-
Notifications
You must be signed in to change notification settings - Fork 16
Description
Checks
- I have checked that this enhancement has not already been requested
How would you categorize this request. You can select multiple if not sure
Other (please describe in Issue Description)
Enhancement Description
Currently to get the real buckaroo experience you need the notebook or a tornado server. This is a heavyweight ask for embedding. For many usecases notably streamlit and xorq integration a simpler experience is desired. This will be a static only buckaroo expreience
Note, initially search, sort, postprocessing functions and the lwocode ui wont work on this experience. search and sort are probably easy to add back in as JS only features.
Pseudo Code Implementation
the user should call prepare_buckaroo_artifact(df:pd.DataFrame|pl.DataFrame|Path('parquet_file')) which produces a single json or b64 encoded string to be served to buckaroo by the frontend.
on the frontend you should be able to call a react component with
<BuckarooTable config={...} artifact={prepared_artifact}/>
Prior Art
I'll just use this for questions.
I have a PR exploring apache flechette and removing hyparquet. #546
I would imagine that parquet is a notably better storage format than flechette, but this needs to be proven out.
Summary stats need to be packaged into the same object.