Philipp Rudiger

Would suggest styling that button to be a little nicer and probably position it to the top right of the output (using absolute or relative positioning).

I'd be fine with both, i.e. document it in Parameters and leave a short note in rx docs.

+1 on keeping `.set`

Definitely unintentional, note that you can use HoloViews dim mapping to achieve the same thing: ```python import holoviews as hv color=hv.dim("species").categorize( {"setosa": "red", "virginica": "blue", "versicolor": "green"} ) ```

Yeah, we should patch this asap. Hopefully we can get a fix in a 0.5.3 release soon.

Putting this on hold for a while, I had large parts of this in my stash already and just wanted to open a PR for it since it came up...

Yup, @ahuang11 is correct, it's a dask specific option to persist the data to memory. I think it would make a lot of sense to have a page dedicated to...

Classic notebook has been effectively broken since February for any operation that transfers a lot of data. This happened due to the removal of `nest_asyncio` which itself had other problems....

As noted above if you datashade each line separately this actually already works: ```python c1 = hv.Curve(np.random.randn(10000).cumsum(), vdims='A') c2 = hv.Curve(np.random.randn(10000).cumsum(), vdims='B') (rasterize(c1, aggregator='any', line_width=1) * rasterize(c2, aggregator='any', line_width=1).opts(cmap=['red'])).opts( multi_y=True,...

@hoxbro @MarcSkovMadsen would love your thoughts. I'd like to think about Narwhals support but am wondering if that may be better implemented as a new parameter type.