Thsi repository contains developing code base for impurity prediction of reactions. At the moment, users need to have installed conda, anaconda or miniconda. Please refer to Reaction Impurity Prediction using a Data Mining Approach for more details.
Navigate to local project directory and:
git clone https://github.com/sustainable-processes/Impurity-Project.git
OR
git clone git@github.com:sustainable-processes/Impurity-Project.git if SSH is configured
To install dependencies, create a conda environment or virtual environment using python 3.9.13. Once the environment is activated, the requirements.txt file can be used:
pip install -r requirements.txt
MainScript.py contains and imports necessary code to run all workflow steps. This is invoked in the tutorial notebooks which the user can go through as explained below.
Two tutorial files are given under the Tutorial_Final folder. Tutorial-Jupyter Notebook_final.html shows the workflow for the paracetamol case study (predicting impurities in paracetamol synthesis) with widgets and results visualized in an html file. It is advised to download this github repository as a .zip, and open Tutorial-Jupyter Notebook_final.html in Google Chrome for best results. Tutorial-Jupyter Notebook_final.ipynb is a jupyter notebook with the same workflow, but with no widget results shown (due to storage space limitations).