pfrstg / smu

Small Molecule Universe

SMU

This repo was forked (not in the git sense of fork, just a copy of the files) from [https://github.com/google-research/google-research/tree/master/smu] then modified to be a more standard python package.

This code is for a still in-progress paper and dataset. More details will be added when the dataset is complete and published.

Get the data

TODO: once the dataset has been released, include links and md5sums

Installation instructions (user)

This is probably the process you want. even if you are writing new python code using the SMU library, you can use this method. It is only if you need to modify the code in the SMU library itself that you need to use the developer mode below

1. Check that you have a compatible version of python

   python3 --version
   

   Should give you a version >=3.6.1. If not, you need to upgrade.

2. Go to directory where you want to do the install

3. Create a python virtual environment

   python3 -m venv smu_venv
   

4. Activate the virtual environment

   source smu_venv/bin/activate
   # Ensure that commands below refer to the virtual enviroment
   rehash
   

5. Install the SMU code.

   # Ensure pip is latest version
   python3 -m pip install --upgrade pip
   # Note that for now we are using test pypi. We'll release to the regular PyPi
   # for final release
   python3 -m pip install --extra-index-url https://test.pypi.org/simple/ smu
   

6. Test your installation

   cd <path where you downloaded the SMU sqlite files>
   python3 -m smu.query_sqlite --input_sqlite 20220621_standard_v4.sqlite --smiles NN
   

TODO: Change to the normal PyPi

Installation instructions (developer)

Are you really sure you need to do this? See previous section.

TODO: write instructions for the new setup

Next steps

1. The documentation for smu.query_sqlite explains how to use the command line tool to read the SMU database. This is especially useful to extract small parts of the database, but can also be used to read the entire thing.

2. The examples directory gives several example scripts for accessing the database with python code. Even if you don't plan on writing python, these examples illustrate some important features of the data, so it is recommended you read them over.