Merge branch 'master' into smaller-font

CHANGELOG.rst

@@ -6,6 +6,7 @@ What's new in version 0.9
   introduced in version 0.7. (:issue:`248`)
 - Ability to disable totals plot with `totals_plot_elements=0`. (:issue:`246`)
 - Ability to set totals y axis label (:issue:`243`)
+- Added ``max_subset_rank`` to get only n most populous subsets.
 
 What's new in version 0.8
 -------------------------

examples/plot_customize_after_plot.py

@@ -21,3 +21,11 @@
 plot_result["totals"].set_ylabel("Category size")
 plot_result["matrix"].set_xlabel("Subsets between categories")
 plt.show()
+
+
+##########################################################################
+# Or we can place the totals label on the x axis
+
+plot_result = plot(example)
+plot_result["totals"].set_xlabel("Category size")
+plt.show()

examples/plot_generated.py

@@ -28,8 +28,8 @@
 
 ##########################################################################
 
-plot(example, show_counts="{:d}")
-plt.suptitle("With counts shown")
+plot(example, show_counts="{:,}")
+plt.suptitle("With counts shown, using a thousands separator")
 plt.show()
 
 ##########################################################################

examples/plot_highlight.py

@@ -34,7 +34,7 @@
 plt.show()
 
 ##########################################################################
-# ... or their size or degree.
+# ... or their size.
 
 upset = UpSet(example)
 upset.style_subsets(
@@ -44,6 +44,17 @@
 plt.suptitle("Hatch subsets with size >1000")
 plt.show()
 
+##########################################################################
+# ... or degree.
+
+upset = UpSet(example)
+upset.style_subsets(min_degree=1, facecolor="blue")
+upset.style_subsets(min_degree=2, facecolor="purple")
+upset.style_subsets(min_degree=3, facecolor="red")
+upset.plot()
+plt.suptitle("Coloring by degree")
+plt.show()
+
 ##########################################################################
 # Multiple stylings can be applied with different criteria in the same
 # plot.

upsetplot/plotting.py

@@ -28,6 +28,7 @@ def _process_data(
     sum_over,
     min_subset_size=None,
     max_subset_size=None,
+    max_subset_rank=None,
     min_degree=None,
     max_degree=None,
     reverse=False,
@@ -41,6 +42,7 @@ def _process_data(
         sum_over=sum_over,
         min_subset_size=min_subset_size,
         max_subset_size=max_subset_size,
+        max_subset_rank=max_subset_rank,
         min_degree=min_degree,
         max_degree=max_degree,
         include_empty_subsets=include_empty_subsets,
@@ -200,6 +202,11 @@ class UpSet:
         a size greater than this threshold will be omitted from plotting.
 
         .. versionadded:: 0.5
+    max_subset_rank : int, optional
+        Limit to the top N ranked subsets in descending order of size.
+        All tied subsets are included.
+
+        .. versionadded:: 0.9
     min_degree : int, optional
         Minimum degree of a subset to be shown in the plot.
 
@@ -270,6 +277,7 @@ def __init__(
         sum_over=None,
         min_subset_size=None,
         max_subset_size=None,
+        max_subset_rank=None,
         min_degree=None,
         max_degree=None,
         facecolor="auto",
@@ -324,6 +332,7 @@ def __init__(
             sum_over=sum_over,
             min_subset_size=min_subset_size,
             max_subset_size=max_subset_size,
+            max_subset_rank=max_subset_rank,
             min_degree=min_degree,
             max_degree=max_degree,
             reverse=not self._horizontal,
@@ -345,6 +354,7 @@ def style_subsets(
         absent=None,
         min_subset_size=None,
         max_subset_size=None,
+        max_subset_rank=None,
         min_degree=None,
         max_degree=None,
         facecolor=None,
@@ -371,6 +381,11 @@ def style_subsets(
             Minimum size of a subset to be styled.
         max_subset_size : int, optional
             Maximum size of a subset to be styled.
+        max_subset_rank : int, optional
+            Limit to the top N ranked subsets in descending order of size.
+            All tied subsets are included.
+
+            .. versionadded:: 0.9
         min_degree : int, optional
             Minimum degree of a subset to be styled.
         max_degree : int, optional
@@ -405,6 +420,7 @@ def style_subsets(
             absent=absent,
             min_subset_size=min_subset_size,
             max_subset_size=max_subset_size,
+            max_subset_rank=max_subset_rank,
             min_degree=min_degree,
             max_degree=max_degree,
         )

upsetplot/reformat.py

@@ -94,14 +94,25 @@ def _scalar_to_list(val):
 
 
 def _get_subset_mask(
-    agg, min_subset_size, max_subset_size, min_degree, max_degree, present, absent
+    agg,
+    min_subset_size,
+    max_subset_size,
+    max_subset_rank,
+    min_degree,
+    max_degree,
+    present,
+    absent,
 ):
     """Get a mask over subsets based on size, degree or category presence"""
     subset_mask = True
     if min_subset_size is not None:
         subset_mask = np.logical_and(subset_mask, agg >= min_subset_size)
     if max_subset_size is not None:
         subset_mask = np.logical_and(subset_mask, agg <= max_subset_size)
+    if max_subset_rank is not None:
+        subset_mask = np.logical_and(
+            subset_mask, agg.rank(method="min", ascending=False) <= max_subset_rank
+        )
     if (min_degree is not None and min_degree >= 0) or max_degree is not None:
         degree = agg.index.to_frame().sum(axis=1)
         if min_degree is not None:
@@ -121,12 +132,21 @@ def _get_subset_mask(
 
 
 def _filter_subsets(
-    df, agg, min_subset_size, max_subset_size, min_degree, max_degree, present, absent
+    df,
+    agg,
+    min_subset_size,
+    max_subset_size,
+    max_subset_rank,
+    min_degree,
+    max_degree,
+    present,
+    absent,
 ):
     subset_mask = _get_subset_mask(
         agg,
         min_subset_size=min_subset_size,
         max_subset_size=max_subset_size,
+        max_subset_rank=max_subset_rank,
         min_degree=min_degree,
         max_degree=max_degree,
         present=present,
@@ -189,6 +209,7 @@ def query(
     absent=None,
     min_subset_size=None,
     max_subset_size=None,
+    max_subset_rank=None,
     min_degree=None,
     max_degree=None,
     sort_by="degree",
@@ -221,6 +242,11 @@ def query(
         Size may be a sum of values, see `subset_size`.
     max_subset_size : int, optional
         Maximum size of a subset to be reported.
+    max_subset_rank : int, optional
+        Limit to the top N ranked subsets in descending order of size.
+        All tied subsets are included.
+
+        .. versionadded:: 0.9
     min_degree : int, optional
         Minimum degree of a subset to be reported.
     max_degree : int, optional
@@ -348,6 +374,7 @@ def query(
         agg,
         min_subset_size=min_subset_size,
         max_subset_size=max_subset_size,
+        max_subset_rank=max_subset_rank,
         min_degree=min_degree,
         max_degree=max_degree,
         present=present,

upsetplot/tests/test_upsetplot.py

@@ -753,6 +753,14 @@ def test_index_must_be_bool(x):
                 (True, True, True): 990,
             },
         ),
+        (
+            {"max_subset_rank": 3},
+            {
+                (True, False, False): 884,
+                (True, True, False): 1547,
+                (True, True, True): 990,
+            },
+        ),
         (
             {"min_subset_size": 800, "max_subset_size": 990},
             {
@@ -822,6 +830,29 @@ def test_filter_subsets(filter_params, expected, sort_by):
     assert upset_full.total == pytest.approx(upset_filtered.total)
 
 
+def test_filter_subsets_max_subset_rank_tie():
+    data = generate_samples(seed=0, n_samples=5, n_categories=3)
+    tested_non_tie = False
+    tested_tie = True
+    full = UpSet(data, subset_size="count").intersections
+    prev = None
+    for max_rank in range(1, 5):
+        cur = UpSet(data, subset_size="count", max_subset_rank=max_rank).intersections
+        if prev is not None:
+            if cur.shape[0] > prev.shape[0]:
+                # check we add rows only when they are new
+                assert cur.min() < prev.min()
+                tested_non_tie = True
+            elif cur.shape[0] != full.shape[0]:
+                assert (cur == cur.min()).sum() > 1
+                tested_tie = True
+
+        prev = cur
+    assert tested_non_tie
+    assert tested_tie
+    assert cur.shape[0] == full.shape[0]
+
+
 @pytest.mark.parametrize(
     "x",
     [