feat: Extend .interactive() with tooltip and legend

altair/vegalite/v5/api.py

@@ -3897,33 +3897,117 @@ def add_selection(self, *params) -> Self:  # noqa: ANN002
         return self.add_params(*params)
 
     def interactive(
-        self, name: str | None = None, bind_x: bool = True, bind_y: bool = True
-    ) -> Self:
+        self,
+        name: str | None = None,
+        bind_x: bool = True,
+        bind_y: bool = True,
+        tooltip: bool = True,
+        legend: bool | LegendChannel_T = False,
+    ) -> Chart:
         """
-        Make chart axes scales interactive.
+        Add common interactive elements to the chart.
 
         Parameters
         ----------
-        name : string
+        name : string or None
             The parameter name to use for the axes scales. This name should be
-            unique among all parameters within the chart.
+            unique among all parameters within the chart
         bind_x : boolean, default True
-            If true, then bind the interactive scales to the x-axis
+            Bind the interactive scales to the x-axis
         bind_y : boolean, default True
-            If true, then bind the interactive scales to the y-axis
+            Bind the interactive scales to the y-axis
+        tooltip : boolean, default True,
+            Add a tooltip containing the encodings used in the chart
+        legend : boolean or string, default True
+            A single encoding channel to be used to create a clickable legend.
+            The deafult is to guess from the spec based on the most commonly used legend encodings.
 
         Returns
         -------
         chart :
-            copy of self, with interactive axes added
+            copy of self, with interactivity added
 
         """
         encodings: list[SingleDefUnitChannel_T] = []
         if bind_x:
             encodings.append("x")
         if bind_y:
             encodings.append("y")
-        return self.add_params(selection_interval(bind="scales", encodings=encodings))
+        chart: Chart = self.copy().add_params(
+            selection_interval(bind="scales", encodings=encodings)
+        )
+        # We can't simply use configure_mark since configure methods
+        # are not allowed in layered specs
+        if tooltip:
+            chart = _add_tooltip(chart)
+        legend_encodings_missing = utils.is_undefined(chart.encoding)
+        if legend and not legend_encodings_missing:
+            facet_encoding: FacetedEncoding = chart.encoding
+            if not isinstance(legend, str):
+                legend = _infer_legend_encoding(facet_encoding)
+
+            facet_legend = facet_encoding[legend]
+            legend_type = facet_legend["type"]
+            if utils.is_undefined(legend_type):
+                legend_type = facet_legend.to_dict(context={"data": chart.data})["type"]
+
+            if legend_type == "nominal":
+                # TODO Ideally this would work for ordinal data too
+                legend_selection = selection_point(bind="legend", encodings=[legend])
+                initial_computed_domain = param(expr=f"domain('{legend}')")
+                nonreactive_domain = param(
+                    react=False, expr=initial_computed_domain.name
+                )
+                scale = facet_legend["scale"]
+                if utils.is_undefined(scale):
+                    scale = {"domain": nonreactive_domain}
+                else:
+                    scale["domain"] = nonreactive_domain
+                chart = chart.add_params(
+                    legend_selection,
+                    initial_computed_domain,
+                    nonreactive_domain,
+                ).transform_filter(legend_selection)
+            else:
+                msg = f"Expected only 'nominal' legend type but got {legend_type!r}"
+                raise NotImplementedError(msg)
+        return chart
+
+
+LegendChannel_T: TypeAlias = Literal[
+    "color",
+    "fill",
+    "shape",
+    "stroke",
+    "opacity",
+    "fillOpacity",
+    "strokeOpacity",
+    "strokeWidth",
+    "strokeDash",
+    "angle",  # TODO Untested
+    "radius",  # TODO Untested
+    "radius2",  # TODO Untested
+    # "size",  # TODO Currently size is not working, renders empty legend
+]
+
+
+def _add_tooltip(chart: _TChart, /) -> _TChart:
+    if isinstance(chart.mark, str):
+        chart.mark = {"type": chart.mark, "tooltip": True}
+    else:
+        chart.mark.tooltip = True
+    return chart
+
+
+def _infer_legend_encoding(encoding: FacetedEncoding, /) -> LegendChannel_T:
+    """Set the legend to commonly used encodings by default."""
+    _channels = t.get_args(LegendChannel_T)
+    it = (ch for ch in _channels if not utils.is_undefined(encoding[ch]))
+    if legend := next(it, None):
+        return legend
+    else:
+        msg = f"Unable to infer target channel for 'legend'.\n\n{encoding!r}"
+        raise NotImplementedError(msg)
 
 
 def _check_if_valid_subspec(
@@ -5003,6 +5087,16 @@ def sphere() -> SphereGenerator:
     return core.SphereGenerator(sphere=True)
 
 
+_TChart = TypeVar(
+    "_TChart",
+    Chart,
+    RepeatChart,
+    ConcatChart,
+    HConcatChart,
+    VConcatChart,
+    FacetChart,
+    LayerChart,
+)
 ChartType: TypeAlias = Union[
     Chart, RepeatChart, ConcatChart, HConcatChart, VConcatChart, FacetChart, LayerChart
 ]

tests/vegalite/v5/test_api.py

@@ -13,7 +13,7 @@
 from datetime import date, datetime
 from importlib.metadata import version as importlib_version
 from importlib.util import find_spec
-from typing import TYPE_CHECKING
+from typing import TYPE_CHECKING, Any, Literal
 
 import ibis
 import jsonschema
@@ -28,15 +28,20 @@
 from tests import skip_requires_vl_convert, slow
 
 if TYPE_CHECKING:
+    from altair.typing import ChartType, Optional
     from altair.vegalite.v5.api import _Conditional, _Conditions
-    from altair.vegalite.v5.schema._typing import Map
+    from altair.vegalite.v5.schema._typing import Map, SingleDefUnitChannel_T
 
 ibis.set_backend("polars")
 
 PANDAS_VERSION = Version(importlib_version("pandas"))
 
+_MakeType = Literal[
+    "layer", "hconcat", "vconcat", "concat", "facet", "facet_encoding", "repeat"
+]
 
-def getargs(*args, **kwargs):
+
+def getargs(*args, **kwargs) -> tuple[tuple[Any, ...], dict[str, Any]]:
     return args, kwargs
 
 
@@ -47,7 +52,7 @@ def getargs(*args, **kwargs):
 }
 
 
-def _make_chart_type(chart_type):
+def _make_chart_type(chart_type: _MakeType) -> ChartType:
     data = pd.DataFrame(
         {
             "x": [28, 55, 43, 91, 81, 53, 19, 87],
@@ -81,8 +86,24 @@ def _make_chart_type(chart_type):
         raise ValueError(msg)
 
 
+@pytest.fixture(
+    params=(
+        "layer",
+        "hconcat",
+        "vconcat",
+        "concat",
+        "facet",
+        "facet_encoding",
+        "repeat",
+    )
+)
+def all_chart_types(request) -> ChartType:
+    """Use the parameter name ``all_chart_types`` to automatically parameterise."""
+    return _make_chart_type(request.param)
+
+
 @pytest.fixture
-def basic_chart():
+def basic_chart() -> alt.Chart:
     data = pd.DataFrame(
         {
             "a": ["A", "B", "C", "D", "E", "F", "G", "H", "I"],
@@ -93,6 +114,18 @@ def basic_chart():
     return alt.Chart(data).mark_bar().encode(x="a", y="b")
 
 
+@pytest.fixture
+def color_data() -> pl.DataFrame:
+    """10 rows, 3 columns ``"x:Q"``, ``"y:Q"``, ``"color:(N|O)"``."""
+    return pl.DataFrame(
+        {
+            "x": [0.32, 0.86, 0.10, 0.30, 0.47, 0.0, 1.0, 0.91, 0.88, 0.12],
+            "y": [0.11, 0.33, 0.01, 0.04, 0.77, 0.1, 0.2, 0.23, 0.05, 0.29],
+            "color": list("ACABBCABBA"),
+        }
+    )
+
+
 @pytest.fixture
 def cars():
     return pd.DataFrame(
@@ -1657,3 +1690,118 @@ def test_ibis_with_vegafusion(monkeypatch: pytest.MonkeyPatch):
         {"a": 2, "b": "2020-01-02T00:00:00.000"},
         {"a": 3, "b": "2020-01-03T00:00:00.000"},
     ]
+
+
+# TODO These chart types don't all work yet
+@pytest.mark.parametrize(
+    "chart_type",
+    [
+        "chart",
+        pytest.param(
+            "layer", marks=pytest.mark.xfail(reason="Not Implemented", raises=TypeError)
+        ),
+        pytest.param(
+            "facet", marks=pytest.mark.xfail(reason="Not Implemented", raises=TypeError)
+        ),
+    ],
+)
+def test_interactive_for_chart_types(chart_type: _MakeType):
+    chart = _make_chart_type(chart_type)
+    assert chart.interactive(legend=True).to_dict()  # type: ignore[call-arg]
+
+
+def test_interactive_with_no_encoding(all_chart_types):
+    # Should not raise error when there is no encoding
+    assert all_chart_types.interactive().to_dict()
+
+
+def test_interactive_tooltip_added_for_all_encodings():
+    # TODO Loop through all possible encodings
+    # and check that tooltip interactivity is added for all of them
+    assert "tooltip" in alt.Chart().mark_point().interactive().to_dict()["mark"]
+    assert (
+        "tooltip"
+        not in alt.Chart().mark_point().interactive(tooltip=False).to_dict()["mark"]
+    )
+
+
+@pytest.mark.parametrize(
+    ("encoding", "err"),
+    [
+        ("xOffset", NotImplementedError),
+        ("yOffset", NotImplementedError),
+        ("x2", NotImplementedError),
+        ("y2", NotImplementedError),
+        ("longitude", NotImplementedError),
+        ("latitude", NotImplementedError),
+        ("longitude2", NotImplementedError),
+        ("latitude2", NotImplementedError),
+        ("theta", NotImplementedError),
+        ("theta2", NotImplementedError),
+        ("radius", None),
+        ("radius2", KeyError),
+        ("color", None),
+        ("fill", None),
+        ("stroke", None),
+        ("opacity", None),
+        ("fillOpacity", None),
+        ("strokeOpacity", None),
+        ("strokeWidth", None),
+        ("strokeDash", None),
+        ("size", NotImplementedError),
+        ("angle", None),
+        ("shape", None),
+        ("key", NotImplementedError),
+        ("text", NotImplementedError),
+        ("href", NotImplementedError),
+        ("url", NotImplementedError),
+        ("description", NotImplementedError),
+    ],
+)
+def test_interactive_legend_made_interactive_for_legend_encodings(
+    color_data, encoding: SingleDefUnitChannel_T, err: type[Exception] | None
+) -> None:
+    """Check that legend interactivity is added only for the allowed legend encodings."""
+    chart = (
+        alt.Chart(color_data).mark_point().encode(x="x", y="y", **{encoding: "color"})  # type: ignore[misc]
+    )
+    if err is None:
+        assert chart.interactive(legend=True).to_dict()
+    else:
+        with pytest.raises(err):
+            chart.interactive(legend=True).to_dict()
+
+
+def test_interactive_legend_made_interactive_for_appropriate_encodings_types(
+    color_data,
+) -> None:
+    chart = alt.Chart(color_data).mark_point().encode(x="x", y="y")
+
+    # TODO Reverse legend=False/True once we revert the default arg to true
+    assert len(chart.encode(color="color:N").interactive().to_dict()["params"]) == 1
+    chart_with_nominal_legend_encoding = (
+        chart.encode(color="color:N").interactive(legend=True).to_dict()
+    )
+    assert len(chart_with_nominal_legend_encoding["params"]) == 4
+    for param in chart_with_nominal_legend_encoding["params"]:
+        if "expr" in param:
+            assert param["expr"] == "domain('color')" or "react" in param
+
+    # TODO These tests currently don't work because we are raising
+    # when the legend is not nominal. To be updated if we change that behavior
+    # TODO Could change this first test if we get interactivity working with nominal
+    # chart_with_ordinal_legend_encoding = (
+    #     chart.encode(color="color:O").interactive(legend=True).to_dict()
+    # )
+    # assert len(chart_with_ordinal_legend_encoding["params"]) == 1
+
+    # chart_with_quantitative_legend_encoding = (
+    #     chart.encode(color="color:Q").interactive(legend=True).to_dict()
+    # )
+    # assert len(chart_with_quantitative_legend_encoding["params"]) == 1
+
+
+def test_interactive_legend_encoding_correctly_picked_from_multiple_encodings():
+    # TODO The legend should be chosen based on the priority order
+    # in the list of possible legend encodings
+    ...