Repository to train, evaluate, and use models for synthesis predictions.
This contains re-factored code previously found in the following repositories
Before you begin, ensure you have met the following requirements:
-
Linux OS (or in principle Windows or macOS).
-
You have installed anaconda or miniconda with python 3.8-3.10
The tool has been developed and fully tested on a Linux platform.
First clone the repository using Git.
The project dependencies can be installed by executing the following commands in the root of the repository:
conda env create -f env-dev.yml
conda activate aizynthmodels
poetry install
If there is an error "ImportError: /lib64/libstdc++.so.6: version `GLIBCXX_3.4.21' not found" it can be mitigated by adding the 'lib' directory from the Conda environment to LD_LIBRARY_PATH
As example:
export LD_LIBRARY_PATH=/path/to/your/conda/envs/chemformer/lib
Finally, rxnutils should be installed. Either clone the reaction_utils repo and install rxnutils in the current (aizynthmodels) environment:
cd ../reaction_utils
poetry install --all-extras
or install it from pypi
python -m pip install reaction-utils
This repository uses pre-commit (https://pre-commit.com) to ensure style consistency. Before you start developing, ensure that pre-commit is installed and hooks are initialised as described in https://pre-commit.com/#install.
pre-commit install
Tests uses the pytest package, and is installed by poetry
Run the tests using:
pytest -v
We welcome contributions, in the form of issues or pull requests.
If you have a question or want to report a bug, please submit an issue.
To contribute with code to the project, follow these steps:
- Fork this repository.
- Create a branch:
git checkout -b <branch_name>. - Make your changes and commit them:
git commit -m '<commit_message>' - Push to the remote branch:
git push - Create the pull request.
Please use pre-commit to ensure proper formatting and linting
The contributors have limited time for support questions, but please do not hesitate to submit an issue (see above).
The software is licensed under the Apache 2.0 license (see LICENSE file), and is free and provided as-is.