Core Technologies
LIVE Environments are made possible by core open-source libraries developed and maintained by core collaborators within the LIVE project, as well as components of the open source Python and javascript data science stack.
glue: multi-dimensional linked-data exploration
The glue library provides a central hub for: (1) providing access to data sets; (2) managing the conceptual links between datasets; and (3) propagating subsets across datasets and viewers.
The well-tested data framework within glue allows it to load datasets in a diverse range of formats, including remote datasets and data sets that are too large to load into memory.
The ability to define conceptual links of arbitrary complexity is key to modern data analysis, where it is critical to be able to combine multiple datasets for analysis but straightforward merging of datasets is not desirable because files are large or because there are multiple interesting ways to combine datasets. The glue library provides the architecture for creating and managing these links.
glue provides for the brushing and linking paradigm that empowers exploratory data analysis by making subsets selected in one viewer propagate automatically to all other relevant viewers and even other (linked) datasets.
JupyterLab: A Next-Generation Notebook Interface
By building on Jupyter components we will meet researchers in their favorite environment for exploring and analyzing their data while also gaining the ability to easily export this analysis to the web. Through Jupyter Server and related infrastructure, LIVE environments will be deployable across a broad range of computational environments. Developments such as JupyterLite (allowing Python kernels to run fully inside a user’s web browser) and new frameworks like Solara (a novel architecture for running Jupyter notebooks as web apps) have greatly expanded the range of deployment solutions available for hosting relatively simple visualization environments seamlessly on the web.
Solara: A pure Python, React-style web framework
Solara is a new Python-based React-style web framework for bulding applications with Jupyter notebooks or a standalone web apps. The accompanying video shows the presentation of Solara at the 2023 JupyterCon.
The two quick demo videos below were created by collaborator Maarten Breddels in the Fall of 2023, as an early demonstration of the power of Solara to create Javascript elements from python sources, in this case the glue-jupyter viewers.
glue-solarathon.movglupyter, in Solara, with tabbed or MDI interface
glue-solarathon-jupyter2.movglupyter, in Solara, inside JupyterLab, showing interaction with Jupyter Notebook