feat: Bokeh support as FigureBokeh

solara/lab/components/__init__.py

@@ -1,5 +1,6 @@
 from .chat import ChatBox, ChatInput, ChatMessage  # noqa: F401
 from .confirmation_dialog import ConfirmationDialog  # noqa: F401
+from .figurebokeh import FigureBokeh  # noqa: F401
 from .input_date import InputDate, InputDateRange  # noqa: F401
 from .input_time import InputTime as InputTime
 from .menu import ClickMenu, ContextMenu, Menu  # noqa: F401 F403

solara/lab/components/bokehloaded.vue

@@ -0,0 +1,27 @@
+<template>
+  <div v-if="!loaded">
+    <div class="loading-text"></div>
+  </div>
+</template>
+
+<script>
+module.exports = {
+  mounted() {
+    const check = () => {
+      if (window.Bokeh) {
+        this.loaded = true;
+        return;
+      }
+      setTimeout(check, 100);
+    };
+    check();
+  },
+};
+</script>
+
+<style>
+.loading-text {
+  margin-top: 10px;
+  font-size: 16px;
+}
+</style>

solara/lab/components/figurebokeh.py

@@ -0,0 +1,73 @@
+from typing import Callable
+
+import solara
+from solara.components.component_vue import component_vue
+from bokeh.io import output_notebook
+from bokeh.models import Plot
+from bokeh.plotting import figure
+from bokeh.themes import Theme
+from jupyter_bokeh import BokehModel
+
+
+@component_vue("bokehloaded.vue")
+def BokehLoaded(loaded: bool, on_loaded: Callable[[bool], None]):
+    pass
+
+
+def FigureBokeh(
+    fig,
+    dependencies=None,
+    light_theme: str | Theme = "light_minimal",
+    dark_theme: str | Theme = "dark_minimal",
+):
+    # NOTE: no docstring because not a component.
+    loaded = solara.use_reactive(False)
+    dark = solara.lab.use_dark_effective()
+    fig_key = solara.use_uuid4([])
+    output_notebook(hide_banner=True)
+    BokehLoaded(loaded=loaded.value, on_loaded=loaded.set)
+    if loaded.value:
+        # TODO: there's an error with deletion on the doc. do we need to modify the underlying class?
+        fig_element = BokehModel.element(model=fig).key(fig_key)
+
+        def update_data():
+            fig_widget: BokehModel = solara.get_widget(fig_element)
+            fig_model: Plot | figure = fig_widget._model  # base class for figure
+            if fig != fig_model:  # don't run through on first startup
+                # pause until all updates complete
+                fig_model.hold_render = True
+
+                # extend renderer set and cull previous
+                length = len(fig_model.renderers)
+                fig_model.renderers.extend(fig.renderers)
+                fig_model.renderers = fig_model.renderers[length:]
+
+                # similarly update plot layout properties
+                places = ["above", "below", "center", "left", "right"]
+                for place in places:
+                    attr = getattr(fig_model, place)
+                    newattr = getattr(fig, place)
+                    length = len(attr)
+                    attr.extend(newattr)
+                    if place == "right":
+                        fig_model.hold_render = False
+                    setattr(fig_model, place, attr[length:])
+            return
+
+        def update_theme():
+            # NOTE: using bokeh.io.curdoc and this _document prop will point to the same object
+            fig_widget: BokehModel = solara.get_widget(fig_element)
+            if dark:
+                fig_widget._document.theme = dark_theme
+            else:
+                fig_widget._document.theme = light_theme
+
+        solara.use_effect(update_data, dependencies or fig)
+        solara.use_effect(update_theme, [dark, loaded.value])
+        return fig_element
+    else:
+        # NOTE: we don't return this as to not break effect callbacks outside this function
+        with solara.Card(margin=0, elevation=0):
+            # the card expands to fit space
+            with solara.Row(justify="center"):
+                solara.SpinnerSolara(size="200px")

solara/website/pages/apps/scatter-bokeh.py

@@ -0,0 +1,129 @@
+import pathlib
+import sys
+
+from typing import Optional, cast
+
+import vaex
+import vaex.datasets
+
+import solara
+import solara.lab
+from bokeh.models import ColumnDataSource
+from bokeh.plotting import figure
+from bokeh.transform import linear_cmap, factor_cmap
+
+github_url = solara.util.github_url(__file__)
+if sys.platform != "emscripten":
+    pycafe_url = solara.util.pycafe_url(path=pathlib.Path(__file__), requirements=["vaex", "bokeh"])
+else:
+    pycafe_url = None
+
+df_sample = vaex.datasets.titanic()
+
+
+class State:
+    color = solara.reactive(cast(Optional[str], None))
+    x = solara.reactive(cast(Optional[str], None))
+    y = solara.reactive(cast(Optional[str], None))
+    df = solara.reactive(cast(Optional[vaex.DataFrame], None))
+
+    @staticmethod
+    def load_sample():
+        State.x.value = "age"
+        State.y.value = "fare"
+        State.color.value = "body"
+        State.df.value = df_sample
+
+    @staticmethod
+    def reset():
+        State.df.value = None
+
+
+@solara.component
+def Page():
+    df = State.df.value
+    selected, on_selected = solara.use_state({"x": [0, 0]})  # noqa: SH101
+
+    # the PivotTable will set this cross filter
+    filter, _ = solara.use_cross_filter(id(df), name="scatter")
+
+    # only apply the filter if the filter or dataframe changes
+    def filter_df():
+        if (filter is not None) and (df is not None):
+            return df[filter]
+        else:
+            return df
+
+    dff = solara.use_memo(filter_df, dependencies=[df, filter])
+
+    with solara.AppBar():
+        solara.lab.ThemeToggle()
+    with solara.Sidebar():
+        with solara.Card("Controls", margin=0, elevation=0):
+            with solara.Column():
+                with solara.Row():
+                    solara.Button("Sample dataset", color="primary", text=True, outlined=True, on_click=State.load_sample, disabled=df is not None)
+                    solara.Button("Clear dataset", color="primary", text=True, outlined=True, on_click=State.reset)
+
+                if df is not None:
+                    columns = df.get_column_names()
+                    solara.Select("Column x", values=columns, value=State.x)
+                    solara.Select("Column y", values=columns, value=State.y)
+                    solara.Select("Color", values=columns, value=State.color)
+
+                    solara.provide_cross_filter()
+                    solara.PivotTable(df, ["pclass"], ["sex"], selected=selected, on_selected=on_selected)
+
+    if dff is not None:
+        source = ColumnDataSource(
+            data={
+                "x": dff[State.x.value].values,
+                "y": dff[State.y.value].values,
+                "z": dff[State.color.value].values,
+            }
+        )
+        if State.x.value and State.y.value:
+            p = figure(x_axis_label=State.x.value, y_axis_label=State.y.value, width_policy="max", height=700)
+
+            # add a scatter, colorbar, and mapper
+            color_expr = dff[State.color.value]
+            if (color_expr.dtype == "string") or (color_expr.dtype == "bool"):
+                mapper = factor_cmap
+                factors = color_expr.unique()
+                try:
+                    factors.remove(None)
+                except ValueError:
+                    pass
+                args = dict(palette=f"Viridis{min(11, max(3, color_expr.nunique()))}", factors=factors)
+            else:
+                mapper = linear_cmap
+                args = dict(palette="Viridis256", low=color_expr.min()[()], high=color_expr.max()[()])
+
+            s = p.scatter(source=source, x="x", y="y", size=12, fill_color=mapper(field_name="z", **args))
+            p.add_layout(s.construct_color_bar(title=State.color.value, label_standoff=6, padding=5, border_line_color=None), "right")
+
+            solara.lab.FigureBokeh(p, dark_theme="carbon")
+
+        else:
+            solara.Warning("Select x and y columns")
+
+    else:
+        solara.Info("No data loaded, click on the sample dataset button to load a sample dataset, or upload a file.")
+
+    with solara.Column(style={"max-width": "400px"}):
+        solara.Button(label="View source", icon_name="mdi-github-circle", attributes={"href": github_url, "target": "_blank"}, text=True, outlined=True)
+        if sys.platform != "emscripten":
+            solara.Button(
+                label="Edit this example live on py.cafe",
+                icon_name="mdi-coffee-to-go-outline",
+                attributes={"href": pycafe_url, "target": "_blank"},
+                text=True,
+                outlined=True,
+            )
+
+
+@solara.component
+def Layout(children):
+    route, routes = solara.use_route()
+    dark_effective = solara.lab.use_dark_effective()
+    return solara.AppLayout(children=children, toolbar_dark=dark_effective, color=None)  # if dark_effective else "primary")

solara/website/pages/documentation/examples/fullscreen/scatter_bokeh.py

@@ -0,0 +1,3 @@
+redirect = "/apps/scatter-bokeh"
+
+Page = True

solara/website/pages/documentation/examples/visualization/bokeh.py

@@ -0,0 +1,48 @@
+"""# Scatter plot using Bokeh
+
+This example shows how to use Bokeh to create a scatter plot and a select box to do some filtering.
+
+Inspired by the bokeh documentation.
+"""
+
+from bokeh.models import ColorBar, DataRange1d, LinearColorMapper
+
+from bokeh.plotting import figure, ColumnDataSource
+from bokeh.sampledata import penguins
+
+import solara
+
+title = "Scatter plot using Bokeh"
+
+df = penguins.data
+
+
+@solara.component
+def Page():
+    all_species = df["species"].unique().tolist()
+    species = solara.use_reactive(all_species[0])
+    with solara.Div() as main:
+        solara.Select(label="Species", value=species, values=all_species)
+        dff = df[df["species"] == species.value]
+
+        source = ColumnDataSource(
+            data={
+                "x": dff["bill_length_mm"].values,
+                "y": dff["bill_depth_mm"].values,
+                "z": dff["body_mass_g"].values,
+            }
+        )
+
+        # make a figure
+        p = figure(
+            x_range=DataRange1d(), y_range=DataRange1d(), x_axis_label="Bill length [mm]", y_axis_label="Bill depth [mm]", width_policy="max", height=400
+        )
+
+        # add a scatter, colorbar, and mapper
+        mapper = LinearColorMapper(palette="Viridis256", low=dff["body_mass_g"].min(), high=dff["body_mass_g"].max())
+        cb = ColorBar(color_mapper=mapper, title="Body mass [g]")
+        p.scatter(source=source, x="x", y="y", marker="circle", size=8, fill_color={"field": "z", "transform": mapper})
+        p.add_layout(cb, "right")
+
+        solara.lab.FigureBokeh(p, dark_theme="carbon", dependencies=[species])
+    return main