synrxn 0.0.7

Reaction Database for benchmarking

SynRXN

Reaction Database for Benchmarking SynRXN is a curated, provenance-tracked collection of reaction datasets and evaluation manifests designed for reproducible benchmarking of reaction-informatics tasks (rebalancing, atom-atom mapping, reaction classification, property prediction, and synthesis/retrosynthesis). It provides standardized splits, manifest files (RNG seeds & split indices), and lightweight utilities to load and inspect datasets for fair, reproducible model comparison.

Installation

1. Python Installation: Ensure that Python 3.11 or later is installed on your system. You can download it from python.org.

2. Creating a Virtual Environment (Optional but Recommended): It's recommended to use a virtual environment to avoid conflicts with other projects or system-wide packages. Use the following commands to create and activate a virtual environment:

python -m venv synrxn-env
source synrxn-env/bin/activate  

Or Conda

conda create --name synrxn-env python=3.11
conda activate synrxn-env

3. Install from PyPi: The easiest way to use SynTemp is by installing the PyPI package synrxn.

pip install synrxn

Optional if you want to install full version

pip install synrxn[all]

Example

from synrxn.data import DataLoader
from synrxn.split.repeated_kfold import RepeatedKFoldsSplitter

# 1) Zenodo (stable release)
from pathlib import Path
from synrxn import DataLoader

dl = DataLoader(
    task="classification",
    source="zenodo",
    version="0.0.6",
    cache_dir=Path("~/.cache/synrxn").expanduser(),
)
print(dl.available_names())   # list available datasets
df = dl.load("schneider_b")
print(len(df), df.columns.tolist())

# 2) GitHub release tag
from pathlib import Path
from synrxn.data import DataLoader

dl = DataLoader(
    task="classification",
    source="github",
    version="v0.0.6",
    cache_dir=Path("~/.cache/synrxn").expanduser(),
    gh_enable=True,
)
print(dl.available_names())
df = dl.load("schneider_b")
print(len(df))

# 3) GitHub commit (pin to SHA)
from pathlib import Path
from synrxn.data import DataLoader

dl = DataLoader(
    task="classification",
    source="commit",
    version="3e1612e2199e8b0e369fce3ed9aff3dda68e4c32",
    cache_dir=Path("~/.cache/synrxn").expanduser(),
    gh_enable=True,
)
print(dl.available_names())
df = dl.load("schneider_b")
print(df.head(2))

# 4) GitHub latest
from pathlib import Path
from synrxn.data import DataLoader

dl = DataLoader(
    task="classification",
    source="github",
    version="latest",
    cache_dir=Path("~/.cache/synrxn").expanduser(),
    gh_enable=True,
)
print(dl.available_names())
df = dl.load("schneider_b")
print(df.shape)

# Simple splitting example (property dataset)
from synrxn.data import DataLoader
from synrxn.split.repeated_kfold import RepeatedKFoldsSplitter
from pathlib import Path

dl = DataLoader(
    task="property",
    source="commit",
    version="latest",
    cache_dir=Path("~/.cache/synrxn").expanduser(),
    gh_enable=True,
)
df = dl.load("b97xd3")

splitter = RepeatedKFoldsSplitter(
    n_splits=5, n_repeats=2, ratio=(8,1,1), shuffle=True, random_state=1
)

splitter.prepare_splits(df, stratify=None)           
train_df, val_df, test_df = splitter.get_split(0, 0, as_frame=True)
print(len(train_df), len(val_df), len(test_df))

Contributing

• Tieu-Long Phan

Publication

SynRXN: A Benchmarking Framework and Open Data Repository for Computer-Aided Synthesis Planning

License

This project is licensed under MIT License - see the License file for details.

Acknowledgments

This project has received funding from the European Unions Horizon Europe Doctoral Network programme under the Marie-Skłodowska-Curie grant agreement No 101072930 (TACsy -- Training Alliance for Computational)