Recently, there has been a growing interest in ML, which holds great promise to transform the drug discovery industry. Unfortunately, the primary barriers to working in this field for machine learning researchers include a lack of domain expertise, pipelines for data pre-processing, and uniform benchmarks. Knowledge graph reasoning, RF learning, deep generative models, and GML are some of the cutting-edge methods used for these tasks. A hierarchical interface on Torch Drug allows both beginners and specialists in this field to customize the product. In the field of modern medicine, precision medicine is expanding quickly, and the effective creation of standardized and automated patient data analysis is expected to depend heavily on open-source machine learning programs. A primary objective of precision cancer medicine accurately anticipate which pharmacological treatments would work best for each patient based on the genetic profiles of their tumour. Here, we describe an open-source software platform that predicts personalized pharmaceutical responses based on gene expression patterns and highly flexible SVM algorithm with classic recursive feature elimination (RFE) technique. This article describes the data required for the hybrid whale optimization project using decision tree WODT prediction, together with an extensive inventory of databases and machine learning techniques that have been suggested and used to forecast DTIs. There is also a brief discussion of the benefits and drawbacks of each set of techniques. In conclusion, we address the difficulties that may arise when predicting DTI by machine learning techniques and provide insight into significant avenues for future research.