Correct bug when color and values correspond to the same color in treemap or sunburst (#2591)

* solve aggregation bug when color and values correspond to the same color

* Update packages/python/plotly/plotly/express/_core.py

* put back deleted code

* added test

* code improvement

* changelog entry + improve name of new column

Author: emmanuelle

Diff for: CHANGELOG.md

@@ -9,6 +9,9 @@ This project adheres to [Semantic Versioning](http://semver.org/).
 - Fixed special cases with `px.sunburst` and `px.treemap` with `path` input ([#2524](https://github.com/plotly/plotly.py/pull/2524))
 - Fixed bug in `hover_data` argument of `px` functions, when the column name is changed with labels and `hover_data` is a dictionary setting up a specific format for the hover data ([#2544](https://github.com/plotly/plotly.py/pull/2544)).
 - Made the Plotly Express `trendline` argument more robust and made it work with datetime `x` values ([#2554](https://github.com/plotly/plotly.py/pull/2554))
+- Fixed bug in `px.sunburst` and `px.treemap`: when the `color` and `values`
+  arguments correspond to the same column, a different aggregation function has
+  to be used for the two arguments ([#2591](https://github.com/plotly/plotly.py/pull/2591))
 - Plotly Express wide mode now accepts mixed integer and float columns ([#2598](https://github.com/plotly/plotly.py/pull/2598))
 - Plotly Express `range_(x|y)` should not impact the unlinked range of marginal subplots ([#2600](https://github.com/plotly/plotly.py/pull/2600))
 - `px.line` now sets `line_group=<variable>` in wide mode by default ([#2599](https://github.com/plotly/plotly.py/pull/2599))

Diff for: packages/python/plotly/plotly/express/_core.py

@@ -474,11 +474,10 @@ def make_trace_kwargs(args, trace_spec, trace_data, mapping_labels, sizeref):
                 # We need to invert the mapping here
                 k_args = invert_label(args, k)
                 if k_args in args["hover_data"]:
-                    if args["hover_data"][k_args][0]:
-                        if isinstance(args["hover_data"][k_args][0], str):
-                            mapping_labels_copy[k] = v.replace(
-                                "}", "%s}" % args["hover_data"][k_args][0]
-                            )
+                    formatter = args["hover_data"][k_args][0]
+                    if formatter:
+                        if isinstance(formatter, str):
+                            mapping_labels_copy[k] = v.replace("}", "%s}" % formatter)
                     else:
                         _ = mapping_labels_copy.pop(k)
         hover_lines = [k + "=" + v for k, v in mapping_labels_copy.items()]
@@ -1507,7 +1506,9 @@ def aggfunc_discrete(x):
 
         if args["color"]:
             if args["color"] == args["values"]:
-                aggfunc_color = "sum"
+                new_value_col_name = args["values"] + "_sum"
+                df[new_value_col_name] = df[args["values"]]
+                args["values"] = new_value_col_name
         count_colname = args["values"]
     else:
         # we need a count column for the first groupby and the weighted mean of color
@@ -1526,7 +1527,7 @@ def aggfunc_discrete(x):
         if not _is_continuous(df, args["color"]):
             aggfunc_color = aggfunc_discrete
             discrete_color = True
-        elif not aggfunc_color:
+        else:
 
             def aggfunc_continuous(x):
                 return np.average(x, weights=df.loc[x.index, count_colname])
@@ -1584,6 +1585,9 @@ def aggfunc_continuous(x):
     if args["color"]:
         if not args["hover_data"]:
             args["hover_data"] = [args["color"]]
+        elif isinstance(args["hover_data"], dict):
+            if not args["hover_data"].get(args["color"]):
+                args["hover_data"][args["color"]] = (True, None)
         else:
             args["hover_data"].append(args["color"])
     return args

Diff for: packages/python/plotly/plotly/tests/test_core/test_px/test_px_functions.py

@@ -149,10 +149,6 @@ def test_sunburst_treemap_with_path():
     fig = px.sunburst(df, path=path, values="values")
     assert fig.data[0].branchvalues == "total"
     assert fig.data[0].values[-1] == np.sum(values)
-    # Continuous colorscale
-    fig = px.sunburst(df, path=path, values="values", color="values")
-    assert "coloraxis" in fig.data[0].marker
-    assert np.all(np.array(fig.data[0].marker.colors) == np.array(fig.data[0].values))
     # Error when values cannot be converted to numerical data type
     df["values"] = ["1 000", "3 000", "2", "4", "2", "2", "1 000", "4 000"]
     msg = "Column `values` of `df` could not be converted to a numerical data type."
@@ -162,6 +158,12 @@ def test_sunburst_treemap_with_path():
     path = [df.total, "regions", df.sectors, "vendors"]
     fig = px.sunburst(df, path=path)
     assert fig.data[0].branchvalues == "total"
+    # Continuous colorscale
+    df["values"] = 1
+    fig = px.sunburst(df, path=path, values="values", color="values")
+    assert "coloraxis" in fig.data[0].marker
+    assert np.all(np.array(fig.data[0].marker.colors) == 1)
+    assert fig.data[0].values[-1] == 8
 
 
 def test_sunburst_treemap_with_path_and_hover():

Diff for: packages/python/plotly/plotly/tests/test_core/test_px/test_px_hover.py

@@ -151,3 +151,15 @@ def test_fail_wrong_column():
         "Ambiguous input: values for 'c' appear both in hover_data and data_frame"
         in str(err_msg.value)
     )
+
+
+def test_sunburst_hoverdict_color():
+    df = px.data.gapminder().query("year == 2007")
+    fig = px.sunburst(
+        df,
+        path=["continent", "country"],
+        values="pop",
+        color="lifeExp",
+        hover_data={"pop": ":,"},
+    )
+    assert "color" in fig.data[0].hovertemplate