use return_type directly when building datasets

Author: FBruzzesi

CHANGELOG.md

@@ -10,7 +10,7 @@ This project adheres to [Semantic Versioning](http://semver.org/).
 
 - Updated plotly.py to use base64 encoding of arrays in plotly JSON to improve performance.
 - Add `subtitle` attribute to all Plotly Express traces
-- Allow to load plotly data directly via pandas, polars and pyarrow, without depending directly on any [#4843](https://github.com/plotly/plotly.py/pull/4843)
+- Make plotly-express dataframe agnostic via Narwhals [#4790](https://github.com/plotly/plotly.py/pull/4790)
 
 ## [5.24.1] - 2024-09-12
 

packages/python/plotly/plotly/tests/test_optional/test_px/conftest.py

@@ -41,3 +41,8 @@ def pyarrow_table_constructor(obj) -> IntoDataFrame:
 @pytest.fixture(params=constructors)
 def constructor(request: pytest.FixtureRequest):
     return request.param  # type: ignore[no-any-return]
+
+
+@pytest.fixture(params=["pandas", "pyarrow", "polars"])
+def backend(request: pytest.FixtureRequest) -> str:
+    return request.param  # type: ignore[no-any-return]

packages/python/plotly/plotly/tests/test_optional/test_px/test_facets.py

@@ -4,9 +4,8 @@
 import random
 
 
-def test_facets(constructor):
-    data = px.data.tips().to_dict(orient="list")
-    df = constructor(data)
+def test_facets(backend):
+    df = px.data.tips(return_type=backend)
 
     fig = px.scatter(df, x="total_bill", y="tip")
     assert "xaxis2" not in fig.layout
@@ -47,9 +46,8 @@ def test_facets(constructor):
     assert fig.layout.yaxis4.domain[0] - fig.layout.yaxis.domain[1] == approx(0.08)
 
 
-def test_facets_with_marginals(constructor):
-    data = px.data.tips().to_dict(orient="list")
-    df = constructor(data)
+def test_facets_with_marginals(backend):
+    df = px.data.tips(return_type=backend)
 
     fig = px.histogram(df, x="total_bill", facet_col="sex", marginal="rug")
     assert len(fig.data) == 4

packages/python/plotly/plotly/tests/test_optional/test_px/test_marginals.py

@@ -5,9 +5,8 @@
 @pytest.mark.parametrize("px_fn", [px.scatter, px.density_heatmap, px.density_contour])
 @pytest.mark.parametrize("marginal_x", [None, "histogram", "box", "violin"])
 @pytest.mark.parametrize("marginal_y", [None, "rug"])
-def test_xy_marginals(constructor, px_fn, marginal_x, marginal_y):
-    data = px.data.tips().to_dict(orient="list")
-    df = constructor(data)
+def test_xy_marginals(backend, px_fn, marginal_x, marginal_y):
+    df = px.data.tips(return_type=backend)
 
     fig = px_fn(
         df, x="total_bill", y="tip", marginal_x=marginal_x, marginal_y=marginal_y
@@ -18,9 +17,8 @@ def test_xy_marginals(constructor, px_fn, marginal_x, marginal_y):
 @pytest.mark.parametrize("px_fn", [px.histogram, px.ecdf])
 @pytest.mark.parametrize("marginal", [None, "rug", "histogram", "box", "violin"])
 @pytest.mark.parametrize("orientation", ["h", "v"])
-def test_single_marginals(constructor, px_fn, marginal, orientation):
-    data = px.data.tips().to_dict(orient="list")
-    df = constructor(data)
+def test_single_marginals(backend, px_fn, marginal, orientation):
+    df = px.data.tips(return_type=backend)
 
     fig = px_fn(
         df, x="total_bill", y="total_bill", marginal=marginal, orientation=orientation

packages/python/plotly/plotly/tests/test_optional/test_px/test_px.py

@@ -6,9 +6,8 @@
 from itertools import permutations
 
 
-def test_scatter(constructor):
-    data = px.data.iris().to_dict(orient="list")
-    iris = nw.from_native(constructor(data))
+def test_scatter(backend):
+    iris = nw.from_native(px.data.iris(return_type=backend))
     fig = px.scatter(iris.to_native(), x="sepal_width", y="sepal_length")
     assert fig.data[0].type == "scatter"
     assert np.all(fig.data[0].x == iris.get_column("sepal_width").to_numpy())
@@ -17,9 +16,8 @@ def test_scatter(constructor):
     assert fig.data[0].mode == "markers"
 
 
-def test_custom_data_scatter(constructor):
-    data = px.data.iris().to_dict(orient="list")
-    iris = nw.from_native(constructor(data))
+def test_custom_data_scatter(backend):
+    iris = nw.from_native(px.data.iris(return_type=backend))
     # No hover, no custom data
     fig = px.scatter(
         iris.to_native(), x="sepal_width", y="sepal_length", color="species"
@@ -67,9 +65,8 @@ def test_custom_data_scatter(constructor):
     )
 
 
-def test_labels(constructor):
-    data = px.data.tips().to_dict(orient="list")
-    tips = nw.from_native(constructor(data))
+def test_labels(backend):
+    tips = nw.from_native(px.data.tips(return_type=backend))
     fig = px.scatter(
         tips.to_native(),
         x="total_bill",
@@ -100,10 +97,8 @@ def test_labels(constructor):
         ({"text": "continent"}, "lines+markers+text"),
     ],
 )
-def test_line_mode(constructor, extra_kwargs, expected_mode):
-    data = px.data.gapminder().to_dict(orient="list")
-    gapminder = constructor(data)
-
+def test_line_mode(backend, extra_kwargs, expected_mode):
+    gapminder = px.data.gapminder(return_type=backend)
     fig = px.line(
         gapminder,
         x="year",
@@ -114,12 +109,11 @@ def test_line_mode(constructor, extra_kwargs, expected_mode):
     assert fig.data[0].mode == expected_mode
 
 
-def test_px_templates(constructor):
+def test_px_templates(backend):
     try:
         import plotly.graph_objects as go
 
-        data = px.data.tips().to_dict(orient="list")
-        tips = constructor(data)
+        tips = px.data.tips(return_type=backend)
 
         # use the normal defaults
         fig = px.scatter()
@@ -245,12 +239,11 @@ def test_px_defaults():
         pio.templates.default = "plotly"
 
 
-def assert_orderings(constructor, days_order, days_check, times_order, times_check):
+def assert_orderings(backend, days_order, days_check, times_order, times_check):
     symbol_sequence = ["circle", "diamond", "square", "cross", "circle", "diamond"]
     color_sequence = ["red", "blue", "red", "blue", "red", "blue", "red", "blue"]
 
-    data = px.data.tips().to_dict(orient="list")
-    tips = nw.from_native(constructor(data))
+    tips = nw.from_native(px.data.tips(return_type=backend))
 
     fig = px.scatter(
         tips.to_native(),
@@ -284,16 +277,16 @@ def assert_orderings(constructor, days_order, days_check, times_order, times_che
 
 @pytest.mark.parametrize("days", permutations(["Sun", "Sat", "Fri", "x"]))
 @pytest.mark.parametrize("times", permutations(["Lunch", "x"]))
-def test_orthogonal_and_missing_orderings(constructor, days, times):
+def test_orthogonal_and_missing_orderings(backend, days, times):
     assert_orderings(
-        constructor, days, list(days) + ["Thur"], times, list(times) + ["Dinner"]
+        backend, days, list(days) + ["Thur"], times, list(times) + ["Dinner"]
     )
 
 
 @pytest.mark.parametrize("days", permutations(["Sun", "Sat", "Fri", "Thur"]))
 @pytest.mark.parametrize("times", permutations(["Lunch", "Dinner"]))
-def test_orthogonal_orderings(constructor, days, times):
-    assert_orderings(constructor, days, days, times, times)
+def test_orthogonal_orderings(backend, days, times):
+    assert_orderings(backend, days, days, times, times)
 
 
 def test_permissive_defaults():
@@ -302,9 +295,8 @@ def test_permissive_defaults():
         px.defaults.should_not_work = "test"
 
 
-def test_marginal_ranges(constructor):
-    data = px.data.tips().to_dict(orient="list")
-    df = constructor(data)
+def test_marginal_ranges(backend):
+    df = px.data.tips(return_type=backend)
     fig = px.scatter(
         df,
         x="total_bill",
@@ -318,9 +310,8 @@ def test_marginal_ranges(constructor):
     assert fig.layout.yaxis3.range is None
 
 
-def test_render_mode(constructor):
-    data = px.data.gapminder().to_dict(orient="list")
-    df = nw.from_native(constructor(data))
+def test_render_mode(backend):
+    df = nw.from_native(px.data.gapminder(return_type=backend))
     df2007 = df.filter(nw.col("year") == 2007)
 
     fig = px.scatter(df2007.to_native(), x="gdpPercap", y="lifeExp", trendline="ols")

packages/python/plotly/plotly/tests/test_optional/test_px/test_px_functions.py

@@ -187,31 +187,28 @@ def test_sunburst_treemap_with_path(constructor):
     assert fig.data[0].values[-1] == 8
 
 
-def test_sunburst_treemap_with_path_and_hover(constructor):
-    data = px.data.tips().to_dict(orient="list")
-    df = constructor(data)
+def test_sunburst_treemap_with_path_and_hover(backend):
+    df = px.data.tips(return_type=backend)
     fig = px.sunburst(
         df, path=["sex", "day", "time", "smoker"], color="smoker", hover_data=["smoker"]
     )
     assert "smoker" in fig.data[0].hovertemplate
 
-    data = px.data.gapminder().query("year == 2007").to_dict(orient="list")
-    df = constructor(data)
-
+    df = nw.from_native(px.data.gapminder(year=2007, return_type=backend))
     fig = px.sunburst(
-        df, path=["continent", "country"], color="lifeExp", hover_data=df.columns
+        df.to_native(),
+        path=["continent", "country"],
+        color="lifeExp",
+        hover_data=df.columns,
     )
     assert fig.layout.coloraxis.colorbar.title.text == "lifeExp"
 
-    data = px.data.tips().to_dict(orient="list")
-    df = constructor(data)
-
+    df = px.data.tips(return_type=backend)
     fig = px.sunburst(df, path=["sex", "day", "time", "smoker"], hover_name="smoker")
     assert "smoker" not in fig.data[0].hovertemplate  # represented as '%{hovertext}'
     assert "%{hovertext}" in fig.data[0].hovertemplate  # represented as '%{hovertext}'
 
-    data = px.data.tips().to_dict(orient="list")
-    df = constructor(data)
+    df = px.data.tips(return_type=backend)
     fig = px.sunburst(df, path=["sex", "day", "time", "smoker"], custom_data=["smoker"])
     assert fig.data[0].customdata[0][0] in ["Yes", "No"]
     assert "smoker" not in fig.data[0].hovertemplate
@@ -414,9 +411,8 @@ def test_funnel():
     assert len(fig.data) == 2
 
 
-def test_parcats_dimensions_max(constructor):
-    data = px.data.tips().to_dict(orient="list")
-    df = constructor(data)
+def test_parcats_dimensions_max(backend):
+    df = px.data.tips(return_type=backend)
 
     # default behaviour
     fig = px.parallel_categories(df)
@@ -449,13 +445,12 @@ def test_parcats_dimensions_max(constructor):
 
 
 @pytest.mark.parametrize("histfunc,y", [(None, None), ("count", "tip")])
-def test_histfunc_hoverlabels_univariate(constructor, histfunc, y):
+def test_histfunc_hoverlabels_univariate(backend, histfunc, y):
     def check_label(label, fig):
         assert fig.layout.yaxis.title.text == label
         assert label + "=" in fig.data[0].hovertemplate
 
-    data = px.data.tips().to_dict(orient="list")
-    df = constructor(data)
+    df = px.data.tips(return_type=backend)
 
     # base case, just "count" (note count(tip) is same as count())
     fig = px.histogram(df, x="total_bill", y=y, histfunc=histfunc)
@@ -481,13 +476,12 @@ def check_label(label, fig):
             check_label("%s (normalized as %s)" % (histnorm, barnorm), fig)
 
 
-def test_histfunc_hoverlabels_bivariate(constructor):
+def test_histfunc_hoverlabels_bivariate(backend):
     def check_label(label, fig):
         assert fig.layout.yaxis.title.text == label
         assert label + "=" in fig.data[0].hovertemplate
 
-    data = px.data.tips().to_dict(orient="list")
-    df = constructor(data)
+    df = px.data.tips(return_type=backend)
 
     # with y, should be same as forcing histfunc to sum
     fig = px.histogram(df, x="total_bill", y="tip")

packages/python/plotly/plotly/tests/test_optional/test_px/test_px_hover.py

@@ -6,9 +6,8 @@
 from collections import OrderedDict  # an OrderedDict is needed for Python 2
 
 
-def test_skip_hover(constructor):
-    data = px.data.iris().to_dict(orient="list")
-    df = constructor(data)
+def test_skip_hover(backend):
+    df = px.data.iris(return_type=backend)
     fig = px.scatter(
         df,
         x="petal_length",
@@ -19,9 +18,8 @@ def test_skip_hover(constructor):
     assert fig.data[0].hovertemplate == "species_id=%{marker.size}<extra></extra>"
 
 
-def test_hover_data_string_column(constructor):
-    data = px.data.tips().to_dict(orient="list")
-    df = constructor(data)
+def test_hover_data_string_column(backend):
+    df = px.data.tips(return_type=backend)
     fig = px.scatter(
         df,
         x="tip",
@@ -31,9 +29,8 @@ def test_hover_data_string_column(constructor):
     assert "sex" in fig.data[0].hovertemplate
 
 
-def test_composite_hover(constructor):
-    data = px.data.tips().to_dict(orient="list")
-    df = constructor(data)
+def test_composite_hover(backend):
+    df = px.data.tips(return_type=backend)
     hover_dict = OrderedDict(
         {"day": False, "time": False, "sex": True, "total_bill": ":.1f"}
     )
@@ -91,9 +88,8 @@ def test_newdatain_hover_data():
         )
 
 
-def test_formatted_hover_and_labels(constructor):
-    data = px.data.tips().to_dict(orient="list")
-    df = constructor(data)
+def test_formatted_hover_and_labels(backend):
+    df = px.data.tips(return_type=backend)
     fig = px.scatter(
         df,
         x="tip",
@@ -176,9 +172,8 @@ def test_fail_wrong_column():
     )
 
 
-def test_sunburst_hoverdict_color(constructor):
-    data = px.data.gapminder().query("year == 2007").to_dict(orient="list")
-    df = constructor(data)
+def test_sunburst_hoverdict_color(backend):
+    df = px.data.gapminder(year=2007, return_type=backend)
     fig = px.sunburst(
         df,
         path=["continent", "country"],
@@ -189,7 +184,7 @@ def test_sunburst_hoverdict_color(constructor):
     assert "color" in fig.data[0].hovertemplate
 
 
-def test_date_in_hover(request, constructor):
+def test_date_in_hover(constructor):
     df = nw.from_native(
         constructor({"date": ["2015-04-04 19:31:30+01:00"], "value": [3]})
     ).with_columns(date=nw.col("date").str.to_datetime(format="%Y-%m-%d %H:%M:%S%z"))