A SQLModel-based repository dedicated to helping users access the PostgreSQL-based qm9star database more easily in a Python environment.

This repository requires a reasonably deployed QM9star database to run. For information on how to deploy a QM9star database, please refer to the download_and_deploy_qm9star.

Once you have deployed correctly and checked network connectivity, you can refer to the query_example to run queries.

git clone http://10.72.201.58:13000/tmj/qm9star_query.git
cd qm9star_query
pip install poetry # if you don't have poetry installed
poetry install

poetry install -E dl

If your cuda version not matches 12.1, you need to remove the dependencies of pyg-lib and other related packages. And then install them again with the correct cuda version. You can find the whl list on pyg-lib

poetry remove pyg-lib torch-scatter torch-sparse torch-cluster torch-spline-conv torch-geometric
poetry add https://data.pyg.org/whl/torch-2.3.0%2Bcu121.html # Use correct cuda version

Note: If you excute poetry install without --E dl again, the extra dependencies will be removed.

• DIG dependency conflict

The DIG package contains a dependency called rdkit-pypi, which may overwrite the original higher version of rdkit when installing, if you encounter this problem, you can use the following methods to solve it.

poetry remove rdkit
poetry add rdkit

Poetry will consider these to be two packages, so it will check that they both exist, but we want the higher version of rdkit to override the lower one.

• Poetry stuck at pending

Sometimes Poetry's dependency resolution gets stuck in hellish wait times, which can be due to network or other reasons. In this case, you can try installing the dependency directly using the following command

pip install .
pip install .[dl] # If you want to use the dl functions