Web-based chemistry databases have long been maintained by a few large centralized providers, and offered to the community largely as read-only collections. Research groups and institutions had the option of putting together something bespoke for their data, contributing to some of the centralized repositories, or using ad-hoc data management techniques locally. These often entailed the ad-hoc use of the file system as a simple database, and relied on group knowledge to communicate data/results. As computational resources grow even for relatively small groups the importance of democratizing web-based chemistry databases grows for both internal use and sharing results generated by groups more widely.<br/><br/>The development of an open source web-based database with JupyterLab for computational chemistry will be described. The core is implemented in the Python programming language, using MongoDB for metadata storage, and file system abstractions for large data storage. It features a RESTful API, a modern HTML5 React-based single page web application, and extensions to JupyterLab. It can run a number of quantum codes such as NWChem, PSI4, and more under development in addition to trained machine learning models. The Jupyter notebook can trigger computational jobs, with automated ingestion of data into the web-based database. The single page application provides access to individual results, including 3D visualization using WebGL-based rendering techniques. Public data can be searched, and authenticated users can do more. It also integrates with a number of web-based chemistry databases, and aims to provide another option for offering data in increasingly data-centric workflows on the web.

Open chemistry: Democratizing web-based chemistry databases

2019 Fall National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.