kedro-datasets: Implement plotly.HTMLDataset

[yury-fedotov]

Description

Currently both of existing plotly datasets save a figure as JSON and load from it.

This has a major limitation because viewing them requires running Python to parse it back as plotly object. The only options are kedro viz or catalog.load("figure_dataset").

I think there is value in having a plotly.HTMLDataset that would save figures on disk as HTML so:

• They can be opened without the need to run Python.
• It would also make them shareable with team members who don't have Kedro project setup.
• Using versioned: True in catalog you'd get a convenient folder of HTML files that you'd be able to scroll through without running Python / Kedro.

Context

Sharing charts with teammates is a task I often encounter and it would be nice if I could share plotly charts as HTMLs rather than having to run kedro viz or catalog.load("figure_dataset") and doing a screenshot manually.

Possible Implementation

• See this minimal project where I did a proof-of-concept implementation.
• Dataset implementation is here.