Docker for reproducing machine learning solubility models.

To run model please install docker.

docker run --rm -v /PATH/TO_FILE/YOU_WANT_TO_WORK_ON/:/data rodrigozepeda/docker-solubility

where:

• /PATH/TO_FILE/YOU_WANT_TO_WORK_ON/ is substituted by path to the csv file conaining the Smiles you want to predict (see To_predict.csv on Github for an example).
• Included models can be specified as:
  • GraphConv
  • Weave
  • MPNN
  • DAG
  • RandomForest Random forest model.
  • KRR
  • XGBoost
  • TextCNN WORKS BUT THROWS ERROR

As an example, assuming the files to predict are included in ~/Dropbox/predict_files and you want the Graph Convolution model you can:

sudo docker run --rm -v ~/Dropbox/predict_files:/data docker-solubility GraphConv

See manual for further instructions