def test_summary_graph_loop():
n = int(RESAMPLED_POINTS)
# Resampling shouldn't get stuck in an infinite loop
g = Graph('foo', {}, {})
for j in range(n):
g.add_data_point(j, 0.1)
g = make_summary_graph([g])
data = g.get_data()
assert len(data) == 1
assert data[0][0] == n
assert abs(data[0][1] - 0.1) < 1e-7
# Resampling should work with long integers
g = Graph('foo', {}, {})
k0 = 2**80
for j in range(2*int(RESAMPLED_POINTS)):
g.add_data_point(k0 + j, 0.1)
g = make_summary_graph([g])
g.get_data()
def test_summary_graph():
n = 2 * int(RESAMPLED_POINTS)
g = Graph('foo', {})
for i in range(n):
g.add_data_point(i, 0.1)
g.add_data_point(n + i, 0.2)
g = make_summary_graph([g])
data = g.get_data()
assert len(data) == 512
for i in range(256):
assert abs(data[i][1] - 0.1) < 1e-7
assert abs(data[256 + i][1] - 0.2) < 1e-7
def test_summary_graph():
n = 2 * int(RESAMPLED_POINTS)
g = Graph('foo', {})
for i in range(n):
g.add_data_point(i, 0.1)
g.add_data_point(n + i, 0.2)
g = make_summary_graph([g])
data = g.get_data()
assert len(data) == 512
for i in range(256):
assert abs(data[i][1] - 0.1) < 1e-7
assert abs(data[256 + i][1] - 0.2) < 1e-7
def test_graph_single():
vals = [
(1, 1, None),
(2, 2, 1),
(3, 3, None),
(4, 4, 7),
(5, 5, None),
(6, None, None),
(7, float('nan'), None),
]
# Should give same data back, excluding missing values at edges
g = Graph('foo', {})
for k, v, dv in vals:
g.add_data_point(k, v, dv)
data = g.get_data()
assert data == vals[:-2]
# Should average duplicate values
g = Graph('foo', {})
g.add_data_point(4, 3)
for k, v, dv in vals:
g.add_data_point(k, v, dv)
g.add_data_point(4, 5)
data = g.get_data()
assert data[3][0] == 4
assert abs(data[3][1] - (3 + 4 + 5)/3.) < 1e-10
# Summary graph should be the same as the main graph
g = Graph('foo', {})
for k, v, dv in vals:
g.add_data_point(k, v, dv)
g = make_summary_graph([g])
data = g.get_data()
assert len(data) == len(vals) - 2
for v, d in zip(vals, data):
kv, xv, dv = v
kd, xd, dd = d
assert kv == kd
assert dd is None
assert abs(xv - xd) < 1e-10
def test_graph_single():
vals = [
(1, 1, None),
(2, 2, 1),
(3, 3, None),
(4, 4, 7),
(5, 5, None),
(6, None, None),
(7, float('nan'), None),
]
# Should give same data back, excluding missing values at edges
g = Graph('foo', {})
for k, v, dv in vals:
g.add_data_point(k, v, dv)
data = g.get_data()
assert data == vals[:-2]
# Should average duplicate values
g = Graph('foo', {})
g.add_data_point(4, 3)
for k, v, dv in vals:
g.add_data_point(k, v, dv)
g.add_data_point(4, 5)
data = g.get_data()
assert data[3][0] == 4
assert abs(data[3][1] - (3 + 4 + 5) / 3.) < 1e-10
# Summary graph should be the same as the main graph
g = Graph('foo', {})
for k, v, dv in vals:
g.add_data_point(k, v, dv)
g = make_summary_graph([g])
data = g.get_data()
assert len(data) == len(vals) - 2
for v, d in zip(vals, data):
kv, xv, dv = v
kd, xd, dd = d
assert kv == kd
assert dd is None
assert abs(xv - xd) < 1e-10
def test_graph_multi():
vals = [
(0, [None, None, None]),
(1, [1, None, float('nan')]),
(2, [2, 5, 4]),
(3, [3, 4, -60]),
(4, [4, 3, 2]),
(5, [None, 2, None]),
(6, [6, 1, None])
]
filled_vals = [
(1, [1, 5, 4]),
(2, [2, 5, 4]),
(3, [3, 4, -60]),
(4, [4, 3, 2]),
(5, [4, 2, 2]),
(6, [6, 1, 2])
]
# Should give same data back, with missing data at edges removed
g = Graph('foo', {})
for k, v in vals:
g.add_data_point(k, v)
data = g.get_data()
assert data[0] == (1, [1, None, None])
assert data[1:] == vals[2:]
# Should average duplicate values
g = Graph('foo', {})
g.add_data_point(4, [1, 2, 3])
for k, v in vals:
g.add_data_point(k, v)
g.add_data_point(4, [3, 2, 1])
data = g.get_data()
assert data[3][0] == 4
assert abs(data[3][1][0] - (1 + 4 + 3)/3.) < 1e-10
assert abs(data[3][1][1] - (2 + 3 + 2)/3.) < 1e-10
assert abs(data[3][1][2] - (3 + 2 + 1)/3.) < 1e-10
# The summary graph is obtained by geometric mean of filled data
g = Graph('foo', {})
for k, v in vals:
g.add_data_point(k, v)
g = make_summary_graph([g])
data = g.get_data()
for v, d in zip(filled_vals, data):
kv, xvs = v
kd, xd = d
assert kv == kd
# geom mean, with some sign convention
expected = _sgn(sum(xvs)) * (abs(xvs[0]*xvs[1]*xvs[2]))**(1./3)
assert abs(xd - expected) < 1e-10
# Test summary over separate graphs -- should behave as if the
# data was in a single graph
g0 = Graph('foo', {})
g1 = Graph('foo', {})
g2 = Graph('foo', {})
for k, v in vals:
g0.add_data_point(k, v)
g1.add_data_point(k, v[0])
g2.add_data_point(k, v[1:])
data0 = make_summary_graph([g0]).get_data()
data = make_summary_graph([g1, g2]).get_data()
assert data == data0
# Check the above is true regardless if some x-values are missing
g0.add_data_point(7, [None, 1, None])
g2.add_data_point(7, [1, None])
g0.add_data_point(4.5, [9, None, None])
g1.add_data_point(4.5, 9)
data0 = make_summary_graph([g0]).get_data()
data = make_summary_graph([g1, g2]).get_data()
assert data == data0
def test_graph_multi():
vals = [
(0, [None, None, None]),
(1, [1, None, float('nan')]),
(2, [2, 5, 4]),
(3, [3, 4, -60]),
(4, [4, 3, 2]),
(5, [None, 2, None]),
(6, [6, 1, None])
]
filled_vals = [
(1, [1, 5, 4]),
(2, [2, 5, 4]),
(3, [3, 4, -60]),
(4, [4, 3, 2]),
(5, [4, 2, 2]),
(6, [6, 1, 2])
]
# Should give same data back, with missing data at edges removed
g = Graph('foo', {}, {})
for k, v in vals:
g.add_data_point(k, v)
data = g.get_data()
assert data[0] == (1, [1, None, None])
assert data[1:] == vals[2:]
# Should average duplicate values
g = Graph('foo', {}, {})
g.add_data_point(4, [1, 2, 3])
for k, v in vals:
g.add_data_point(k, v)
g.add_data_point(4, [3, 2, 1])
data = g.get_data()
assert data[3][0] == 4
assert abs(data[3][1][0] - (1 + 4 + 3)/3.) < 1e-10
assert abs(data[3][1][1] - (2 + 3 + 2)/3.) < 1e-10
assert abs(data[3][1][2] - (3 + 2 + 1)/3.) < 1e-10
# The summary graph is obtained by geometric mean of filled data
g = Graph('foo', {}, {})
for k, v in vals:
g.add_data_point(k, v)
g = make_summary_graph([g])
data = g.get_data()
for v, d in zip(filled_vals, data):
kv, xvs = v
kd, xd = d
assert kv == kd
# geom mean, with some sign convention
expected = _sgn(sum(xvs)) * (abs(xvs[0]*xvs[1]*xvs[2]))**(1./3)
assert abs(xd - expected) < 1e-10
# Test summary over separate graphs -- should behave as if the
# data was in a single graph
g0 = Graph('foo', {}, {})
g1 = Graph('foo', {}, {})
g2 = Graph('foo', {}, {})
for k, v in vals:
g0.add_data_point(k, v)
g1.add_data_point(k, v[0])
g2.add_data_point(k, v[1:])
data0 = make_summary_graph([g0]).get_data()
data = make_summary_graph([g1, g2]).get_data()
assert data == data0
# Check the above is true regardless if some x-values are missing
g0.add_data_point(7, [None, 1, None])
g2.add_data_point(7, [1, None])
g0.add_data_point(4.5, [9, None, None])
g1.add_data_point(4.5, 9)
data0 = make_summary_graph([g0]).get_data()
data = make_summary_graph([g1, g2]).get_data()
assert data == data0
