir(parent_dir)
+    g['log_file'] = target_log_path
+    return g
+
+
 def resolve_site_jobs(g):
     raw_mode = str(g.get('mode', 'intersection')).strip()
     mode, mode_expression, set_stat_type = _parse_mode_and_expression(raw_mode)
@@ -3335,6 +3359,7 @@ def main_site(g):
     OS_tensor = substitution.get_substitution_tensor(state_tensor=g['state_cdn'], mode='syn', g=g, mmap_attr='S')
     OS_tensor = substitution.apply_min_sub_pp(g, OS_tensor)
     g = resolve_site_jobs(g)
+    g = _maybe_relocate_site_log_file(g)
     for site_job in g['site_jobs']:
         branch_ids = _normalize_branch_ids(site_job['branch_ids'])
         g['single_branch_mode'] = site_job['single_branch_mode']

csubst/runtime.py

@@ -12,6 +12,7 @@
 _DEFAULT_OUTPUT_DIR = "."
 _DEFAULT_OUTPUT_PREFIX = "csubst"
 _DEFAULT_IQTREE_OUTDIR = "csubst_iqtree"
+_DEFAULT_SITE_LOG_DIR = "csubst_site"
 _MISSING = object()
 
 
@@ -144,6 +145,15 @@ def ensure_iqtree_layout(g, create_dir=False):
     return g
 
 
+def default_site_log_path(base_dir=None, create_dir=False):
+    if base_dir is None:
+        base_dir = os.getcwd()
+    site_log_dir = os.path.abspath(os.path.join(str(base_dir), _DEFAULT_SITE_LOG_DIR))
+    if create_dir:
+        os.makedirs(site_log_dir, exist_ok=True)
+    return os.path.join(site_log_dir, _DEFAULT_OUTPUT_PREFIX + ".log")
+
+
 def _strip_gzip_suffix(path):
     path_txt = str(path).strip()
     if path_txt.lower().endswith(".gz"):

tests/test_cli_logging.py

@@ -62,6 +62,14 @@ def test_simulate_help_uses_simulate_default_log_name(tmp_path):
     assert not (tmp_path / "csubst.log").exists()
 
 
+def test_site_help_uses_site_default_log_name(tmp_path):
+    result = _run_csubst(["site", "-h"], cwd=tmp_path)
+    assert result.returncode == 0
+    log_file = tmp_path / "csubst_site" / "csubst.log"
+    assert log_file.exists()
+    assert not (tmp_path / "csubst.log").exists()
+
+
 def test_subcommand_output_namespace_defaults_are_command_specific():
     repo_root = Path(__file__).resolve().parents[1]
     script_path = repo_root / "csubst" / "csubst"

tests/test_cli_validation.py

@@ -19,6 +19,8 @@ def _resolve_log_path(repo_root, args):
             return repo_root / "csubst_analyze" / "csubst.log"
         if args[0] == "inspect":
             return repo_root / "csubst_inspect" / "csubst.log"
+        if args[0] == "site":
+            return repo_root / "csubst_site" / "csubst.log"
     return repo_root / "csubst.log"
 
 

tests/test_main_site_and_substitution.py

@@ -5,6 +5,7 @@
 from pathlib import Path
 
 from csubst import main_site
+from csubst import runtime
 from csubst import substitution
 from csubst import substitution_sparse
 from csubst import tree
@@ -587,6 +588,25 @@ def test_resolve_site_jobs_rejects_duplicate_branch_ids(tiny_tree):
         main_site.resolve_site_jobs(g)
 
 
+def test_maybe_relocate_site_log_file_moves_default_log_into_single_job_outdir(tmp_path, monkeypatch):
+    default_log = Path(runtime.default_site_log_path(base_dir=tmp_path, create_dir=True))
+    default_log.write_text("hello\n", encoding="utf-8")
+    site_outdir = tmp_path / "csubst_site.branch_id1,2"
+    g = {
+        "log_file": str(default_log),
+        "site_jobs": [{"site_outdir": str(site_outdir)}],
+    }
+
+    monkeypatch.chdir(tmp_path)
+    out = main_site._maybe_relocate_site_log_file(g)
+
+    relocated_log = site_outdir / "csubst.log"
+    assert out["log_file"] == str(relocated_log.resolve())
+    assert relocated_log.exists()
+    assert relocated_log.read_text(encoding="utf-8") == "hello\n"
+    assert not default_log.exists()
+
+
 def test_normalize_branch_ids_rejects_non_integer_like_values():
     with pytest.raises(ValueError, match="integer-like"):
         main_site._normalize_branch_ids(np.array([1.5]))