updating linking order

Author: Andrewq11

.github/workflows/test.yml

@@ -11,13 +11,16 @@ on:
   schedule:
     - cron: "0 4 * * *"
 
+env:
+  ENV_FILE: "env.yml"
+
 jobs:
   test:
     strategy:
       fail-fast: false
       matrix:
-        os: [windows-latest] # ubuntu-latest, macos-latest,
-        python-version: ["3.10", "3.11"] # -> Will re-enable support for py312 once pyg is released
+        os: [windows-latest] # ubuntu-latest, macos-latest
+        python-version: ["3.11"] # -> Will re-enable support for py312 once pyg is released, "3.10",
 
     runs-on: ${{ matrix.os }}
     timeout-minutes: 30
@@ -27,29 +30,32 @@ jobs:
         shell: bash -l {0}
 
     name: |
-        regular_env -
-        python=${{ matrix.python-version }} -
-        os=${{ matrix.os }}
+      regular_env -
+      python=${{ matrix.python-version }} -
+      os=${{ matrix.os }}
 
     steps:
       - name: Checkout the code
         uses: actions/checkout@v3
+      
+      - name: Set OS-specific environment file to use
+        run: |
+          if [ "${{ matrix.os }}" == "windows-latest" ]; then
+            echo "ENV_FILE=env_windows.yml" >> $GITHUB_ENV;
+          fi
 
       - name: Setup mamba
         uses: mamba-org/setup-micromamba@v1
         with:
-          environment-file: env.yml
+          environment-file: ${{ env.ENV_FILE }}
           environment-name: graphium
           cache-environment: true
           cache-downloads: true
           create-args: >-
             python=${{ matrix.python-version }}
-          condarc: |
-            - conda-forge
-            - pytorch
 
       - name: Install library
-        run: python -m pip install --no-deps --no-build-isolation -e . # `-e` required for correct `coverage` run.
+        run: python -m pip install --no-deps --no-build-isolation -e . -v # `-e` required for correct `coverage` run.
 
       - name: Run tests
         run: pytest

env.yml

@@ -1,7 +1,5 @@
 channels:
   - conda-forge
-  - pytorch
-  # - pyg # Add for Windows
 
 dependencies:
   - python >3.9,<3.12
@@ -35,7 +33,7 @@ dependencies:
   - lightning >=2.0
   - torchmetrics
   - ogb
-  - pytorch_geometric >=2.0 # Use `pyg` for Windows instead of `pytorch_geometric`
+  - pytorch_geometric >=2.0
   - wandb >=0.18.5
   - mup
   - pytorch_sparse >=0.6

env_windows.yml

@@ -0,0 +1,86 @@
+channels:
+  - conda-forge
+  - pytorch
+  - pyg 
+
+dependencies:
+  - python >3.9,<3.12
+  - pip
+  - typer
+  - loguru
+  - omegaconf >=2.0.0
+  - hydra-core >=1.3.2
+  - tqdm
+  - platformdirs
+
+  # scientific
+  - numpy <=1.25.0
+  - scipy >=1.4
+  - pandas >=1.0
+  - scikit-learn
+  - fastparquet
+
+  # viz
+  - matplotlib >=3.0.1
+  - seaborn
+
+  # cloud IO
+  - fsspec >=2021.6
+  - s3fs >=2021.6
+  - gcsfs >=2021.6
+
+  # ML packages
+  - cuda-version # works also with CPU-only system.
+  - pytorch >=1.12,<2.5
+  - lightning >=2.0
+  - torchmetrics
+  - ogb
+  - pyg >=2.0
+  - wandb >=0.18.5
+  - mup
+  - pytorch-sparse >=0.6
+  - pytorch-cluster >=1.5
+  - pytorch-scatter >=2.0
+
+  # chemistry
+  - rdkit <=2024.03.4
+  - datamol >=0.10
+  - boost # needed by rdkit
+
+  # Optional deps
+  - sympy
+  - tensorboard
+  - pydantic <2  # because of lightning. See https://github.com/Lightning-AI/lightning/issues/18026 and https://github.com/Lightning-AI/lightning/pull/18022
+
+  # Dev
+  - pytest >=6.0
+  - pytest-xdist
+  - pytest-cov
+  - pytest-forked
+  - nbconvert
+  - black >=23
+  - jupyterlab
+  - ipywidgets
+
+  # Doc
+  - mkdocs
+  - mkdocs-material
+  - mkdocs-material-extensions
+  - mkdocstrings
+  - mkdocstrings-python
+  - mkdocs-jupyter
+  - markdown-include
+  - mike >=1.0.0
+  - doxygen
+
+  # graphium_cpp build dependencies
+  - pybind11
+  - librdkit-dev 
+
+  # Optional
+  - pytdc
+
+  - pip:
+      # Build deps
+      - setuptools-scm
+      - build

graphium/graphium_cpp/features.h

@@ -195,8 +195,8 @@ template<> struct FeatureValues<double> {
         static_assert(std::is_floating_point_v<T>);
         return double(inputType);
     }
-
-    static constexpr bool is_finite(double v) {
+    
+    static bool is_finite(double v) {
         return std::isfinite(v);
     }
 

graphium/graphium_cpp/labels.cpp

@@ -7,13 +7,14 @@
 
 
 #include "labels.h"
-
 #include "features.h"
 
 // C++ standard library headers
 #include <filesystem>
 #include <thread>
 #include <unordered_map>
+#define NOMINMAX // Disabling MVSC-defined min/max macros
+#include <algorithm>
 
 // RDKit headers
 #include <GraphMol/ROMol.h>

setup.py

@@ -19,22 +19,33 @@
 
 python_version = str(sys.version_info[0]) + str(sys.version_info[1])
 
-torch_dir = torch.__path__[0]
-rdkit_lib_index = rdkit.__path__[0].split("/").index("lib")
-rdkit_prefix = "/".join(rdkit.__path__[0].split("/")[:rdkit_lib_index])
+# Base variables required for compilation
+path_separator = "/"
+lib_folder_name = "lib"
+boost_include = "include/boost"
+package_compile_args = []
 
+# Updating variables used during compilation based on OS
 system = platform.system()
-package_compile_args = [
-    "-O3",
-    "-Wall",
-    "-Wmissing-field-initializers",
-    "-Wmaybe-uninitialized",
-    "-Wuninitialized",
-]
+if system == "Darwin" or system == "Linux":
+    package_compile_args += ["-O3", "-Wall", "-Wmaybe-uninitialized", "-Wuninitialized"]
+
 if system == "Darwin":
-    package_compile_args.append("-mmacosx-version-min=10.15")
+    package_compile_args += ["-mmacosx-version-min=10.15"]
 elif system == "Windows":
-    pass
+    path_separator = "\\"
+    lib_folder_name = "Lib"
+    package_compile_args += ["/Wall", "/O3"]
+
+# Extracting paths to torch and rdkit dependencies
+torch_dir = torch.__path__[0]
+rdkit_lib_index = rdkit.__path__[0].split(path_separator).index(lib_folder_name)  # breaks on windows
+rdkit_prefix = "/".join(rdkit.__path__[0].split(path_separator)[:rdkit_lib_index])
+
+# Windows-specific changed to rdkit path
+if system == "Windows":
+    rdkit_prefix += "/Library"
+    boost_include = "include"
 
 ext_modules = [
     Pybind11Extension(
@@ -57,10 +68,11 @@
             os.path.join(torch_dir, "include"),
             os.path.join(torch_dir, "include/torch/csrc/api/include"),
             os.path.join(rdkit_prefix, "include/rdkit"),
-            os.path.join(rdkit_prefix, "include/boost"),
+            os.path.join(rdkit_prefix, boost_include),
             numpy.get_include(),
         ],
         libraries=[
+            "RDKitRDGeneral",
             "RDKitAlignment",
             "RDKitDataStructs",
             "RDKitDistGeometry",
@@ -73,13 +85,13 @@
             "RDKitInchi",
             "RDKitRDInchiLib",
             "RDKitRDBoost",
-            "RDKitRDGeneral",
             "RDKitRDGeometryLib",
             "RDKitRingDecomposerLib",
             "RDKitSmilesParse",
             "RDKitSubstructMatch",
             "torch_cpu",
             "torch_python",
+            "c10",
             f"boost_python{python_version}",
         ],
         library_dirs=[os.path.join(rdkit_prefix, "lib"), os.path.join(torch_dir, "lib")],

tests/test_datamodule.py

@@ -80,7 +80,9 @@ def test_ogb_datamodule(self):
         # test module
         assert ds.num_edge_feats == 5
         assert ds.num_node_feats == 50
-        assert len(ds) == 642
+        assert (
+            len(ds) == 642 or len(ds) == 644
+        )  # Accounting for differences in csv file reads across Linux & OSX
 
         # test batch loader
         batch = next(iter(ds.train_dataloader()))
@@ -183,7 +185,9 @@ def test_caching(self):
         # test module
         assert ds.num_edge_feats == 5
         assert ds.num_node_feats == 50
-        assert len(ds) == 642
+        assert (
+            len(ds) == 642 or len(ds) == 644
+        )  # Accounting for differences in csv file reads across Linux & OSX
 
         # test batch loader
         batch = next(iter(ds.train_dataloader()))