St.deck_gl_chart parameters

[davideps]

If I understand correctly, you are busy further developing the connection to deck_gl so the API and functionality may be in flux. I noticed that with this code:

layers = [
{“data”: these_cars, “type”: “ScatterplotLayer”, “opacity”:0.1, “getRadius”: 200, “getFillColor”:[0,0,0]},
{“data”: these_pickups, “type”: “ScatterplotLayer”, “opacity”:0.1, “getRadius”:100, “getFillColor”:[50,255,50]},
{“data”: these_dropoffs, “type”: “ScatterplotLayer”, “opacity”:1.0, “getRadius”:50, “getFillColor”:[255,0,0]}
]
st.deck_gl_chart(viewport=viewport, layers=layers)

That “getRadius” has no effect and the last “opacity” setting overrides the setting for all layers. Am I defining the layers incorrectly?

Also, do you intend to support selecting features (items) of layers?

[Jonathan_Rhone]

Hi @davideps,

Looks like you have a number of questions here…

• getRadius: https://streamlit.io/docs/api.html?highlight=getradius#streamlit.deck_gl_chart

you can use a “getRadius” field to individually set the radius of each circle in the plot. So here you would set “getRadius”: “my_column” where “my_column” is the name of the column containing the radius data.

Example usage: https://github.com/streamlit/streamlit/blob/develop/lib/streamlit/hello/demos.py#L79

• opacity: could you post a full code snippet so I can copy/paste and debug your code? Including data for these_cars, etc…

• selecting features/items of layers: apologies but I’m not sure exactly what you mean, could you explain this a bit further?

[davideps]

Thanks @Jonathan_Rhone. By selecting features, I mean allowing users to select one or more dots in a scatterplot (for example) to change their display color, to calculate an average value, or to exclude them from further calculations.

[Jonathan_Rhone]

Hi @davideps,

Selecting features in a chart is not currently supported, but is being discussed/designed internally!

Here are some similar threads from users about this functionality:

Can you select rows in a table?

Hi @gaboc4. Unfortunately that functionality is not yet supported but we are designing it! In the meantime, a workaround would be to use an st.multiselect, e.g.: selected_indices = st.multiselect('Select rows:', data.index) Which will work up to about 1000 rows or so and which will look like this: [image] You can see an example in this gist. This feature request is being tracked here. Please feel free to follow the issue to keep up with the latest developments. Also please feel free to com…

Is it currently possible to select features or draw a selection box on a plot canvas?

@cwerner are you building an image annotation pipeline? As someone that’s building their own, I would love a selectable canvas. Currently I’m on a capped free license on label box. Labelbox provides polygon draw and box draw: [image] What if we had: obj_list = st.canvas_select(img, box = True, polygon = True) Im pretty sure there is even nicer abstraction that can handle a wider variety of interactions. Imagine, for example something like: geometry_list = st.geom_canvas(img, box = True, …

Is it possible to process click events on charts?

I am eager to rewrite a dash app I develop using streamlit. The only bit I am missing is the ability to react to click or select events on a chart, like updating another chart base on the click coordinates. Is it possible to do things like this in streamlit?

Please let us know if you’ll be providing sample code to resolve your opacity issue!

[thiago]

Just wanted to drop in and mention that we’re soon going to deprecate st.deck_gl_chart in favor of st.pydeck_chart.

st.deck_gl_chart uses an API we threw together in a rush in order to support DeckGL (which is a JavaScript library) in Python, and frankly that API is just not that great.

But in the time since we implemented that feature a new Python package called PyDeck has popped up which does a much better job at creating DeckGL specs from Python! So we’ll be migrating to that ASAP.

I expect this to land in Streamlit in a week or two.

[davideps]

Thank you! I’ll get back to the opacity issue in a few days and post if I can’t get it working.