clean up

.github/workflows/test_pr.yml

@@ -37,58 +37,41 @@ jobs:
     strategy:
       matrix:
         python-version: ["3.11"]
-        os: [macos-latest, ubuntu-latest]
+        os: [macos-latest, ubuntu-latest, windows-latest]
     steps:
       - uses: actions/checkout@v3
         with:
           submodules: true
-
       - name: Set up Python ${{ matrix.python-version }}
         uses: actions/setup-python@v4
         with:
           python-version: ${{ matrix.python-version }}
           cache: "pip"
-
-      - name: Install libomp (macOS only)
-        run: |
-          brew install libomp
-        if: ${{ runner.os == 'macOS' }}
-
       - name: Set libomp environment variables (macOS only)
         run: |
+          brew install libomp
           LIBOMP_PATH="/opt/homebrew/opt/libomp"
           echo "LDFLAGS=-L${LIBOMP_PATH}/lib" >> $GITHUB_ENV
           echo "CPPFLAGS=-I${LIBOMP_PATH}/include" >> $GITHUB_ENV
           echo "DYLD_LIBRARY_PATH=${LIBOMP_PATH}/lib:\$DYLD_LIBRARY_PATH" >> $GITHUB_ENV
           echo "OMP_PATH=${LIBOMP_PATH}/include" >> $GITHUB_ENV
         if: ${{ runner.os == 'macOS' }}
-
       - name: Install dependencies
         run: |
           python -m pip install -U pip
           pip install -r prereq.txt
-          pip install pytest-timeout
-
-      - name: Dump GitHub Action environment
-        run: |
-          echo "OS: $(uname -a)"
-          python --version
-          pip freeze
-
       - name: Limit OpenMP threads
         run: |
           echo "OMP_NUM_THREADS=2" >> $GITHUB_ENV
-
-      # - name: Test Core
-      #   run: |
-      #     pip install .[testing]
-      #     pip freeze
-      #     pytest -vvvsx --timeout=2000 --timeout-method=thread -m "not slow" --durations=50
+      - name: Test Core
+        run: |
+          pip install .[testing]
+          pip freeze
+          pytest -vvvsx -m "not slow" --durations=50
       - name: Set macOS deployment target
         run: |
           echo "MACOSX_DEPLOYMENT_TARGET=10.13" >> $GITHUB_ENV
         if: ${{ runner.os == 'macOS' }}
-
       - name: Test GOGGLE
         run: |
           pip install .[testing,goggle]

setup.cfg

@@ -89,7 +89,6 @@ testing =
     nbformat
     pytest-benchmark
     pytest-xdist[psutil]
-    pytest_timeout
     pytest-xprocess
     igraph
     py # pytest 7.2.0 bug https://github.com/pytest-dev/pytest-xprocess/issues/110

src/synthcity/metrics/eval_performance.py

@@ -1066,7 +1066,7 @@ def evaluate(
             model = XGBClassifier(
                 tree_method="approx",
                 n_jobs=2,
-                verbosity=2,
+                verbosity=0,
                 depth=3,
                 random_state=self._random_state,
             )

src/synthcity/utils/compression.py

@@ -48,7 +48,7 @@ def compress_dataset(
             model = model = XGBClassifier(
                 tree_method="approx",
                 n_jobs=2,
-                verbosity=2,
+                verbosity=0,
                 depth=3,
             )
             try:

tests/conftest.py

@@ -21,13 +21,3 @@ def run_before_tests() -> Generator:
     workspace = Path("workspace")
     if workspace.exists():
         shutil.rmtree(workspace, ignore_errors=True)
-
-
-# Hook to modify the test result if it exceeds a timeout
-def pytest_runtest_makereport(item: pytest.Item, call: pytest.CallInfo) -> None:
-    """Modify the test result if it exceeds the timeout to skip instead of failing."""
-    if call.when == "call" and call.excinfo is not None:
-        print(f"Call info: {call}")
-        # Check if the test was stopped due to a timeout using call.result
-        if "Timeout" in str(call.excinfo.value):
-            pytest.skip(f"Test skipped due to exceeding the timeout: {item.nodeid}")

tests/metrics/test_attacks.py

@@ -1,5 +1,4 @@
 # stdlib
-import sys
 from typing import Type
 
 # third party
@@ -16,7 +15,6 @@
 from synthcity.plugins.core.dataloader import GenericDataLoader
 
 
-@pytest.mark.skipif(sys.platform == "darwin", reason="Test skipped on MacOS")
 @pytest.mark.parametrize("reduction", ["mean", "max", "min"])
 @pytest.mark.parametrize(
     "evaluator_t",
@@ -54,10 +52,6 @@ def test_reduction(reduction: str, evaluator_t: Type) -> None:
     assert def_score == score[reduction]
 
 
-@pytest.mark.skipif(
-    sys.platform == "darwin",
-    reason="Test skipped on MacOS",
-)
 @pytest.mark.parametrize(
     "evaluator_t",
     [

tests/metrics/test_detection.py

@@ -1,7 +1,5 @@
 # stdlib
-import os
 import sys
-from pathlib import Path
 from typing import Type
 
 # third party
@@ -173,17 +171,8 @@ def test_detect_synth_timeseries(test_plugin: Plugin, evaluator_t: Type) -> None
 @pytest.mark.slow_1
 @pytest.mark.slow
 def test_image_support_detection() -> None:
-    # Get the MNIST dataset directory from an environment variable
-    mnist_dir = os.getenv(
-        "MNIST_DATA_DIR", "."
-    )  # Default to current directory if not set
-
-    # Check if the MNIST dataset is already downloaded
-    mnist_path = Path(mnist_dir) / "MNIST" / "processed"
-    if not mnist_path.exists():
-        dataset = datasets.MNIST(mnist_dir, download=True)
-    else:
-        dataset = datasets.MNIST(mnist_dir, train=True)
+
+    dataset = datasets.MNIST(".", download=True)
 
     X1 = ImageDataLoader(dataset).sample(100)
     X2 = ImageDataLoader(dataset).sample(100)

tests/metrics/test_performance.py

@@ -1,7 +1,5 @@
 # stdlib
-import os
 import sys
-from pathlib import Path
 from typing import Optional, Type
 
 # third party
@@ -484,17 +482,8 @@ def test_evaluate_performance_time_series_survival(
 @pytest.mark.slow_1
 @pytest.mark.slow
 def test_image_support_perf() -> None:
-    # Get the MNIST dataset directory from an environment variable
-    mnist_dir = os.getenv(
-        "MNIST_DATA_DIR", "."
-    )  # Default to current directory if not set
-
-    # Check if the MNIST dataset is already downloaded
-    mnist_path = Path(mnist_dir) / "MNIST" / "processed"
-    if not mnist_path.exists():
-        dataset = datasets.MNIST(mnist_dir, download=True)
-    else:
-        dataset = datasets.MNIST(mnist_dir, train=True)
+
+    dataset = datasets.MNIST(".", download=True)
 
     X1 = ImageDataLoader(dataset).sample(100)
     X2 = ImageDataLoader(dataset).sample(100)

tests/metrics/test_privacy.py

@@ -1,7 +1,5 @@
 # stdlib
-import os
 import sys
-from pathlib import Path
 from typing import Type
 
 # third party
@@ -86,17 +84,8 @@ def test_evaluator(evaluator_t: Type, test_plugin: Plugin) -> None:
 
 @pytest.mark.skipif(sys.platform != "linux", reason="Linux only for faster results")
 def test_image_support() -> None:
-    # Get the MNIST dataset directory from an environment variable
-    mnist_dir = os.getenv(
-        "MNIST_DATA_DIR", "."
-    )  # Default to current directory if not set
 
-    # Check if the MNIST dataset is already downloaded
-    mnist_path = Path(mnist_dir) / "MNIST" / "processed"
-    if not mnist_path.exists():
-        dataset = datasets.MNIST(mnist_dir, download=True)
-    else:
-        dataset = datasets.MNIST(mnist_dir, train=True)
+    dataset = datasets.MNIST(".", download=True)
 
     X1 = ImageDataLoader(dataset).sample(100)
     X2 = ImageDataLoader(dataset).sample(100)

tests/metrics/test_sanity.py

@@ -1,7 +1,5 @@
 # stdlib
-import os
 import sys
-from pathlib import Path
 from typing import Callable, Tuple
 
 # third party
@@ -199,17 +197,8 @@ def test_evaluate_distant_values(test_plugin: Plugin) -> None:
 
 @pytest.mark.skipif(sys.platform != "linux", reason="Linux only for faster results")
 def test_image_support() -> None:
-    # Get the MNIST dataset directory from an environment variable
-    mnist_dir = os.getenv(
-        "MNIST_DATA_DIR", "."
-    )  # Default to current directory if not set
-
-    # Check if the MNIST dataset is already downloaded
-    mnist_path = Path(mnist_dir) / "MNIST" / "processed"
-    if not mnist_path.exists():
-        dataset = datasets.MNIST(mnist_dir, download=True)
-    else:
-        dataset = datasets.MNIST(mnist_dir, train=True)
+
+    dataset = datasets.MNIST(".", download=True)
 
     X1 = ImageDataLoader(dataset).sample(100)
     X2 = ImageDataLoader(dataset).sample(100)

tests/metrics/test_statistical.py

@@ -1,7 +1,5 @@
 # stdlib
-import os
 import sys
-from pathlib import Path
 from typing import Any, Tuple, Type
 
 # third party
@@ -289,17 +287,8 @@ def test_evaluate_survival_km_distance(test_plugin: Plugin) -> None:
 
 @pytest.mark.skipif(sys.platform != "linux", reason="Linux only for faster results")
 def test_image_support() -> None:
-    # Get the MNIST dataset directory from an environment variable
-    mnist_dir = os.getenv(
-        "MNIST_DATA_DIR", "."
-    )  # Default to current directory if not set
-
-    # Check if the MNIST dataset is already downloaded
-    mnist_path = Path(mnist_dir) / "MNIST" / "processed"
-    if not mnist_path.exists():
-        dataset = datasets.MNIST(mnist_dir, download=True)
-    else:
-        dataset = datasets.MNIST(mnist_dir, train=True)
+
+    dataset = datasets.MNIST(".", download=True)
 
     X1 = ImageDataLoader(dataset).sample(100)
     X2 = ImageDataLoader(dataset).sample(100)

tests/plugins/core/models/test_convnet.py

@@ -1,7 +1,3 @@
-# stdlib
-import os
-from pathlib import Path
-
 # third party
 import numpy as np
 import pytest
@@ -67,17 +63,7 @@ def test_train_clf() -> None:
             transforms.Normalize(mean=(0.5,), std=(0.5,)),
         ]
     )
-    # Get the MNIST dataset directory from an environment variable
-    mnist_dir = os.getenv(
-        "MNIST_DATA_DIR", "."
-    )  # Default to current directory if not set
-
-    # Check if the MNIST dataset is already downloaded
-    mnist_path = Path(mnist_dir) / "MNIST" / "processed"
-    if not mnist_path.exists():
-        dataset = datasets.MNIST(mnist_dir, download=True, transform=data_transform)
-    else:
-        dataset = datasets.MNIST(mnist_dir, train=True, transform=data_transform)
+    dataset = datasets.MNIST(".", download=True, transform=data_transform)
     dataset = Subset(dataset, np.arange(len(dataset))[:100])
 
     classes = 10

tests/plugins/core/models/test_image_gan.py

@@ -1,7 +1,3 @@
-# stdlib
-import os
-from pathlib import Path
-
 # third party
 import numpy as np
 import pytest
@@ -30,15 +26,7 @@
 # Load MNIST dataset as tensors
 batch_size = 128
 
-# Get the MNIST dataset directory from an environment variable
-mnist_dir = os.getenv("MNIST_DATA_DIR", ".")  # Default to current directory if not set
-
-# Check if the MNIST dataset is already downloaded
-mnist_path = Path(mnist_dir) / "MNIST" / "processed"
-if not mnist_path.exists():
-    dataset = datasets.MNIST(mnist_dir, download=True, transform=data_transform)
-else:
-    dataset = datasets.MNIST(mnist_dir, train=True, transform=data_transform)
+dataset = datasets.MNIST(".", download=True, transform=data_transform)
 
 dataset = Subset(dataset, np.arange(len(dataset))[:100])
 dataset = FlexibleDataset(dataset)

tests/plugins/core/test_dataloader.py

@@ -1,8 +1,6 @@
 # stdlib
-import os
 import sys
 from datetime import datetime
-from pathlib import Path
 from typing import Any
 
 # third party
@@ -642,17 +640,8 @@ def test_time_series_survival_pack_unpack_padding(as_numpy: bool) -> None:
 @pytest.mark.parametrize("height", [55, 64])
 @pytest.mark.parametrize("width", [32, 22])
 def test_image_dataloader_sanity(height: int, width: int) -> None:
-    # Get the MNIST dataset directory from an environment variable
-    mnist_dir = os.getenv(
-        "MNIST_DATA_DIR", "."
-    )  # Default to current directory if not set
-
-    # Check if the MNIST dataset is already downloaded
-    mnist_path = Path(mnist_dir) / "MNIST" / "processed"
-    if not mnist_path.exists():
-        dataset = datasets.MNIST(mnist_dir, download=True)
-    else:
-        dataset = datasets.MNIST(mnist_dir, train=True)
+
+    dataset = datasets.MNIST(".", download=True)
 
     loader = ImageDataLoader(
         data=dataset,
@@ -693,17 +682,8 @@ def test_image_dataloader_sanity(height: int, width: int) -> None:
 
 @pytest.mark.skipif(sys.platform != "linux", reason="Linux only for faster results")
 def test_image_dataloader_create_from_info() -> None:
-    # Get the MNIST dataset directory from an environment variable
-    mnist_dir = os.getenv(
-        "MNIST_DATA_DIR", "."
-    )  # Default to current directory if not set
-
-    # Check if the MNIST dataset is already downloaded
-    mnist_path = Path(mnist_dir) / "MNIST" / "processed"
-    if not mnist_path.exists():
-        dataset = datasets.MNIST(mnist_dir, download=True)
-    else:
-        dataset = datasets.MNIST(mnist_dir, train=True)
+
+    dataset = datasets.MNIST(".", download=True)
 
     loader = ImageDataLoader(
         data=dataset,