fix failing tests

.github/workflows/autorun-tests.yml

@@ -15,18 +15,24 @@ jobs:
 
     steps:
     - uses: actions/checkout@v3
-    - name: Set up Python 3.10
+
+    - name: Set up Python 3.11
       uses: actions/setup-python@v3
       with:
-        python-version: "3.10"
+        python-version: "3.11"
+
+    - name: Install uv
+      uses: astral-sh/setup-uv@v7
+
     - name: Install dependencies
       run: |
-        python -m pip install --upgrade pip
-        pip install -e ".[test,dev]"
+        uv sync --extra test --extra dev && uv lock
+
     - name: Lint and format with ruff
       run: |
-        ruff check . --statistics
-        ruff format .
+        uv run ruff check . --statistics
+        uv run ruff format .
+
     - name: Test with pytest
       run: |
-        pytest --cov=dscript --cov-report=xml --cov-report=term-missing
+        uv run pytest --cov-report=xml --cov-report=term-missing

dscript/__init__.py

@@ -1,4 +1,6 @@
-__version__ = "0.3.1"
+from importlib.metadata import version as _get_version
+
+__version__ = _get_version("dscript")
 __citation__ = """Sledzieski, Singh, Cowen, Berger. "D-SCRIPT translates genome to phenome with sequence-based, structure-aware, genome-scale predictions of protein-protein interactions." Cell Systems 12, no. 10 (2021): 969-982.
 
 Devkota, Singh, Sledzieski, Berger, Cowen, Topsy-Turvy: integrating a global view into sequence-based PPI prediction, Bioinformatics, In Press."""

dscript/tests/test_language_model.py

@@ -28,15 +28,15 @@ class TestLanguageModelUnit:
 
         # Mock the proj layer
         model.proj = Mock()
-        model.proj.weight = torch.randn(100, 100)
-        model.proj.bias = torch.zeros(100)
+        model.proj.weight = torch.randn(6165, 6165)
+        model.proj.bias = torch.zeros(6165)
 
         # Mock transform to return realistic embeddings
         def mock_transform(x):
             batch_size = x.shape[0]
             seq_len = x.shape[1]
-            # Return (batch, seq_len, embedding_dim=100)
-            return torch.randn(batch_size, seq_len, 100)
+            # Return (batch, seq_len, embedding_dim=6165)
+            return torch.randn(batch_size, seq_len, 6165)
 
         model.transform = Mock(side_effect=mock_transform)
         return model
@@ -49,10 +49,10 @@ class TestLanguageModelUnit:
         test_seq = "MKTAYIAKQRQISFVKSHFSRQ"
         x = lm_embed(test_seq, use_cuda=False)
 
-        # Should be (batch=1, seq_len, embedding_dim=100)
+        # Should be (batch=1, seq_len, embedding_dim=6165)
         assert x.shape[0] == 1
         assert x.shape[1] == len(test_seq)
-        assert x.shape[2] == 100
+        assert x.shape[2] == 6165
 
     @patch("dscript.language_model.get_pretrained")
     def test_lm_embed_returns_tensor(self, mock_get_pretrained, mock_model):
@@ -73,7 +73,7 @@ class TestLanguageModelUnit:
         x = lm_embed(short_seq, use_cuda=False)
 
         assert x.shape[1] == 2
-        assert x.shape[2] == 100
+        assert x.shape[2] == 6165
 
     @patch("dscript.language_model.get_pretrained")
     def test_lm_embed_single_amino_acid(self, mock_get_pretrained, mock_model):
@@ -84,7 +84,7 @@ class TestLanguageModelUnit:
         x = lm_embed(single_aa, use_cuda=False)
 
         assert x.shape[1] == 1
-        assert x.shape[2] == 100
+        assert x.shape[2] == 6165
 
     @patch("dscript.language_model.get_pretrained")
     def test_embed_from_fasta_creates_h5(self, mock_get_pretrained, mock_model, tmp_path):
@@ -215,7 +215,7 @@ class TestLanguageModelIntegration:
 
         assert x.shape[0] == 1
         assert x.shape[1] == len(test_seq)
-        assert x.shape[2] == 100
+        assert x.shape[2] == 6165
         assert isinstance(x, torch.Tensor)
 
     def test_embed_from_fasta_real(self, tmp_path):
@@ -239,4 +239,4 @@ class TestLanguageModelIntegration:
                 embedding = f[name][:]
                 assert embedding.shape[0] == 1
                 assert embedding.shape[1] == len(seq)
-                assert embedding.shape[2] == 100
+                assert embedding.shape[2] == 6165

dscript/tests/test_main.py

@@ -89,7 +89,7 @@ class TestMainFunction:
             "test.tsv",
             "--embedding",
             "embed.h5",
-            "--output",
+            "--outfile",
             "output",
             "--save-prefix",
             "model",
@@ -162,13 +162,13 @@ class TestMainFunction:
         test_args = [
             "dscript",
             "predict_bipartite",
-            "--seqs0",
-            "seqs0.fasta",
-            "--seqs1",
-            "seqs1.fasta",
-            "--embeddings0",
+            "--protA",
+            "protA.txt",
+            "--protB",
+            "protB.txt",
+            "--embedA",
             "embed0.h5",
-            "--embeddings1",
+            "--embedB",
             "embed1.h5",
             "--model",
             "model_path",
@@ -188,7 +188,7 @@ class TestMainFunction:
         test_args = [
             "dscript",
             "evaluate",
-            "--pairs",
+            "--test",
             "pairs.tsv",
             "--embeddings",
             "embed.h5",
@@ -261,7 +261,7 @@ class TestMainFunction:
             "test.tsv",
             "--embedding",
             "embed.h5",
-            "--output",
+            "--outfile",
             "output",
             "--save-prefix",
             "model",

dscript/tests/test_utils.py

@@ -95,7 +95,7 @@ class TestLog:
         mock_file.flush = Mock()
 
         log("flush test", file=mock_file)
-        mock_file.flush.assert_called_once()
+        mock_file.flush.assert_called()
 
 
 class TestRBF:

dscript/utils.py

@@ -53,7 +53,7 @@ def log(m, file=None, timestamped=True, print_also=False):
         file.flush()
 
 
-def RBF(D, sigma=None):
+def RBF(D, sigma=None, pseudocount=1e-10):
     """
     Convert distance matrix into similarity matrix using Radial Basis Function (RBF) Kernel.
 
@@ -66,6 +66,7 @@ def RBF(D, sigma=None):
     :return: Similarity matrix
     :rtype: np.ndarray
     """
+    D += pseudocount
     sigma = sigma or np.sqrt(np.max(D))
     return np.exp(-1 * (np.square(D) / (2 * sigma**2)))
 
@@ -161,7 +162,11 @@ def collate_paired_sequences(args):
     """
     Collate function for PyTorch data loader.
     """
+    if not len(args):
+        return [], [], torch.tensor([])
+
     x0 = [a[0] for a in args]
     x1 = [a[1] for a in args]
-    y = [a[2] for a in args]
-    return x0, x1, torch.stack(y, 0)
+    y = torch.stack([a[2] for a in args], 0)
+
+    return x0, x1, y

pyproject.toml

@@ -1,10 +1,10 @@
 [build-system]
-requires = ["setuptools>=45", "wheel"]
-build-backend = "setuptools.build_meta"
+requires = ["uv_build>=0.9.15,<0.10.0"]
+build-backend = "uv_build"
 
 [project]
 name = "dscript"
-dynamic = ["version"]
+version = "0.3.1"
 description = "D-SCRIPT: protein-protein interaction prediction"
 authors = [
     {name = "Samuel Sledzieski", email = "samsl@mit.edu"}
@@ -43,6 +43,7 @@ docs = [
     "jinja2<3.1",
     "Sphinx==3.4",
     "sphinx-rtd-theme==1.0.0",
+    "pydata-sphinx-theme",
     "sphinxcontrib-applehelp==1.0.2",
     "sphinxcontrib-devhelp==1.0.2",
     "sphinxcontrib-htmlhelp==1.0.3",
@@ -57,11 +58,14 @@ Homepage = "http://dscript.csail.mit.edu"
 [project.scripts]
 dscript = "dscript.__main__:main"
 
-[tool.setuptools.dynamic]
-version = {attr = "dscript.__version__"}
+[tool.uv]
+package = true
 
-[tool.setuptools.packages.find]
-include = ["dscript*"]
+[tool.uv.sources]
+
+[tool.uv.build-backend]
+module-name = "dscript"
+module-root = ""
 
 [tool.ruff]
 line-length = 90
@@ -89,6 +93,9 @@ known-first-party = ["dscript"]
 [tool.coverage.paths]
 source = ["dscript"]
 
+[tool.coverage.run]
+omit = ["dscript/tests/*"]
+
 [tool.pytest.ini_options]
 filterwarnings = [
     "ignore::UserWarning",