Make datasette-lite a Python package
rclement opened this issue · 1 comments
It would be nice for datasette-lite
to be distributed as a Python package like jupyterlite
in order to build a datasette-lite
website distribution, instead of having to fork this repository. This could even be Datasette plugin, in the vein of datasette-publish-vercel
, providing a command to build the static website.
For instance, a possible workflow:
pip install datasette datasette-lite
datasette lite build
- Deploy on static hosting
Some nice to have features that could be integrated with the build command:
datasette lite build --output dir
: specify the output directory:datasette lite build --install datasette-dashboards
: pre-installing additional dependencies (without having to use theinstall
query argument) by templatingwebworker.js
to include additionalmicropip.install
instructions
The build command could then be responsible for generating at least the following files:
index.html
webworker.js
Having datasette-lite
as a Python package to build a distribution could also be used for advanced use-cases.
For instance, embedding a Datasette-Lite distribution within an app or a static website (MkDocs/Sphinx/JupyterBook/Whatever). I'm doing this with MkDocs + JupyterLite and it works great. Would love to be able to do it with Datasette-Lite!