fix: add python 3.11 to 3.14 and drop 3.9 support

.github/actions/setup_nox/action.yml

@@ -5,7 +5,7 @@ inputs:
   cache_number:
     description: "A manually specified cache number. Useful for triggering new caches."
     required: True
-    default: 0
+    default: "0"
   poetry_version:
     description: "Which poetry version to use."
     required: True
@@ -20,6 +20,18 @@ runs:
     - uses: actions/setup-python@v5
       with:
         python-version: "3.10"
+    - uses: actions/setup-python@v5
+      with:
+        python-version: "3.11"
+    - uses: actions/setup-python@v5
+      with:
+        python-version: "3.12"
+    - uses: actions/setup-python@v5
+      with:
+        python-version: "3.13"
+    - uses: actions/setup-python@v5
+      with:
+        python-version: "3.14"
     - name: "Install nox"
       run: pipx install nox
       shell: bash

.github/workflows/cicd.yaml

@@ -46,7 +46,7 @@ jobs:
         with:
           cache_number: ${{ env.CACHE_NUMBER }}
           poetry_version: ${{ env.POETRY_VERSION }}
-      - name: Check for Dependency Vulnerabilities
+      - name: Lint code
         run: nox -s lint
 
   type:
@@ -60,7 +60,7 @@ jobs:
         with:
           cache_number: ${{ env.CACHE_NUMBER }}
           poetry_version: ${{ env.POETRY_VERSION }}
-      - name: Check for Dependency Vulnerabilities
+      - name: Type check
         run: nox -s type
 
   security:
@@ -103,7 +103,7 @@ jobs:
         run: nox -s tests
       - name: Upload to codecov
         if: matrix.os == 'ubuntu-latest'
-        uses: codecov/codecov-action@v2
+        uses: codecov/codecov-action@v4
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
           files: ./coverage.xml
@@ -114,7 +114,7 @@ jobs:
   release:
     runs-on: ubuntu-latest
     if: ${{ github.ref == 'refs/heads/main' }}
-    needs: [lint, test, security]
+    needs: [lint, type, test, security]
     outputs:
       released: ${{ steps.release.outputs.released }}
       version: ${{ steps.release.outputs.version }}
@@ -129,7 +129,7 @@ jobs:
       - name: Set up Python
         uses: actions/setup-python@v5
         with:
-          python-version: '3.9'
+          python-version: '3.14'
 
       - name: Install dependencies
         run: |
@@ -149,8 +149,6 @@ jobs:
           semantic-release version
           poetry build
           semantic-release publish
-          # Mark that a release was created
-          echo "released=true" >> $GITHUB_OUTPUT
 
       - name: Upload distribution artifacts
         uses: actions/upload-artifact@v4

CHANGELOG.md

@@ -1,6 +1,38 @@
 # CHANGELOG
 
 
+## v3.6.0 (2026-06-14)
+
+### Bug Fixes
+
+- **data_handling.py**: Fix deprecated pandas `groupby(axis=...)` API for pandas 2.0 compatibility
+
+- **data_handling.py**: Replace `np.NAN` with `np.nan` for NumPy 2.0 compatibility
+
+### Build System
+
+- **jaccard.py**: Vendor `boolean-jaccard` module locally; removes the external dependency which was
+  unmaintained and capped at Python < 3.11
+
+- **pyproject.toml**: Drop Python 3.8 and 3.9 support; add Python 3.11, 3.12, and 3.13 support
+
+- **pyproject.toml**: Bump pandas to `^2.0` and numpy to `^2.0`
+
+- **pyproject.toml**: Update dev tooling — flake8, pytest, coverage, black, Sphinx; replace `safety`
+  with `pip-audit`
+
+### Continuous Integration
+
+- **noxfile.py**: Update test matrix from Python 3.9/3.10 to 3.10–3.13; update default session
+  Python to 3.13
+
+- **noxfile.py**: Replace `safety` with `pip-audit` in security session; update `poetry export`
+  flags from `--dev` / `--no-dev` to `--with dev` / `--only main` for Poetry v1.2+ compatibility
+
+- **cicd.yaml**: Add Python 3.11–3.14 to CI setup; upgrade codecov action to v4; gate release job
+  on type check passing; fix CI step display names
+
+
 ## v3.5.3 (2025-11-09)
 
 ### Bug Fixes

docs/conf.py

@@ -12,7 +12,7 @@
 project = "LTA"
 author = "Ryan B Patterson-Cross"
 copyright = "2021, IMS-MRL Bioinformatics and Biostatistic Core"
-version = "3.5.2"
+version = "3.5.3"
 extensions = [
     "sphinx_rtd_theme",
     "sphinx.ext.autodoc",

lta/__init__.py

@@ -7,4 +7,5 @@
     The version number,
     specified in the form 'major.minor.patch'
 """
-__version__ = "3.5.2"
+
+__version__ = "3.5.3"

lta/cli.py

@@ -1,5 +1,6 @@
 # -*- coding: utf-8 -*-
 """Provide entry point for CLI."""
+
 import logging
 from datetime import datetime
 from pathlib import Path

lta/commands/run.py

@@ -1,5 +1,6 @@
 # -*- coding: utf-8 -*-
 """A simple sample function for the CLI."""
+
 import logging
 
 import configargparse

lta/helpers/data_handling.py

@@ -7,6 +7,7 @@
 and, thus, testable -
 by removing it from the harder to test context of the object.
 """
+
 import logging
 from pathlib import Path
 from typing import Any, List, Literal, Optional, Tuple
@@ -109,11 +110,17 @@ def not_zero(
     pd.DataFrame
         The processed data.
     """
-    df = (
-        (df == 0)
-        .groupby(axis=axis, level=[compartment, level])
-        .agg(lambda x: x.sum() <= (thresh * len(x)))
-    )
+    bool_df = df == 0
+    if axis == "index":
+        df = bool_df.groupby(level=[compartment, level]).agg(
+            lambda x: x.sum() <= (thresh * len(x))
+        )
+    else:
+        df = (
+            bool_df.T.groupby(level=[compartment, level])
+            .agg(lambda x: x.sum() <= (thresh * len(x)))
+            .T
+        )
     if axis == "index":
         df = df.loc[:, df.any(axis=axis)]
     if axis == "columns":
@@ -168,7 +175,10 @@ def enfc(
     if not order:
         logging.debug("Order not passed. Defaulting to ('experimental', 'control')")
         order = ("experimental", "control")
-    mean = df.groupby(axis=axis, level=level).mean()
+    if axis == "index":
+        mean = df.groupby(level=level).mean()
+    else:
+        mean = df.T.groupby(level=level).mean().T
     logging.debug(f"Grouping/filtering on {axis}.")
     # Replace inf (x/0) with NaN
     # Replace 0 (0/x) with NaN
@@ -177,20 +187,30 @@ def enfc(
         logfc = np.log10(
             mean.loc[order[0], :]
             .div(mean.loc[order[1], :])
-            .replace([np.inf, -np.inf, 0], np.NAN)
+            .replace([np.inf, -np.inf, 0], np.nan)
         )
     else:
         logfc = np.log10(
             mean.loc[:, order[0]]
             .div(mean.loc[:, order[1]])
-            .replace([np.inf, -np.inf, 0], np.NAN)
+            .replace([np.inf, -np.inf, 0], np.nan)
+        )
+    if axis == "index":
+        error = (
+            df.groupby(level=level)
+            .std(numeric_only=True)
+            .pow(2)
+            .sum(axis=axis)
+            .div(2)
+            .pow(0.5)
+        )
+    else:
+        error = (
+            df.T.groupby(level=level)
+            .std(numeric_only=True)
+            .T.pow(2)
+            .sum(axis=axis)
+            .div(2)
+            .pow(0.5)
         )
-    error = (
-        df.groupby(axis=axis, level=level)
-        .std(numeric_only=True)
-        .pow(2)
-        .sum(axis=axis)
-        .div(2)
-        .pow(0.5)
-    )
     return logfc.div(error)

lta/helpers/jaccard.py

@@ -0,0 +1,112 @@
+# -*- coding: utf-8 -*-
+"""Jaccard similarities and their p-values.
+
+Vendored from boolean-jaccard 0.1.1 (https://github.com/rbpatt2019/boolean-jaccard),
+which is unmaintained and capped at Python <3.11.
+Original: python port of the R jaccard package by N. Chung.
+"""
+
+import logging
+from typing import Optional
+
+import numpy as np
+import pandas as pd
+
+logger = logging.getLogger(__name__)
+
+
+def similarity(
+    x: np.ndarray,
+    y: np.ndarray,
+    center: bool = False,
+    px: Optional[float] = None,
+    py: Optional[float] = None,
+) -> float:
+    """Calculate Jaccard similarity."""
+    if x.ndim != 1 or y.ndim != 1:
+        logging.error(
+            f"All vectors must be 1-d. ndims: {[x.ndim, y.ndim]}.", stack_info=True
+        )
+        raise IndexError
+    if x.shape != y.shape:  # type: ignore[operator]
+        logging.error(
+            f"All vectors must have the same length. shape: {[x.shape, y.shape]}",
+            stack_info=True,
+        )
+        raise IndexError
+    if x.dtype != bool or y.dtype != bool:
+        logging.error(
+            f"All vectors must be boolean. dtypes: {[x.dtype, y.dtype]}",
+            stack_info=True,
+        )
+        raise TypeError
+
+    if px is None:
+        px = float(x.mean())
+    if py is None:
+        py = float(y.mean())
+
+    intersect = (x & y).sum()
+    union = x.sum() + y.sum() - intersect
+
+    denominator = px + py - (px * py)
+    if denominator == 0:
+        return np.nan
+
+    if union == 0:
+        j = (px * py) / denominator
+    else:
+        j = intersect / union
+
+    if center:
+        return j - ((px * py) / denominator)
+    return j
+
+
+def distance(
+    x: np.ndarray, y: np.ndarray, px: Optional[float] = None, py: Optional[float] = None
+) -> float:
+    """Calculate Jaccard distance (1 - similarity)."""
+    if px is None:
+        px = float(x.mean())
+    if py is None:
+        py = float(y.mean())
+    return 1 - similarity(x, y, center=False, px=px, py=py)
+
+
+def bootstrap(
+    x: np.ndarray,
+    y: np.ndarray,
+    px: Optional[float] = None,
+    py: Optional[float] = None,
+    n: int = 1000,
+    seed: int = 42,
+) -> pd.Series:
+    """Bootstrap p-value for Jaccard similarity."""
+    j = similarity(x, y, center=False, px=px, py=py)
+    if px is None:
+        px = float(x.mean())
+    if py is None:
+        py = float(y.mean())
+    if px == 1 or py == 1 or len(x) == x.sum() or len(y) == y.sum():
+        logger.warning("Bootstrap is degenerate as at least one vector is all 1.")
+        return pd.Series([j, 1], index=["J-sim", "p-val"])
+    if px == 0 or py == 0 or x.sum() == 0 or y.sum() == 0:
+        logger.warning("Bootstrap is degenerate as at least one vector is all 0.")
+        return pd.Series([j, 1], index=["J-sim", "p-val"])
+
+    j_obs = similarity(x, y, center=True, px=px, py=py)
+
+    rng = np.random.default_rng(seed)
+    vals = (
+        similarity(
+            rng.choice(x, size=len(x), replace=True, shuffle=False),
+            rng.choice(y, size=len(y), replace=True, shuffle=False),
+            center=True,
+        )
+        for _ in range(n)
+    )
+    j_null = np.fromiter(vals, dtype=np.float32, count=n)
+    np.abs(j_null, dtype=np.float32, out=j_null)
+    p_val = (j_null >= np.abs(j_obs)).sum() / n
+    return pd.Series([j, p_val], index=["J-sim", "p-val"])

lta/helpers/pipeline.py

@@ -1,15 +1,16 @@
 # -*- coding: utf-8 -*-
 """A dataclass that allows for an object oriented pipeline."""
+
 import itertools
 import logging
 from dataclasses import dataclass
 from pathlib import Path
 from typing import Dict
 
 import pandas as pd
-from jaccard import jaccard as jac
 
 import lta.helpers.data_handling as dh
+from lta.helpers import jaccard as jac
 from lta.helpers import utils
 
 logger = logging.getLogger(__name__)

lta/parser.py

@@ -53,6 +53,7 @@
 lta_parser : argparse.ArgumentParser
     The argument parser for the root command.
 """
+
 from pathlib import Path
 
 import configargparse