def test_velocity_dispersion(self):
# produce a test cluster
tab = random_cluster(3, id='i', x='f', y='f', z='f')
tab['vx'] = [1., 1., 1.]
tab['vy'] = [1., 2., 2.]
tab['vz'] = [1., 2., 2.]
# 3D
s = gcp.Snapshot(tab)
sig2 = (np.std(s['vx'])**2., np.std(s['vy'])**2., np.std(s['vz'])**2.)
assert gcp.velocity_dispersion(s)[0] == np.sqrt(np.sum(sig2))
# 2D
sig2 = (np.std(s['vx'])**2., np.std(s['vy'])**2.)
assert gcp.velocity_dispersion(s, dim=2)[0] == np.sqrt(np.sum(sig2))
# 2D + isotropy
sig2 = (np.std(s['vx'])**2., np.std(s['vy'])**2., np.std(s['vz'])**2.)
sig_to_test = gcp.velocity_dispersion(s, dim=2, isotropy=True)[0]
assert sig_to_test == (np.sqrt(np.sum(sig2) * 2. / 3.))
# 1D
sig2 = (np.std(s['vz'])**2.)
assert gcp.velocity_dispersion(s, dim=1)[0] == np.sqrt(np.sum(sig2))
# # masked
# s_mskd = gcp.Snapshot(tab)
# s_mskd.filter(id=(0,1), by_range={"id":False})
# assert gcp.velocity_dispersion(s_mskd)[0] == np.sqrt(np.std([3.,9.]))
# empty cluster
tab = Table(names=('x', 'y', 'z', 'vx', 'vy', 'vz'))
s = gcp.Snapshot(tab)
assert np.isnan(gcp.velocity_dispersion(s)[0])
assert np.isnan(gcp.velocity_dispersion(s)[1])
def test_randomization(self):
# Test rotation (only coordinates)
tab = random_cluster(1000, x='f', y='f', z='f')
# original cluster
s1 = gcp.Snapshot(tab)
# randomized cluster
s2 = gcp.Snapshot(tab, seed=999)
# x, y and z coordinates must all differ
assert np.all(s1['x'] != s2['x'])
assert np.all(s1['y'] != s2['y'])
assert np.all(s1['z'] != s2['z'])
# but not the radii (allowing a 1e-15 precision error)
assert np.all(np.abs(s1['_r'] - s2['_r']) < np.zeros(len(tab)) + 1e-15)
# Test rotation (also velocities)
tab = random_cluster(1000, x='f', y='f', z='f', vx='f', vy='f', vz='f')
# original cluster
s1 = gcp.Snapshot(tab)
# randomized cluster
s2 = gcp.Snapshot(tab, seed=9)
# x, y and z coordinates and velocities must all differ
assert np.all(s1['x'] != s2['x'])
assert np.all(s1['y'] != s2['y'])
assert np.all(s1['z'] != s2['z'])
assert np.all(s1['vx'] != s2['vx'])
assert np.all(s1['vy'] != s2['vy'])
assert np.all(s1['vz'] != s2['vz'])
# but not the radii (allowing a 1e-15 precision error)
assert np.all(np.abs(s1['_r'] - s2['_r']) < np.zeros(len(tab)) + 1e-15)
# and not the velocities squared
v1 = np.sqrt(s1['vx']**2. + s1['vy']**2. + s1['vz']**2.)
v2 = np.sqrt(s2['vx']**2. + s2['vy']**2. + s2['vz']**2.)
assert np.all(np.abs(v1 - v2) < np.zeros(len(tab)) + 1e-15)
def test_sort(self):
tab = random_cluster(50, id='i', m='f', x='f', y='f', z='f')
m_sorted = np.sort(tab['m'])
s = gcp.Snapshot(tab)
s.sort('m')
assert np.all(s['m'] == m_sorted)
def test_mass_function(self):
# produce a test cluster with 2 star populations
tab = random_cluster(100, id='i', x='f', y='f', z='f')
tab["m"] = np.ones(len(tab))
tab["m"][50:] = 2.
s = gcp.Snapshot(tab)
assert np.all(gcp.mass_function(s, bins=2)[0] == np.array([50, 50]))
def test_default_init(self):
# 1. right construction
tab = random_cluster(100,
id='i',
kw=14,
x='f',
y='f',
z='f',
vx='f',
vy='f',
vz='f')
s = gcp.Snapshot(tab)
assert (len(s.original) == len(tab))
# 2. wrong construction
with pytest.raises(ValueError):
s = gcp.Snapshot([np.arange(100)], names=['m'])
def test_lagr_radii(self):
tab = random_cluster(100, x='f', y='f', z='f')
s = gcp.Snapshot(tab)
# what if an incomplete cluster is passed?
with pytest.raises(ValueError):
gcp.lagr_rad(s, 50.)
# produce a realistic GC with rh = 2.5pc
rh = 2.5
tab = limepy_cluster(10000,
seed=np.random.randint(0, 999),
M=33070,
rh=rh)
# pass it to an instance of Snapshot
s = gcp.Snapshot(tab)
# test if numbers make sense, namely the relative error for the
# half-mass radius of a N=10K better be less than 10%
rh_calc = gcp.lagr_rad(s, 50.)
assert abs(2. * (rh_calc - rh) / (rh_calc + rh)) < 0.1
# what if the lagr. percentage is 100 or greater ?
assert gcp.lagr_rad(s, 100.) >= np.max(s["_r"])
assert gcp.lagr_rad(s, 105.) == gcp.lagr_rad(s, 100.)
s.filter(_r=(0., gcp.lagr_rad(s, 100.)))
assert len(s[:]) == 10000
# what if the lagr. radii are light based, but no luminosity is given?
with pytest.raises(ValueError):
gcp.lagr_rad(s, 50., light=True)
# what if an almost empty cluster is passed?
tab = random_cluster(2, m='f', x='f', y='f', z='f')
s = gcp.Snapshot(tab)
rh = gcp.lagr_rad(s, 50.)
assert np.isnan(rh)
# what if a small (but not almost empty) cluster is passed?
tab = random_cluster(3, m='f', x='f', y='f', z='f')
s = gcp.Snapshot(tab)
lagrs = gcp.lagr_rad(s, np.arange(100.))
assert not np.all(np.isnan(lagrs))
def test_com(self):
m = (1., 1., 1.)
x = (0., 2., 4.)
y = (0., 0., 0.)
# intrinsic com
z = (0., 0., 0.)
s = gcp.Snapshot(data=[m, x, y, z], names=['m', 'x', 'y', 'z'])
assert gcp.center_of_mass(s) == [2., 0., 0.]
# masked com
i = (0, 1, 2)
s = gcp.Snapshot(data=[i, m, x, y, z],
names=['id', 'm', 'x', 'y', 'z'])
s.filter(id=(0, 1), by_range={"id": False})
assert gcp.center_of_mass(s, masked=True) == [1., 0., 0.]
# projected com
z = (8., 10., -99.)
s = gcp.Snapshot(data=[m, x, y, z],
names=['m', 'x', 'y', 'z'],
project=True)
assert gcp.center_of_mass(s) == [2., 0.]
def test_density_center(self):
# produce a test cluster with 10 stars (separated by 1 pc)
# with equal masses (=1 Msun)
tab = Table()
tab['x'] = np.arange(0., 10., 1) # all the stars along x
tab['y'] = np.zeros_like(tab['x']) * 0.
tab['z'] = np.zeros_like(tab['x']) * 0.
tab['m'] = np.ones_like(tab['x']) * 1. # m = 1 for each star
s = gcp.Snapshot(tab)
s.add_local_density()
rho = s["_rho"]
dc = gcp.density_center(s)
assert dc[0] == np.dot(tab['x'], rho) / np.sum(rho)
def test_local_density(self):
# produce a test cluster with 10 stars (separated by 1 pc)
# with equal masses (=1 Msun)
tab = Table()
tab['x'] = np.arange(0., 10., 1) # all the stars along x
tab['y'] = np.zeros_like(tab['x']) * 0.
tab['z'] = np.zeros_like(tab['x']) * 0.
tab['m'] = np.ones_like(tab['x']) * 1. # m = 1 for each star
# not masked
s = gcp.Snapshot(tab)
s.add_local_density()
neighbors_radii = np.array([6., 5., 4., 3., 3., 3., 3., 4., 5., 6.])
rho = 5. / (4. / 3 * np.pi * neighbors_radii**3.)
assert all([s["_rho"][i] == rho[i] for i in range(10)])
# masked
s_mskd = gcp.Snapshot(tab)
s_mskd.filter(x=(0., 8.1))
s_mskd.add_local_density()
neighbors_radii = np.array([6., 5., 4., 3., 3., 3., 4., 5., 6.])
rho = 5. / (4. / 3 * np.pi * neighbors_radii**3.)
assert all([s_mskd["_rho"][i] == rho[i] for i in [6, 7, 8]])
def test_filter(self):
tab = random_cluster(10000,
id='i',
kw=14,
m='f',
x='f',
y='f',
z='f',
vx='f',
vy='f',
vz='f')
# 1. default selection
s = gcp.Snapshot(tab)
s.filter(m=(0.2, 0.5))
# are elements in a given range ?
assert np.all(s['m'] < 0.5)
assert np.all(s['m'] >= 0.2)
# 2. overwriting selection
s.filter(m=(0.6, 1.0))
# are element in a range different from before ?
assert np.all(s['m'] < 1.0)
assert np.all(s['m'] >= 0.6)
# 3. inplace selection
s.filter(m=(0.6, 1.0),
id=np.arange(50),
by_range={
'm': True,
'id': False
})
# are element in the same range as before, but also with the new
# selection criterium applied ?
assert np.all(s['m'] < 1.0)
assert np.all(s['m'] >= 0.6)
assert np.all(s['id'] <= 50)
# 4. wrong selection
# 4.a wrong key
with pytest.raises(KeyError):
s.filter(mass=(0., 0.5))
# 4.b wrong values
with pytest.raises(ValueError):
s.filter(m=0.5)
with pytest.raises(ValueError):
s.filter(m='0.5')
with pytest.raises(ValueError):
s.filter(m=['0.5', '0.8'])
with pytest.raises(ValueError):
s.filter(m=[0.8, 1.0, 1.4])
def test_density_radius(self):
# produce a test cluster with 10 stars (separated by 1 pc)
# with equal masses (=1 Msun)
tab = Table()
tab['x'] = np.arange(0., 10., 1) # all the stars along x
tab['y'] = np.zeros_like(tab['x']) * 0.
tab['z'] = np.zeros_like(tab['x']) * 0.
tab['m'] = np.ones_like(tab['x']) * 1. # m = 1 for each star
s = gcp.Snapshot(tab)
s.add_local_density()
rho = s["_rho"]
# default density radius
rd = gcp.density_radius(s)
assert rd == np.dot(tab['x'], rho * rho) / np.sum(rho * rho)
# mass weighted density radius
rd = gcp.density_radius(s, mass_weighted=True)
assert rd == np.dot(tab['x'], rho) / np.sum(rho)
def test_density(self):
# produce a test cluster with 5 stars (separated by 1 pc)
# with equal masses (=1) and luminosity (=1)
tab = Table()
tab['x'] = np.arange(0., 5., 1) # all the stars along x
tab['y'] = np.zeros_like(tab['x']) * 0.
tab['z'] = np.zeros_like(tab['x']) * 0.
tab['m'] = np.ones_like(tab['x']) * 2. # m = 1 for each star
s = gcp.Snapshot(tab)
# intrinsic number count
assert gcp.density(s) == 5. / (4. / 3. * np.pi * tab['x'][-1]**3.)
# intrinsic mass
assert gcp.density(s, quantity='m')== np.sum(tab['m']) / \
(4./3. * np.pi * tab['x'][-1]**3.)
# wrong quantity for density
with pytest.raises(ValueError):
tmp = gcp.density(s, quantity='xxxx')
# masked cluster
s.filter(_r=(0., 3.1))
assert gcp.density(s) == 4. / (4. / 3. * np.pi * tab['x'][-2]**3.)