def test_three_sets_int(self):
self.assertDictEqual(
iterables.subsets([{0, 1}, {2, 3}, {1, 3, 4}]), {
'010': set([2]),
'011': set([3]),
'001': set([4]),
'111': set([]),
'110': set([]),
'100': set([0]),
'101': set([1])
})
def upset_wrap(N, labels=None, drop_empty=False, drop_single=False, **kwargs):
'''
Wrapper for UpSetPlot package. Mostly just generates the Boolean
multi-indexed ``pandas.Series`` the ``upsetplot.plot`` function
needs as input.
* Arguments:
- *N* [list]: Or any iterable type containing [set] objects.
- *labels* [list]: Optional, ``None`` by default. Labels for the
sets following the same order as provided in *N*. If none is
passed they will be labelled ``'set0'``, ``'set1'`` and so on.
- *drop_empty* [bool]: Optional, ``False`` by default. Whether
to remove the empty set intersections from the plot or not.
- *drop_single* [bool]: Optional, ``False`` by default. Whether
to drop the non-intersecting sets from the plot.
- *\*\*kwargs*: Optional. Additional keyword arguments passed to
``upsetplot.UpSet`` class.
* Returns:
[dict]: Contains the ``matplotlib.axes.Axes`` instances for the
UpSetPlot figure.
'''
ss = subsets(N)
ids = [tuple([bool(int(j)) for j in i]) for i in ss.keys()]
counts = list(map(len, ss.values()))
if labels is None:
labels = ['set%d' % i for i in range(len(N))]
idx = pd.MultiIndex.from_tuples(ids, names=labels)
series = pd.Series(counts, index=idx)
if drop_empty:
series[series == 0] = np.nan
series.dropna(inplace=True)
if drop_single:
series = series.iloc[[
n for n, i in enumerate(series.index) if sum(i) > 1
]]
return usp.plot(series, **kwargs)
def venn(N, labels=['A', 'B', 'C', 'D', 'E'], c=['C0', 'C1', 'C2', 'C3', 'C4'],
pct=False, sizes=False, title=None, filename=None, figsize=None):
'''
Plots a Venn diagram from a list of sets *N*. Number of sets must be
between 2 and 5 (inclusive).
* Arguments:
- *N* [list]: Or any iterable type containing [set] objects.
- *labels* [list]: Optional, ``['A', 'B', 'C', 'D', 'E']`` by
default. Labels for the sets following the same order as
provided in *N*.
- *c* [list]: Optional, ``['C0', 'C1' 'C2', 'C3', 'C4']`` by
default (matplotlib default colors). Any iterable containing
color arguments tolerated by matplotlib (e.g.: ``['r', 'b']``
for red and blue). Must contain at least the same number of
elements as *N* (if more are provided, they will be ignored).
- *pct* [bool]: Optional, ``False`` by default. Indicates
whether to show percentages instead of absolute counts.
- *sizes* [bool]: Optional, ``False`` by default. Whether to
include the size of the sets in the legend or not.
- *title* [str]: Optional, ``None`` by default. Defines the plot
title.
- *filename* [str]: Optional, ``None`` by default. If passed,
indicates the file name or path where to store the figure.
Format must be specified (e.g.: .png, .pdf, etc)
- *figsize* [tuple]: Optional, ``None`` by default (default
matplotlib size). Any iterable containing two values denoting
the figure size (in inches) as [width, height].
* Returns:
- [matplotlib.figure.Figure]: The figure object containing a
combination of box and scatter plots of the gene-set scores,
unless *filename* is provided.
* Example:
>>> N = [{0, 1}, {2, 3}, {1, 3, 4}] # Sets A, B, C
>>> venn(N)
.. image:: ../figures/venn_example.png
:align: center
:scale: 100
'''
def ellipse(ax, x, y, w, h, a, color, alpha=1, label=None):
e = matplotlib.patches.Ellipse(xy=(x, y), width=w, height=h, angle=a,
color=color, alpha=alpha, label=label)
ax.add_patch(e)
if len(N) == 2:
# Ellipse parameters
x = [-.25, .25]
y = [0, 0]
w = [1, 1]
h = [1.5, 1.5]
a = [0, 0]
# Text (counts) parameters
xt = [-.5, .5, 0]
yt = [0, 0, 0]
keys = ['10', '01', '11']
elif len(N) == 3:
# Ellipse parameters
x = [0, -.25, .25]
y = [.33, -.33, -.33]
w = [1, 1, 1]
h = [1.5, 1.5, 1.5]
a = [0, 0, 0]
# Text (counts) parameters
xt = [0, -.5, .5, -.33, .33, 0, 0]
yt = [.6, -.5, -.5, .15, .15, -.6, -.1]
keys = ['100', '010', '001', '110', '101', '011', '111']
elif len(N) == 4:
# Ellipse parameters
x = [-.15, -.35, .15, .35]
y = [.15, -.25, .15, -.25]
w = [2, 2, 2, 2]
h = [1, 1, 1, 1]
a = [-60, -60, 60, 60]
# Text (counts) parameters
xt = [-.5, -.8, .5, .8, -.6, 0, .4, -.4, 0, .6, -.3, .2, .3, -.2, 0]
yt = [.8, -.1, .8, -.1, .33, .4, -.55, -.55, -.9, .33, 0, -.68, 0,
-.68, -.33]
keys = ['1000', '0100', '0010', '0001', '1100', '1010', '1001', '0110',
'0101', '0011', '1110', '1101', '1011', '0111', '1111']
elif len(N) == 5:
# Ellipse parameters
x = [0, -.2125, -.2375, -.03125, .125]
y = [0, -.05, -.275, -.35, -.1875]
w = [1.25, 1.25, 1.25, 1.25, 1.25]
h = [2, 2, 2, 2, 2]
a = [0, 71, 154, 37, 108]
# Text (counts) parameters
xt = [0, -1, -.7, .5, .9, -.41, .1, .2, .51, -.87, -.67, .69, -.25,
-.69, .6, -.1, -.51, .54, -.06, .33, .3, -.72, -.79, .63, -.48,
-.38, .3, .49, -.13, -.69, -.1]
yt = [.8, .1, -1, -1, .05, .55, .55, -.91, .3, -.3, .3, -.25, -1, -.65,
-.65, .5, .43, 0, -.95, .4, -.78, 0, -.41, -.47, -.76, .33, .27,
-.4, -.82, -.25, -.2]
keys = ['10000', '01000', '00100', '00010', '00001', '11000', '10100',
'10010', '10001', '01100', '01010', '01001', '00110', '00101',
'00011', '11100', '11010', '11001', '10110', '10101', '10011',
'01110', '01101', '01011', '00111', '11110', '11101', '11011',
'10111', '01111', '11111']
else:
return 'The maximum number of sets supported is 5.'
ssets = subsets(N)
# Subset counts
text = dict(zip(ssets.keys(), map(len, ssets.values())))
if pct:
total = float(sum(text.values()))
text = dict(zip(text.keys(),
np.round(100 * np.array(text.values()) / total,
decimals=2)))
fig, ax = plt.subplots(figsize=figsize)
for i in range(len(N)):
ellipse(ax, x[i], y[i], w[i], h[i], a[i], alpha=.25, color=c[i],
label='%s (%d)' % (labels[i], len(N[i])) if sizes
else labels[i])
for i in range(len(text)):
ax.text(xt[i], yt[i], text[keys[i]], fontdict={'ha': 'center'})
ax.set_xlim(-1.5, 1.5)
ax.set_ylim(-1.5, 1.5)
ax.set_title(title)
ax.legend()
ax.axis('off')
fig.tight_layout()
if filename:
fig.savefig(filename)
else:
return fig
def test_two_sets_int(self):
self.assertDictEqual(iterables.subsets([{0, 1, 2}, {2, 3, 4}]), {
'11': set([2]),
'10': set([0, 1]),
'01': set([3, 4])
})