Skip to content

Installation

ALineMol supports Python 3.9, 3.10, and 3.11. The default install is lean (splitter API + SplitAnalyzer only); heavier components are opt-in extras.

uv is a fast Python package installer and resolver.

git clone https://github.com/HFooladi/ALineMol.git
cd ALineMol

# CPU (default)
./install.sh

# Or with CUDA support
./install.sh cu121   # CUDA 12.1
./install.sh cu118   # CUDA 11.8
./install.sh cu124   # CUDA 12.4

source .venv/bin/activate

# Install uv
curl -LsSf https://astral.sh/uv/install.sh | sh

git clone https://github.com/HFooladi/ALineMol.git
cd ALineMol

uv venv --python 3.11
source .venv/bin/activate

# CPU
uv pip install -e ".[all]" \
  -f https://download.pytorch.org/whl/cpu \
  -f https://data.dgl.ai/wheels/repo.html

# Or CUDA 12.1
uv pip install -e ".[all]" \
  -f https://download.pytorch.org/whl/cu121 \
  -f https://data.dgl.ai/wheels/repo.html

Using conda

git clone https://github.com/HFooladi/ALineMol.git
cd ALineMol

conda env create -f environment.yml
conda activate alinemol
pip install --no-deps -e .

Optional extras

The lean base install ships only the splitter API and SplitAnalyzer. Pull in heavier components on demand:

Extra Pulls in When you need it
[gnn] torch, dgl, dgllife, torch-geometric Training/inference with GNN models
[ml] statsmodels, POT, astartes alinemol.utils, OT-based graph utilities, astartes-backed splitters
[datasail] datasail The datasail splitter
[all] gnn + ml + datasail Everything (used by install.sh)
[docs] mkdocs-material, mkdocstrings, mike, … Building this documentation
[dev] ruff, pre-commit, mypy Contributing
[test] pytest, pytest-cov Running the test suite 
# Splitters + GNN training only
uv pip install -e ".[gnn]"

# Everything, plus dev and test tooling
uv pip install -e ".[all,dev,test]"

GNN stack pin

DGL 2.1.0 ships prebuilt GraphBolt binaries only for torch 2.0.0–2.2.1, so the [gnn] extra caps torch at <2.2.2. Set DGL_SKIP_GRAPHBOLT=1 if you hit a GraphBolt import error.

Verify the install

python -c "from alinemol.splitters import get_splitter_names; print(get_splitter_names())"

You should see the list of available splitters. Next, head to the Quickstart.