Rapid

This repo contains code, data and jupyter notebook related to RAPID.

Click this button to access the jupyter notebook without downloading/installing

If the above link does not launch. Try an alternative server:

Repository contents

The following sections indicate the folders which contain code and related data

All raw data files are located in the data folder

cifs - Contains the Crystallographic Information Files
images - Contains side vial images of each experiment performed
xrd/xy - Contains xy files for XRD data
0042.perovskitedata_RAPID.csv - Escalate generated data file including 8 experimental features (with "rxn" as header prefix) and 67 chemical features (with "feat" as header prefix). The detailed explanations of these features are listed in "Explanation of Features-Descriptors" section in "Perovskite Dataset Description.pdf". This CSV file is used in visualization and machine learning.
0042.perovskitedata_RAPID_full.csv - This escalate generated data file contains the same experiments as "0042.perovskitedata_RAPID.csv" but has all 787 features, including additional "raw" features describing experiment details (see "Explanation of Features-Descriptors" section in "Perovskite Dataset Description" for the explanations of "raw" prefix). The csv file is not used for visualization or machine learning.
image_list.json - Keeps track of all image files in the image folder
ml_data.pkl - Python pickle file containing ML results
inventory.csv - Chemical inventory data
organic_inchikey.csv - Inchi keys and chemical names
s_spaces.json - Co-ordinates of state space for each amine

The following python scripts are used in the RAPID.ipynb notebook to generate visualizations

plots.py - Generates the reaction outcomes 3D plot widget
xrd_plot.py - Generates the xrd plot widget
ml_section.py - Generates machine learning outcomes widget
cif_plots.py - Generates the cif plot widget. Note that jsmol is used to create the widget