Update from pkg-resources to importlib and bump to version 1.2.3

Author: Old-Shatterhand

.github/workflows/test.yaml

@@ -34,7 +34,7 @@ jobs:
       - name: Install environment
         shell: bash -l {0}
         run: |
-          mamba install -c conda-forge -c bioconda -y numpy"<2" pandas networkx matplotlib pytest setuptools pyscipopt scip"<=9.1" foldseek mmseqs2 cd-hit mash tmalign diamond cvxpy=1.5.3 pytest-cov rdkit">=2023.09.1" pytest-cases scikit-learn pyyaml h5py pyarrow tqdm
+          mamba install -c conda-forge -c bioconda -y numpy"<2" pandas networkx matplotlib pytest setuptools pyscipopt scip"<=9.1" foldseek mmseqs2 cd-hit mash tmalign diamond cvxpy=1.5.3 pytest-cov rdkit">=2023.09.1" pytest-cases scikit-learn pyyaml h5py pyarrow tqdm importlib_resources
           pip install grakel
         
       - name: Run tests

.github/workflows/test_all_pyv.yaml

@@ -40,7 +40,7 @@ jobs:
       - name: Install environment
         shell: bash -l {0}
         run: |
-          mamba install -c conda-forge -c bioconda -y numpy"<2" pandas networkx matplotlib pytest setuptools pyscipopt scip"<=9.1" foldseek mmseqs2 cd-hit mash tmalign diamond cvxpy=1.5.3 pytest-cov rdkit">=2023.09.1" pytest-cases scikit-learn pyyaml h5py pyarrow tqdm
+          mamba install -c conda-forge -c bioconda -y numpy"<2" pandas networkx matplotlib pytest setuptools pyscipopt scip"<=9.1" foldseek mmseqs2 cd-hit mash tmalign diamond cvxpy=1.5.3 pytest-cov rdkit">=2023.09.1" pytest-cases scikit-learn pyyaml h5py pyarrow tqdm importlib_resources
           pip install grakel
 
       - name: Run tests

.github/workflows/test_rdkit.yaml

@@ -36,7 +36,7 @@ jobs:
       - name: Install environment
         shell: bash -l {0}
         run: |
-          mamba install -c conda-forge -c bioconda -y numpy"<2" pandas networkx matplotlib pytest setuptools pyscipopt scip"<=9.1" foldseek mmseqs2 cd-hit mash tmalign diamond cvxpy=1.5.3 pytest-cov rdkit">=2023.03.1,<2023.09.1" pytest-cases scikit-learn pyyaml h5py pyarrow tqdm
+          mamba install -c conda-forge -c bioconda -y numpy"<2" pandas networkx matplotlib pytest setuptools pyscipopt scip"<=9.1" foldseek mmseqs2 cd-hit mash tmalign diamond cvxpy=1.5.3 pytest-cov rdkit">=2023.03.1,<2023.09.1" pytest-cases scikit-learn pyyaml h5py pyarrow tqdm importlib_resources
           pip install grakel
 
       - name: Run tests

CHANGELOG.md

@@ -12,6 +12,11 @@
 - [ ] Include MASH for amino acid sequences
 - [ ] Custom clustering methods ([Issue #25](https://github.com/kalininalab/DataSAIL/issues/25))
 
+## v1.2.3 (2025-10-21)
+
+- Bug fixed in overflow handling.
+- Switch from `pkg-resources` to `importlib`.
+
 ## v1.2.2 (2025-10-14)
 
 - Bug fixed in the evaluation module.

base_recipe.yaml

@@ -1,5 +1,5 @@
 package:
-  version: '1.2.2'
+  version: '1.2.3'
 
 source:
   path: ..
@@ -35,6 +35,7 @@ requirements:
     - h5py
     - pyarrow
     - tqdm
+    - importlib_resources
 
 about:
   home: https://github.com/kalininalab/DataSAIL

datasail/settings.py

@@ -7,7 +7,7 @@
 import platform
 
 import cvxpy
-import pkg_resources
+import importlib_resources
 
 
 def get_default(data_type: str, data_format: str) -> Tuple[Optional[str], Optional[str]]:
@@ -134,17 +134,17 @@ def format2ending(fmt: str) -> str:
 
 # YAML-files storing the arguments to the tools
 YAML_FILE_NAMES = {
-    MMSEQS: pkg_resources.resource_filename("datasail", "args/mmseqs2.yaml"),
-    MMSEQS2: pkg_resources.resource_filename("datasail", "args/mmseqs2.yaml"),
-    MMSEQSPP: pkg_resources.resource_filename("datasail", "args/mmseqspp.yaml"),
-    CDHIT: pkg_resources.resource_filename("datasail", "args/cdhit.yaml"),
-    CDHIT_EST: pkg_resources.resource_filename("datasail", "args/cdhit_est.yaml"),
-    DIAMOND: pkg_resources.resource_filename("datasail", "args/diamond.yaml"),
-    FOLDSEEK: pkg_resources.resource_filename("datasail", "args/foldseek.yaml"),
-    ECFP: pkg_resources.resource_filename("datasail", "args/.yaml"),
-    MASH: pkg_resources.resource_filename("datasail", "args/mash.yaml"),
-    MASH_SKETCH: pkg_resources.resource_filename("datasail", "args/mash_sketch.yaml"),
-    MASH_DIST: pkg_resources.resource_filename("datasail", "args/mash_dist.yaml"),
+    MMSEQS: importlib_resources.files("datasail") / "args/mmseqs2.yaml",
+    MMSEQS2: importlib_resources.files("datasail") / "args/mmseqs2.yaml",
+    MMSEQSPP: importlib_resources.files("datasail") / "args/mmseqspp.yaml",
+    CDHIT: importlib_resources.files("datasail") / "args/cdhit.yaml",
+    CDHIT_EST: importlib_resources.files("datasail") / "args/cdhit_est.yaml",
+    DIAMOND: importlib_resources.files("datasail") / "args/diamond.yaml",
+    FOLDSEEK: importlib_resources.files("datasail") / "args/foldseek.yaml",
+    ECFP: importlib_resources.files("datasail") / "args/.yaml",
+    MASH: importlib_resources.files("datasail") / "args/mash.yaml",
+    MASH_SKETCH: importlib_resources.files("datasail") / "args/mash_sketch.yaml",
+    MASH_DIST: importlib_resources.files("datasail") / "args/mash_dist.yaml",
 }
 
 KW_CACHE = "cache"

datasail/version.py

@@ -1 +1 @@
-__version__ = "1.2.2"
+__version__ = "1.2.3"

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "datasail"
-version = "1.2.2"
+version = "1.2.3"
 repository = "https://github.com/kalininalab/DataSAIL"
 readme = "README.md"
 description = "A package to compute hard out-of-distribution data splits for machine learning, challenging generalization of models."
@@ -29,6 +29,7 @@ h5py = "*"
 pyarrow = "*"
 tqdm = "*"
 grakel = "<0.1.10"
+importlib_resources = "*"
 
 [tool.poetry.dev-dependencies]
 setuptools = "*"