Requirements

• Python >= 3.6
• Pytorch >= 1.7

Folder Specification

• ddi_main.py: Please excuate this file to run our DDI model.
• ddi_predictor5.py: It includes our main architecture code.
• Dataset This folder includes deep_drug_dict.json containing all dictionary lists of drugs, deep_train.json, deep_valid.json, and deep_test.json, including annotation and unannotated relationship information between pair drugs.
• Utils: This folder includes ddi_get_mol_graphs.py responsible for transforming drug SMOLES sequence data into drug moleculers and ddi_dataset_wo_type.py splitting datasets into train, validated and test datasets for DDI.
• Utils: This folder includes ddi_layers.py containing import representation learning blocks and ddi_train.py optimizing framework via chi-square.

Run the Code

To excecute this framework, please run the following command. The contents include train and inference:

cd cd
python ddi_main.py

Acknowledgement

We sincerely thank Weiyu Shi for providing code. Please contact us with Email: standyshi@qq.com

If you are interested in Natural Language Processing and Large Language Models, feel free to contact us by Email: zhangyijia@dlmu.edu.cn