def intersect_rows(array1, array2, index=None):
"""Return intersection of rows"""
if (array1.shape[0] == 0):
if index == True:
return (array1, sp.zeros((0, )), sp.zeros((0, )))
else:
return array1
if (array2.shape[0] == 0):
if index == True:
return (array2, sp.zeros((0, )), sp.zeros((0, )))
else:
return array2
array1_v = array1.view([('', array1.dtype)] * array1.shape[1])
array2_v = array2.view([('', array2.dtype)] * array2.shape[1])
array_i = sp.intersect1d(array1_v, array2_v)
if index == True:
a1_i = sp.where(sp.in1d(array1_v, array_i))[0]
a2_i = sp.where(sp.in1d(array2_v, array_i))[0]
return (array_i.view(array1.dtype).reshape(array_i.shape[0],
array1.shape[1]), a1_i,
a2_i)
else:
return array_i.view(array1.dtype).reshape(array_i.shape[0],
array1.shape[1])
def _generate_masked_mesh(self, cell_mask=None):
r"""
Generates the mesh based on the cell mask provided
"""
#
if cell_mask is None:
cell_mask = sp.ones(self.data_map.shape, dtype=bool)
#
# initializing arrays
self._edges = sp.ones(0, dtype=str)
self._merge_patch_pairs = sp.ones(0, dtype=str)
self._create_blocks(cell_mask)
#
# building face arrays
mapper = sp.ravel(sp.array(cell_mask, dtype=int))
mapper[mapper == 1] = sp.arange(sp.count_nonzero(mapper))
mapper = sp.reshape(mapper, (self.nz, self.nx))
mapper[~cell_mask] = -sp.iinfo(int).max
#
boundary_dict = {
'bottom':
{'bottom': mapper[0, :][cell_mask[0, :]]},
'top':
{'top': mapper[-1, :][cell_mask[-1, :]]},
'left':
{'left': mapper[:, 0][cell_mask[:, 0]]},
'right':
{'right': mapper[:, -1][cell_mask[:, -1]]},
'front':
{'front': mapper[cell_mask]},
'back':
{'back': mapper[cell_mask]},
'internal':
{'bottom': [], 'top': [], 'left': [], 'right': []}
}
#
# determining cells linked to a masked cell
cell_mask = sp.where(~sp.ravel(cell_mask))[0]
inds = sp.in1d(self._field._cell_interfaces, cell_mask)
inds = sp.reshape(inds, (len(self._field._cell_interfaces), 2))
inds = inds[:, 0].astype(int) + inds[:, 1].astype(int)
inds = (inds == 1)
links = self._field._cell_interfaces[inds]
#
# adjusting order so masked cells are all on links[:, 1]
swap = sp.in1d(links[:, 0], cell_mask)
links[swap] = links[swap, ::-1]
#
# setting side based on index difference
sides = sp.ndarray(len(links), dtype='<U6')
sides[sp.where(links[:, 1] == links[:, 0]-self.nx)[0]] = 'bottom'
sides[sp.where(links[:, 1] == links[:, 0]+self.nx)[0]] = 'top'
sides[sp.where(links[:, 1] == links[:, 0]-1)[0]] = 'left'
sides[sp.where(links[:, 1] == links[:, 0]+1)[0]] = 'right'
#
# adding each block to the internal face dictionary
inds = sp.ravel(mapper)[links[:, 0]]
for side, block_id in zip(sides, inds):
boundary_dict['internal'][side].append(block_id)
self.set_boundary_patches(boundary_dict, reset=True)
def make_unique_by_event(event_list):
# function event_list = make_unique_by_event(event_list)
#
# This script removes all events that share the sam alternative evnt coordinates
# but differ in the flanking size. The longest of several equal events is kept.
rm_idx = []
last_kept = 0
for i in range(1, event_list.shape[0]):
if i % 1000 == 0:
print '.',
if i % 10000 == 0:
print '%i' % i
old_coords = event_list[last_kept].get_inner_coords(trafo=True)
curr_coords = event_list[i].get_inner_coords(trafo=True)
if old_coords.shape[0] == curr_coords.shape[0] and sp.all(
old_coords == curr_coords):
### assertion that we did everything right
assert (event_list[last_kept].chr == event_list[i].chr)
assert (event_list[last_kept].strand == event_list[i].strand)
### check, which event is longer -> keep longer event
len1 = event_list[last_kept].get_len()
len2 = event_list[i].get_len()
if len1 > len2:
keep_idx = last_kept
not_keep_idx = i
else:
keep_idx = i
not_keep_idx = last_kept
### check if we would loose strains
idx = sp.where(~sp.in1d(event_list[not_keep_idx].strain,
event_list[keep_idx].strain))[0]
if idx.shape[0] > 0:
event_list[keep_idx].strain = sp.r_[
event_list[keep_idx].strain,
event_list[not_keep_idx].strain[idx]]
### TODO !!!!!!!!!!!!! make sure that we keep different coordinates if the strains differ ...
event_list[keep_idx].gene_name = sp.union1d(
event_list[keep_idx].gene_name,
event_list[not_keep_idx].gene_name)
rm_idx.append(not_keep_idx)
last_kept = keep_idx
else:
last_kept = i
print 'events dropped: %i' % len(rm_idx)
keep_idx = sp.where(~sp.in1d(sp.arange(event_list.shape[0]), rm_idx))[0]
event_list = event_list[keep_idx]
return event_list
def plot_overlap_ps(result_file, ss_file='/Users/bjarnivilhjalmsson/data/GIANT/GIANT_HEIGHT_Wood_et_al_2014_publicrelease_HapMapCeuFreq.txt',
fig_filename='/Users/bjarnivilhjalmsson/data/tmp/manhattan_combPC_HGT.png', method='combPC',
ylabel='Comb. PC (HIP,WC,HGT,BMI) $-log_{10}(P$-value$)$', xlabel='Height $-log_{10}(P$-value$)$', p_thres=0.00001):
# Parse results ans SS file
res_table = pandas.read_table(result_file)
ss_table = pandas.read_table(ss_file)
# Parse
res_sids = sp.array(res_table['SNPid'])
if method == 'MVT':
comb_ps = sp.array(res_table['pval'])
elif method == 'combPC':
comb_ps = sp.array(res_table['combPC'])
if 'MarkerName' in ss_table.keys():
ss_sids = sp.array(ss_table['MarkerName'])
elif 'SNP' in ss_table.keys():
ss_sids = sp.array(ss_table['SNP'])
else:
raise Exception("Don't know where to look for rs IDs")
marg_ps = sp.array(ss_table['p'])
# Filtering boring p-values
res_p_filter = comb_ps < p_thres
res_sids = res_sids[res_p_filter]
comb_ps = comb_ps[res_p_filter]
# ss_p_filter = marg_ps<p_thres
# ss_sids = ss_sids[ss_p_filter]
# marg_ps = marg_ps[ss_p_filter]
common_sids = sp.intersect1d(res_sids, ss_sids)
print 'Found %d SNPs in common' % (len(common_sids))
ss_filter = sp.in1d(ss_sids, common_sids)
res_filter = sp.in1d(res_sids, common_sids)
ss_sids = ss_sids[ss_filter]
res_sids = res_sids[res_filter]
marg_ps = marg_ps[ss_filter]
comb_ps = comb_ps[res_filter]
print 'Now sorting'
ss_index = sp.argsort(ss_sids)
res_index = sp.argsort(res_sids)
marg_ps = -sp.log10(marg_ps[ss_index])
comb_ps = -sp.log10(comb_ps[res_index])
with plt.style.context('fivethirtyeight'):
plt.plot(marg_ps, comb_ps, 'b.', alpha=0.2)
(x_min, x_max) = plt.xlim()
(y_min, y_max) = plt.ylim()
plt.plot([x_min, x_max], [y_min, y_max], 'k--', alpha=0.2)
plt.ylabel(ylabel)
plt.xlabel(xlabel)
plt.tight_layout()
plt.savefig(fig_filename)
plt.clf()
def test_trim_extend():
pn = OpenPNM.Network.Cubic(shape=[5, 5, 5])
assert sp.all(sp.in1d(pn.find_neighbor_pores(pores=0), [1, 5, 25]))
assert [pn.Np, pn.Nt] == [125, 300]
pn.trim(pores=[0])
assert sp.all(sp.in1d(pn.find_neighbor_pores(pores=0), [1, 5, 25]))
assert [pn.Np, pn.Nt] == [124, 297]
pn.extend(pore_coords=[0, 0, 0], throat_conns=[[124, 0]])
assert [pn.Np, pn.Nt] == [125, 298]
assert sp.all(sp.in1d(pn.find_neighbor_pores(pores=0), [1, 5, 25, 124]))
def test_trim_extend():
pn = OpenPNM.Network.Cubic(shape=[5, 5, 5])
assert sp.all(sp.in1d(pn.find_neighbor_pores(pores=0), [1, 5, 25]))
assert [pn.Np, pn.Nt] == [125, 300]
pn.trim(pores=[0])
assert sp.all(sp.in1d(pn.find_neighbor_pores(pores=0), [1, 5, 25]))
assert [pn.Np, pn.Nt] == [124, 297]
pn.extend(pore_coords=[0, 0, 0], throat_conns=[[124, 0]])
assert [pn.Np, pn.Nt] == [125, 298]
assert sp.all(sp.in1d(pn.find_neighbor_pores(pores=0), [1, 5, 25, 124]))
def make_unique_by_event(event_list):
# function event_list = make_unique_by_event(event_list)
#
# This script removes all events that share the sam alternative evnt coordinates
# but differ in the flanking size. The longest of several equal events is kept.
rm_idx = []
last_kept = 0
for i in range(1, event_list.shape[0]):
if i % 1000 == 0:
print '.',
if i % 10000 == 0:
print '%i' % i
old_coords = event_list[last_kept].get_inner_coords(trafo=True)
curr_coords = event_list[i].get_inner_coords(trafo=True)
if old_coords.shape[0] == curr_coords.shape[0] and sp.all(old_coords == curr_coords):
### assertion that we did everything right
assert(event_list[last_kept].chr == event_list[i].chr)
assert(event_list[last_kept].strand == event_list[i].strand)
### check, which event is longer -> keep longer event
len1 = event_list[last_kept].get_len()
len2 = event_list[i].get_len()
if len1 > len2:
keep_idx = last_kept
not_keep_idx = i
else:
keep_idx = i
not_keep_idx = last_kept
### check if we would loose strains
idx = sp.where(~sp.in1d(event_list[not_keep_idx].strain, event_list[keep_idx].strain))[0]
if idx.shape[0] > 0:
event_list[keep_idx].strain = sp.r_[event_list[keep_idx].strain, event_list[not_keep_idx].strain[idx]]
### TODO !!!!!!!!!!!!! make sure that we keep different coordinates if the strains differ ...
event_list[keep_idx].gene_name = sp.union1d(event_list[keep_idx].gene_name, event_list[not_keep_idx].gene_name)
rm_idx.append(not_keep_idx)
last_kept = keep_idx
else:
last_kept = i
print 'events dropped: %i' % len(rm_idx)
keep_idx = sp.where(~sp.in1d(sp.arange(event_list.shape[0]), rm_idx))[0]
event_list = event_list[keep_idx]
return event_list
def calculate_hapmap_pcs(hapmap_file, pc_weights_dict, snps_filter=None):
"""
Calculates the principal components for the hapmap project
:param hapmap_file: Hapmap file in HDF5 format
:param pc_weights_dict: dictionary with SNP weights (key = snpid)
:param snps_filter: list of snp-ids to subset (optional)
:return: dictionary with pcs and number of snps that were used
"""
log.info('Calculating Principal components for Hapmap file %s' % hapmap_file)
ok_sids = np.asarray(list(pc_weights_dict.keys()))
log.info('Loaded PC weight for %d SNPs' % (len(ok_sids)))
# Load genotypes
log.info('Load Hapmap dataset')
h5f = h5py.File(hapmap_file, 'r')
num_indivs = len(h5f['indivs']['continent'][...])
log.info('Found genotypes for %d individuals' % num_indivs)
pcs = sp.zeros((num_indivs, 2))
num_nt_issues = 0
num_snps_used = 0
log.info('Calculating PCs')
for chrom in range(1, 23):
log.info('Working on Chromosome %d' % chrom)
chrom_str = 'chr%d' % chrom
log.info('Identifying overlap')
ok_snp_filter = sp.in1d(ok_sids, snps_filter[chrom_str])
ok_chrom_sids = ok_sids.compress(ok_snp_filter, axis=0)
sids = h5f[chrom_str]['variants']['ID'][...]
ok_snp_filter = sp.in1d(sids, ok_chrom_sids)
# assert sids[ok_snp_filter]==ok_sids, 'WTF?'
sids = sids.compress(ok_snp_filter, axis=0)
log.info('Loading SNPs')
snps = h5f[chrom_str]['calldata']['snps'][...]
snps = snps.compress(ok_snp_filter, axis=0)
length = len(h5f[chrom_str]['variants/REF'])
nts = np.hstack((h5f[chrom_str]['variants/REF'][:].reshape(length, 1),
h5f[chrom_str]['variants/ALT'][:].reshape(length, 1)))
nts = nts.compress(ok_snp_filter, axis=0)
log.info('Updating PCs')
pcs_per_chr = _calc_pcs(pc_weights_dict, sids, nts, snps)
pcs += pcs_per_chr['pcs']
num_nt_issues += pcs_per_chr['num_nt_issues']
num_snps_used += pcs_per_chr['num_snps_used']
h5f.close()
log.info('%d SNPs were excluded from the analysis due to nucleotide issues.' % (num_nt_issues))
log.info('%d SNPs were used for the analysis.' % (num_snps_used))
return {'pcs': pcs, 'num_snps_used': num_snps_used}
def intersect_rows(array1, array2, index=False):
"""Return array with rows that intersect between array1 and array2"""
tmp1 = sp.array(['-'.join(array1[i, :].astype('str')) for i in range(array1.shape[0])])
tmp2 = sp.array(['-'.join(array2[i, :].astype('str')) for i in range(array2.shape[0])])
idx = sp.where(sp.in1d(tmp1, tmp2))[0]
if index:
idx2 = sp.where(sp.in1d(tmp2, tmp1))[0]
if index:
return (array1[idx, :], idx, idx2)
else:
return (array1[idx, :], None, None)
def find_interface_throats(self, labels=[]):
r"""
Finds the throats that join two pore labels.
Parameters
----------
labels : list of strings
The labels of the two pore groups whose interface is sought
Returns
-------
An array of throat numbers that connect the given pore groups
Notes
-----
This method is meant to find interfaces between TWO groups, regions or
clusters of pores (as defined by their label). If the input labels
overlap or are not adjacent, an empty array is returned.
Examples
--------
>>> import OpenPNM
>>> pn = OpenPNM.Network.TestNet()
>>> pn['pore.domain1'] = False
>>> pn['pore.domain2'] = False
>>> pn['pore.domain1'][[0, 1, 2]] = True
>>> pn['pore.domain2'][[5, 6, 7]] = True
>>> pn.find_interface_throats(labels=['domain1', 'domain2'])
array([1, 4, 7])
TODO: It might be a good idea to allow overlapping regions
"""
Tind = sp.array([], ndmin=1)
if sp.shape(labels)[0] != 2:
logger.error('Exactly two labels must be given')
pass
else:
P1 = self.pores(labels=labels[0])
P2 = self.pores(labels=labels[1])
# Check if labels overlap
if sp.sum(sp.in1d(P1, P2)) > 0:
logger.error('Some labels overlap, iterface cannot be found')
pass
else:
T1 = self.find_neighbor_throats(P1)
T2 = self.find_neighbor_throats(P2)
Tmask = sp.in1d(T1, T2)
Tind = T1[Tmask]
return Tind
def find_interface_throats(self, labels=[]):
r"""
Finds the throats that join two pore labels.
Parameters
----------
labels : list of strings
The labels of the two pore groups whose interface is sought
Returns
-------
An array of throat numbers that connect the given pore groups
Notes
-----
This method is meant to find interfaces between TWO groups, regions or
clusters of pores (as defined by their label). If the input labels
overlap or are not adjacent, an empty array is returned.
Examples
--------
>>> import OpenPNM
>>> pn = OpenPNM.Network.TestNet()
>>> pn['pore.domain1'] = False
>>> pn['pore.domain2'] = False
>>> pn['pore.domain1'][[0, 1, 2]] = True
>>> pn['pore.domain2'][[5, 6, 7]] = True
>>> pn.find_interface_throats(labels=['domain1', 'domain2'])
array([1, 4, 7])
TODO: It might be a good idea to allow overlapping regions
"""
Tind = sp.array([], ndmin=1)
if sp.shape(labels)[0] != 2:
logger.error('Exactly two labels must be given')
pass
else:
P1 = self.pores(labels=labels[0])
P2 = self.pores(labels=labels[1])
# Check if labels overlap
if sp.sum(sp.in1d(P1, P2)) > 0:
logger.error('Some labels overlap, iterface cannot be found')
pass
else:
T1 = self.find_neighbor_throats(P1)
T2 = self.find_neighbor_throats(P2)
Tmask = sp.in1d(T1, T2)
Tind = T1[Tmask]
return Tind
def on_shifted_dwp_curves(self, t):
a = P4Rm()
if a.AllDataDict['model'] == 0:
temp_1 = arange(2, len(a.ParamDict['dwp'])+1)
temp_2 = temp_1 * t / (len(a.ParamDict['dwp']))
P4Rm.ParamDict['x_dwp'] = t - temp_2
shifted_dwp = a.ParamDict['dwp'][:-1:]
temp_3 = in1d(around(a.ParamDict['depth'], decimals=3),
around(a.ParamDict['x_dwp'], decimals=3))
temp_4 = a.ParamDict['DW_i'][temp_3]
P4Rm.ParamDict['scale_dw'] = shifted_dwp / temp_4
P4Rm.ParamDict['scale_dw'][a.ParamDict['scale_dw'] == 0] = 1.
P4Rm.ParamDict['DW_shifted'] = shifted_dwp/a.ParamDict['scale_dw']
P4Rm.ParamDict['dw_out'] = a.ParamDict['dwp'][-1]
elif a.AllDataDict['model'] == 1:
temp_1 = arange(0, len(a.ParamDict['dwp'])+1-3)
temp_2 = temp_1 * t / (len(a.ParamDict['dwp'])-3)
P4Rm.ParamDict['x_dwp'] = t - temp_2
shifted_dwp = a.ParamDict['dwp'][1:-1:]
temp_3 = in1d(around(a.ParamDict['depth'], decimals=3),
around(a.ParamDict['x_dwp'], decimals=3))
temp_4 = a.ParamDict['DW_i'][temp_3]
P4Rm.ParamDict['scale_dw'] = shifted_dwp / temp_4
P4Rm.ParamDict['scale_dw'][a.ParamDict['scale_dw'] == 0] = 1.
P4Rm.ParamDict['DW_shifted'] = shifted_dwp/a.ParamDict['scale_dw']
temp_5 = array([a.ParamDict['dwp'][0], a.ParamDict['dwp'][-1]])
P4Rm.ParamDict['dw_out'] = temp_5
elif a.AllDataDict['model'] == 2:
x_dw_temp = []
x_dw_temp.append(t*(1-a.ParamDict['dwp'][1]))
x_dw_temp.append(t*(1-a.ParamDict['dwp'][1] +
a.ParamDict['dwp'][2]/2))
x_dw_temp.append(t*(1-a.ParamDict['dwp'][1] -
a.ParamDict['dwp'][3]/2))
x_dw_temp.append(t*0.05)
P4Rm.ParamDict['x_dwp'] = x_dw_temp
y_dw_temp = []
y_dw_temp.append(a.ParamDict['dwp'][0])
y_dw_temp.append(1. - (1-a.ParamDict['dwp'][0])/2)
y_dw_temp.append(1. - (1-a.ParamDict['dwp'][0])/2 -
(1-a.ParamDict['dwp'][6])/2)
y_dw_temp.append(a.ParamDict['dwp'][6])
P4Rm.ParamDict['DW_shifted'] = y_dw_temp
def on_shifted_dwp_curves(self, t):
a = P4Rm()
if a.AllDataDict['model'] == 0:
temp_1 = arange(2, len(a.ParamDict['dwp'])+1)
temp_2 = temp_1 * t / (len(a.ParamDict['dwp']))
P4Rm.ParamDict['x_dwp'] = t - temp_2
shifted_dwp = a.ParamDict['dwp'][:-1:]
temp_3 = in1d(around(a.ParamDict['depth'], decimals=3),
around(a.ParamDict['x_dwp'], decimals=3))
temp_4 = a.ParamDict['DW_i'][temp_3]
P4Rm.ParamDict['scale_dw'] = shifted_dwp / temp_4
P4Rm.ParamDict['scale_dw'][a.ParamDict['scale_dw'] == 0] = 1.
P4Rm.ParamDict['DW_shifted'] = shifted_dwp/a.ParamDict['scale_dw']
P4Rm.ParamDict['dw_out'] = a.ParamDict['dwp'][-1]
elif a.AllDataDict['model'] == 1:
temp_1 = arange(0, len(a.ParamDict['dwp'])+1-3)
temp_2 = temp_1 * t / (len(a.ParamDict['dwp'])-3)
P4Rm.ParamDict['x_dwp'] = t - temp_2
shifted_dwp = a.ParamDict['dwp'][1:-1:]
temp_3 = in1d(around(a.ParamDict['depth'], decimals=3),
around(a.ParamDict['x_dwp'], decimals=3))
temp_4 = a.ParamDict['DW_i'][temp_3]
P4Rm.ParamDict['scale_dw'] = shifted_dwp / temp_4
P4Rm.ParamDict['scale_dw'][a.ParamDict['scale_dw'] == 0] = 1.
P4Rm.ParamDict['DW_shifted'] = shifted_dwp/a.ParamDict['scale_dw']
temp_5 = array([a.ParamDict['dwp'][0], a.ParamDict['dwp'][-1]])
P4Rm.ParamDict['dw_out'] = temp_5
elif a.AllDataDict['model'] == 2:
x_dw_temp = []
x_dw_temp.append(t*(1-a.ParamDict['dwp'][1]))
x_dw_temp.append(t*(1-a.ParamDict['dwp'][1] +
a.ParamDict['dwp'][2]/2))
x_dw_temp.append(t*(1-a.ParamDict['dwp'][1] -
a.ParamDict['dwp'][3]/2))
x_dw_temp.append(t*0.05)
P4Rm.ParamDict['x_dwp'] = x_dw_temp
y_dw_temp = []
y_dw_temp.append(a.ParamDict['dwp'][0])
y_dw_temp.append(1. - (1-a.ParamDict['dwp'][0])/2)
y_dw_temp.append(1. - (1-a.ParamDict['dwp'][0])/2 -
(1-a.ParamDict['dwp'][6])/2)
y_dw_temp.append(a.ParamDict['dwp'][6])
P4Rm.ParamDict['DW_shifted'] = y_dw_temp
def solve(self, A=None, b=None, iterative_solver=None, **kwargs):
r"""
Executes the right algorithm for the solution: regular solution of a
linear system or iterative solution over the nonlinear source terms.
Parameters
----------
A : sparse matrix
2D Coefficient matrix
b : dense matrix
1D RHS vector
iterative_sovler : string
Name of solver to use. If not solve is specified, sp.solve is used
which is a direct solver (SuperLU on default Scipy installation)
kwargs : list of keyword arguments
These arguments and values are sent to the sparse solver, so read
the specific documentation for the solver chosen
"""
self._iterative_solver = iterative_solver
# Executes the right algorithm
if any("pore.source_nonlinear" in s for s in self.props()):
X = self._do_one_outer_iteration(**kwargs)
else:
X = self._do_one_inner_iteration(A, b, **kwargs)
self.X = X
self._Neumann_super_X = self.X[
-sp.in1d(sp.arange(0, self._coeff_dimension), self.pores())]
#Removing the additional super pore variables from the results
self[self._quantity] = self.X[self.pores()]
logger.info('Writing the results to ' + '[\'' + self._quantity +
'\'] in the ' + self.name + ' algorithm.')
def _check_trapping(self, inv_val):
r"""
Determine which pores and throats are trapped by invading phase. This
method is called by ``run`` if 'trapping' is set to True.
"""
# Generate a list containing boolean values for throat state
Tinvaded = self['throat.inv_Pc'] < sp.inf
# Add residual throats, if any, to list of invaded throats
Tinvaded = Tinvaded + self['throat.residual']
# Invert logic to find defending throats
Tdefended = ~Tinvaded
[pclusters, tclusters] = self._net.find_clusters2(mask=Tdefended,
t_labels=True)
# See which outlet pores remain uninvaded
outlets = self['pore.outlets'] * (self['pore.inv_Pc'] == sp.inf)
# Identify clusters connected to remaining outlet sites
def_clusters = sp.unique(pclusters[outlets])
temp = sp.in1d(sp.unique(pclusters), def_clusters, invert=True)
trapped_clusters = sp.unique(pclusters)[temp]
trapped_clusters = trapped_clusters[trapped_clusters >= 0]
# Find defending clusters NOT connected to the outlet pores
pmask = np.in1d(pclusters, trapped_clusters)
# Store current applied pressure in newly trapped pores
pinds = (self['pore.trapped'] == sp.inf) * (pmask)
self['pore.trapped'][pinds] = inv_val
# Find throats on the trapped defending clusters
tinds = self._net.find_neighbor_throats(pores=pinds,
mode='intersection')
self['throat.trapped'][tinds] = inv_val
self['throat.entry_pressure'][tinds] = 1000000
def make_introns_feasible(introns, genes, CFG):
# introns = make_introns_feasible(introns, genes, CFG)
tmp1 = sp.array([x.shape[0] for x in introns[:, 0]])
tmp2 = sp.array([x.shape[0] for x in introns[:, 1]])
unfeas = sp.where((tmp1 > 200) | (tmp2 > 200))[0]
print >> CFG['fd_log'], 'found %i unfeasible genes' % unfeas.shape[0]
while unfeas.shape[0] > 0:
### make filter more stringent
CFG['read_filter']['exon_len'] = min(36, CFG['read_filter']['exon_len'] + 4)
CFG['read_filter']['mincount'] = 2 * CFG['read_filter']['mincount']
CFG['read_filter']['mismatch'] = max(CFG['read_filter']['mismatch'] - 1, 0)
### get new intron counts
tmp_introns = get_intron_list(genes[unfeas], CFG)
introns[unfeas, :] = tmp_introns
### stil unfeasible?
tmp1 = sp.array([x.shape[0] for x in introns[:, 0]])
tmp2 = sp.array([x.shape[0] for x in introns[:, 1]])
still_unfeas = sp.where((tmp1 > 200) | (tmp2 > 200))[0]
idx = sp.where(~sp.in1d(unfeas, still_unfeas))[0]
for i in unfeas[idx]:
print >> CFG['fd_log'], '[feasibility] set criteria for gene %s to: min_ex %i, min_conf %i, max_mism %i' % (genes[i].name, CFG['read_filter']['exon_len'], CFG['read_filter']['mincount'], CFG['read_filter']['mismatch'])
unfeas = still_unfeas;
return introns
def get_usgs_n(self):
if self.get_usgsrc() == 0:
return
self.get_values(
) # Fetch usgsq,usgsh,handq,handh,handarea,handrad,handslope, handstage
# Find indices for integer stageheight values in usgsh, and apply to usgsq
usgsidx = scipy.where(scipy.equal(scipy.mod(
self.usgsh, 1), 0)) # Find indices of integer values in usgsh
usgsh = self.usgsh[usgsidx]
usgsq = self.usgsq[usgsidx]
# Find indices where usgsh[usgsidx] occur in handstage, and apply to handarea and handrad
handidx = scipy.where(scipy.in1d(self.handstage, usgsh))
area = self.handarea[handidx]
hydrad = self.handrad[handidx]
# Remove usgsq values for duplicate usgsh heights (keep first instance only)
if usgsh.shape != area.shape:
for i in range(usgsh.shape[0]):
if i == 0: pass
elif usgsh[i] == usgsh[i - 1]:
usgsq = scipy.delete(usgsq, i)
# Calculate average manning's n after converting discharge units
disch = usgsq #*0.0283168 # Convert cfs to cms
self.usgsroughness_array = self.mannings_n(area=area,
hydrad=hydrad,
slope=self.handslope,
disch=disch)
self.usgsroughness = scipy.average(self.usgsroughness_array)
print 'Average roughness: {0:.2f}'.format(self.usgsroughness)
def test_find_nearby_pores_distance_2_flattened_inclself(self):
a = self.net.find_nearby_pores(pores=[0, 1],
distance=2,
flatten=True,
excl_self=False)
assert sp.size(a) == 17
assert sp.all(sp.in1d([0, 1], a))
def intersect_rows(array1, array2, index=False):
"""Return array with rows that intersect between array1 and array2"""
tmp1 = sp.array(
['-'.join(array1[i, :].astype('str')) for i in range(array1.shape[0])])
tmp2 = sp.array(
['-'.join(array2[i, :].astype('str')) for i in range(array2.shape[0])])
idx = sp.where(sp.in1d(tmp1, tmp2))[0]
if index:
idx2 = sp.where(sp.in1d(tmp2, tmp1))[0]
if index:
return (array1[idx, :], idx, idx2)
else:
return (array1[idx, :], None, None)
def validate_vulnerability_set(self):
"""The vulnerability set must provide curves for all sites in this
object. A Vulnerability_Function needs to be defined to match each
attributes['STRUCTURE_CLASSIFICATION'] identifier.
Raises a RuntimeError if it cannot find a match.
"""
if self.vulnerability_set is None:
raise RuntimeError('Vulnerability Set must not be None')
# Function IDs for the vulnerability set
curves_defined = self.vulnerability_set.vulnerability_functions.keys()
# Sites STRUCTURE_CLASSIFICATIONs
structure_classifications = self.attributes['STRUCTURE_CLASSIFICATION']
structure_classifications = unique(structure_classifications)
# Are there any unique structure classifications that are not in the
# curves defined?
in_curves_defined = in1d(structure_classifications, curves_defined)
if not alltrue(in_curves_defined):
msg = 'The following structures do not have a vulnerability curve: '
msg += '%s' % structure_classifications[where(
in_curves_defined == False)]
raise RuntimeError(msg)
def getSizeFactor(fn_anno, data, gid, mode = 'sum', withXYMT = True, filterbyPC = True):
'''
input annotation, counts and gene ids
output sum of protein coding gene levels excluding sex chromosomes and mitochondria genes
'''
anno = sp.loadtxt(fn_anno, delimiter = '\t', dtype = 'string', usecols=[0,2,8])
anno = anno[anno[:,1] == 'gene', :]
if not withXYMT: ### filter xymt
anno = anno[anno[:,0] != 'MT',:]
anno = anno[anno[:,0] != 'Y',:]
anno = anno[anno[:,0] != 'X',:]
agid = [x.split(';')[0] for x in anno[:,2]] ### clean gene id's
agid = sp.array([x.split(" ")[1].strip('\"') for x in agid])
if filterbyPC: ### filter protein coding
gtpe = [x.split(';')[2] for x in anno[:,2]]
gtpe = sp.array([x.split('\"')[1].split('\"')[0] for x in gtpe])
iPC = sp.where(gtpe == 'protein_coding')[0]
agid = agid[iPC]
iGn = sp.in1d(gid, agid)
libsize = sp.sum(data[iGn,:], axis = 0)
if mode == 'uq':
libsize = sp.array([sp.percentile(x[x!=0] ,75) for x in data[iGn,:].T]) * iGn.sum()
return libsize
def get_sid_pos_map(sids, hdf5_kgenomes_file=None):
"""
Returns a SNP map, with information for each SNP
"""
if hdf5_kgenomes_file is None:
hdf5_kgenomes_file = __default_kgenomes_file__
h5f = h5py.File(hdf5_kgenomes_file, 'r')
snp_info_map = {}
for chrom_i in range(1, 23):
cg = h5f['chrom_%d' % chrom_i]
sids_1k = cg['sids'][...]
sids_filter_1k = sp.in1d(sids_1k, sp.array(sids))
common_sids = sids_1k[sids_filter_1k]
common_positions = cg['positions'][sids_filter_1k]
eur_mafs = cg['eur_mafs'][sids_filter_1k]
nts = cg['nts'][sids_filter_1k]
for sid, pos, eur_maf, nt in izip(common_sids, common_positions,
eur_mafs, nts):
snp_info_map[sid] = {
'pos': pos,
'chrom': chrom_i,
'eur_maf': eur_maf,
'nts': nt
}
return snp_info_map
def make_introns_feasible(introns, genes, CFG):
# introns = make_introns_feasible(introns, genes, CFG)
tmp1 = sp.array([x.shape[0] for x in introns[:, 0]])
tmp2 = sp.array([x.shape[0] for x in introns[:, 1]])
unfeas = sp.where((tmp1 > 200) | (tmp2 > 200))[0]
print >> CFG['fd_log'], 'found %i unfeasible genes' % unfeas.shape[0]
while unfeas.shape[0] > 0:
### make filter more stringent
CFG['read_filter']['exon_len'] = min(36, CFG['read_filter']['exon_len'] + 4)
CFG['read_filter']['mincount'] = 2 * CFG['read_filter']['mincount']
CFG['read_filter']['mismatch'] = max(CFG['read_filter']['mismatch'] - 1, 0)
### get new intron counts
tmp_introns = get_intron_list(genes[unfeas], CFG)
introns[unfeas, :] = tmp_introns
### stil unfeasible?
tmp1 = sp.array([x.shape[0] for x in introns[:, 0]])
tmp2 = sp.array([x.shape[0] for x in introns[:, 1]])
still_unfeas = sp.where((tmp1 > 200) | (tmp2 > 200))[0]
idx = sp.where(~sp.in1d(unfeas, still_unfeas))[0]
for i in unfeas[idx]:
print >> CFG['fd_log'], '[feasibility] set criteria for gene %s to: min_ex %i, min_conf %i, max_mism %i' % (genes[i].name, CFG['read_filter']['exon_len'], CFG['read_filter']['mincount'], CFG['read_filter']['mismatch'])
unfeas = still_unfeas;
return introns
def regenerate(self, prop_list='',mode=None):
r'''
This updates all properties using the selected methods
Parameters
----------
prop_list : string or list of strings
The names of the properties that should be updated, defaults to all
mode : string
Control how the regeneration occurs.
Examples
--------
>>> pn = OpenPNM.Network.TestNet()
>>> pind = pn.get_pore_indices()
>>> geom = OpenPNM.Geometry.Stick_and_Ball(network=pn, name='geo_test', locations=pind)
>>> geom.regenerate() # Regenerate all properties at once
>>> geom.regenerate('pore_seed') # only one property
>>> geom.regenerate(['pore_seed', 'pore_diameter']) # or several
'''
if prop_list == '':
prop_list = self._prop_list
elif type(prop_list) == str:
prop_list = [prop_list]
if mode == 'exclude':
a = sp.array(self._prop_list)
b = sp.array(prop_list)
c = a[sp.where(~sp.in1d(a,b))[0]]
prop_list = list(c)
for item in prop_list:
self._logger.debug('Refreshing: '+item)
getattr(self,item)()
def test_find_nearby_pores_distance_2_flattened_inclself(self):
a = self.net.find_nearby_pores(pores=[0, 1],
distance=2,
flatten=True,
excl_self=False)
assert sp.size(a) == 17
assert sp.all(sp.in1d([0, 1], a))
def _check_trapping(self, inv_val):
r"""
Determine which pores and throats are trapped by invading phase. This
method is called by ``run`` if 'trapping' is set to True.
"""
# Generate a list containing boolean values for throat state
Tinvaded = self['throat.inv_Pc'] < sp.inf
# Add residual throats, if any, to list of invaded throats
Tinvaded = Tinvaded + self['throat.residual']
# Invert logic to find defending throats
Tdefended = ~Tinvaded
[pclusters, tclusters] = self._net.find_clusters2(mask=Tdefended,
t_labels=True)
# See which outlet pores remain uninvaded
outlets = self['pore.outlets']*(self['pore.inv_Pc'] == sp.inf)
# Identify clusters connected to remaining outlet sites
def_clusters = sp.unique(pclusters[outlets])
temp = sp.in1d(sp.unique(pclusters), def_clusters, invert=True)
trapped_clusters = sp.unique(pclusters)[temp]
trapped_clusters = trapped_clusters[trapped_clusters >= 0]
# Find defending clusters NOT connected to the outlet pores
pmask = np.in1d(pclusters, trapped_clusters)
# Store current applied pressure in newly trapped pores
pinds = (self['pore.trapped'] == sp.inf) * (pmask)
self['pore.trapped'][pinds] = inv_val
# Find throats on the trapped defending clusters
tinds = self._net.find_neighbor_throats(pores=pinds,
mode='intersection')
self['throat.trapped'][tinds] = inv_val
self['throat.entry_pressure'][tinds] = 1000000
def calculate_ld(nt_map_file, kgenomes_file, output_folder, window_size):
"""
Calculate LD in windows for a reference genome dataset for a given set of SNPIds that are defined in the genotype_file
"""
log.info('Calculating LD')
# Load 1K genome
kg_h5f = h5py.File(kgenomes_file, 'r')
# load map file.
with open(nt_map_file, 'rb') as f:
snp_map_dict = pickle.load(f, encoding='latin1')
# Figure out overlap (all genotype SNPs should be in the 1K genomes data)..
for chrom in range(1, 23):
log.info('Working on Chromosome %s' % chrom)
chrom_str1 = 'chr%s' % chrom
kg_cg = kg_h5f[chrom_str1]
kg_sids = kg_cg['snp_ids'][...]
chrom_dict = snp_map_dict[chrom_str1]
g_sids = chrom_dict['sids']
kg_filter = sp.in1d(kg_sids, g_sids)
assert sp.sum(kg_filter) == len(g_sids), '..bug...'
assert sp.all(kg_sids[kg_filter] == g_sids), '...bug'
snps = kg_cg['snps'][...]
snps = snps.compress(kg_filter, axis=0)
snp_stds = kg_cg['snp_stds'][...]
snp_stds = snp_stds.compress(kg_filter, axis=0)
snp_means = kg_cg['snp_means'][...]
snp_means = snp_means.compress(kg_filter, axis=0)
norm_snps = sp.array((snps - snp_means) / snp_stds, dtype='single')
# Iterate over SNPs and calculate LD
num_snps, num_indivs = snps.shape
ld_mats = []
boundaries = []
for snp_i in range(num_snps):
start_i = max(0, snp_i - window_size / 2)
end_i = min(snp_i + (window_size / 2) + 1, num_snps)
X = norm_snps[start_i:end_i]
D = sp.dot(X, X.T) / num_indivs
ld_mats.append(D)
boundaries.append([start_i, end_i])
ld_dict = {'Ds':ld_mats, 'boundaries':boundaries, 'snp_means':snp_means, 'snp_stds':snp_stds, 'window_size':window_size}
# Store things
ld_file = '%s/LD' % output_folder + '_' + chrom_str1 + '.pickled.gz'
log.info('Saving LD in %s' % ld_file)
with gzip.open(ld_file, 'w') as f:
pickle.dump(ld_dict, f, protocol=2)
def test_find_nearby_pores_distance_2_flattened_include_input(self):
a = self.net.find_nearby_pores(pores=[0, 1],
r=2,
flatten=True,
include_input=True)
assert sp.size(a) == 17
assert sp.all(sp.in1d([0, 1], a))
def remove_isolated_clusters(conns, nonzero_locs, num_to_keep):
r"""
Identifies and removes all disconnected clusters except the number of
groups specified by "num_to_keep". num_to_keep=N retains the N largest
clusters
"""
#
adj_mat = generate_adjacency_matrix(conns, nonzero_locs)
#
logger.info('determining connected components...')
cs_ids = csgraph.connected_components(csgraph=adj_mat, directed=False)[1]
groups, counts = sp.unique(cs_ids, return_counts=True)
order = sp.argsort(counts)[::-1]
groups = groups[order]
counts = counts[order]
#
msg = ' {} component groups for {} total nodes'
logger.debug(msg.format(groups.size, cs_ids.size))
msg = ' largest group number: {}, size {}'
logger.debug(msg.format(groups[0], counts[0]))
msg = ' {} % of nodes contained in largest group'
logger.debug(msg.format(counts[0]/cs_ids.size*100))
msg = ' {} % of nodes contained in {} retained groups'
num = sp.sum(counts[0:num_to_keep])/cs_ids.size*100
logger.debug(msg.format(num, num_to_keep))
#
inds = sp.where(sp.in1d(cs_ids, groups[0:num_to_keep]))[0]
num = nonzero_locs.size
nonzero_locs = nonzero_locs[inds]
msg = ' removed {} disconnected nodes'
logger.debug(msg.format(num - nonzero_locs.size))
#
return nonzero_locs
def fill_masked_pixels(dll, ll, delta, diff, iv, no_apply_filling):
if no_apply_filling: return ll, delta, diff, iv, 0
ll_idx = ll.copy()
ll_idx -= ll[0]
ll_idx /= dll
ll_idx += 0.5
index = sp.array(ll_idx, dtype=int)
index_all = range(index[-1] + 1)
index_ok = sp.in1d(index_all, index)
delta_new = sp.zeros(len(index_all))
delta_new[index_ok] = delta
ll_new = sp.array(index_all, dtype=float)
ll_new *= dll
ll_new += ll[0]
diff_new = sp.zeros(len(index_all))
diff_new[index_ok] = diff
iv_new = sp.ones(len(index_all))
iv_new *= 0.0
iv_new[index_ok] = iv
nb_masked_pixel = len(index_all) - len(index)
return ll_new, delta_new, diff_new, iv_new, nb_masked_pixel
def evaluate_trapping(self, p_outlets):
r"""
Finds trapped pores and throats after a full ordinary
percolation simulation has been run.
Parameters
----------
p_outlets : array_like
A list of pores that define the wetting phase outlets.
Disconnection from these outlets results in trapping.
Returns
-------
It creates arrays called ``pore.trapped`` and ``throat.trapped``, but
also adjusts the ``pore.inv_Pc`` and ``throat.inv_Pc`` arrays to set
trapped locations to have infinite invasion pressure.
"""
self['pore.trapped'] = sp.zeros([
self.Np,
], dtype=float)
self['throat.trapped'] = sp.zeros([
self.Nt,
], dtype=float)
try:
# Get points used in OP
inv_points = sp.unique(self['pore.inv_Pc'])
except:
raise Exception('Orindary percolation has not been run!')
tind = self._net.throats()
conns = self._net.find_connected_pores(tind)
for inv_val in inv_points[0:-1]:
# Find clusters of defender pores
Pinvaded = self['pore.inv_Pc'] <= inv_val
Cstate = sp.sum(Pinvaded[conns], axis=1)
Tinvaded = self['throat.inv_Pc'] <= inv_val
# 0 = all open, 1=1 pore filled,
# 2=2 pores filled 3=2 pores + 1 throat filled
Cstate = Cstate + Tinvaded
clusters = self._net.find_clusters(Cstate == 0)
# Clean up clusters (invaded = -1, defended >=0)
clusters = clusters * (~Pinvaded) - (Pinvaded)
# Identify clusters connected to outlet sites
out_clusters = sp.unique(clusters[p_outlets])
trapped_pores = ~sp.in1d(clusters, out_clusters)
trapped_pores[Pinvaded] = False
if sum(trapped_pores) > 0:
inds = (self['pore.trapped'] == 0) * trapped_pores
self['pore.trapped'][inds] = inv_val
trapped_throats = self._net.find_neighbor_throats(
trapped_pores)
trapped_throat_array = np.asarray([False] * len(Cstate))
trapped_throat_array[trapped_throats] = True
inds = (self['throat.trapped'] == 0) * trapped_throat_array
self['throat.trapped'][inds] = inv_val
inds = (self['throat.trapped'] == 0) * (Cstate == 2)
self['throat.trapped'][inds] = inv_val
self['pore.inv_Pc'][self['pore.trapped'] > 0] = sp.inf
self['throat.inv_Pc'][self['throat.trapped'] > 0] = sp.inf
def whitelisting(options, header, data):
whitelist = sp.loadtxt(options.fn_white, delimiter = '\t', dtype = 'string')
midx_m = sp.in1d(header, whitelist)
tags = sp.array([x.split('-')[3] for x in header])
midx_n = np.core.defchararray.startswith(tags, '1')
header = header[midx_m | midx_n]
data = data[:, midx_m | midx_n]
return header, data
def test_add_boundary_pores(self):
net = op.Network.CubicDual(shape=[5, 5, 5], label_1='primary',
label_2='secondary')
Ps = net.pores(labels=['surface', 'bottom'], mode='intersection')
net.add_boundary_pores(pores=Ps, offset=[0, 0, -0.5])
Ps2 = net.pores(labels=['boundary'], mode='intersection')
assert Ps.size == Ps2.size
assert ~sp.any(sp.in1d(Ps, Ps2))
def test_add_boundary_pores_cubicdual(self):
net = OpenPNM.Network.CubicDual(shape=[5, 5, 5],
label_1='primary',
label_2='secondary')
Ps = net.pores(labels=['surface', 'bottom'], mode='intersection')
net.add_boundary_pores(pores=Ps, offset=[0, 0, -0.5])
Ps2 = net.pores(labels=['boundary'], mode='intersection')
assert Ps.size == Ps2.size
assert ~sp.any(sp.in1d(Ps, Ps2))
def rate(self,pores='',throats=''):
if throats!='':
p1 = self._net.find_connected_pores(throats)[:,0]
p2 = self._net.find_connected_pores(throats)[:,1]
elif pores!='':
throats = self._net.find_neighbor_throats(pores,flatten=True,mode='not_intersection')
p1 = self._net.find_connected_pores(throats)[:,0]
p2 = self._net.find_connected_pores(throats)[:,1]
pores1 = sp.copy(p1)
pores2 = sp.copy(p2)
pores1[-sp.in1d(p1,pores)] = p2[-sp.in1d(p1,pores)]
pores2[-sp.in1d(p1,pores)] = p1[-sp.in1d(p1,pores)]
X1 = self._result[pores1]
X2 = self._result[pores2]
g = self._conductance[throats]
R = sp.sum(sp.multiply(g,(X1-X2)))
return(R)
def _build_RHS_matrix(self,
modified_RHS_pores=None,
RHS_added_data=None,
mode='overwrite'):
r'''
This builds the right-hand-side matrix for the linear solver.
'''
if mode == 'overwrite':
A_dim = self._coeff_dimension
b = sp.zeros([A_dim, 1])
try:
Dir_pores = self.pores(self._phase.name + '_Dirichlet')
Dir_pores_vals = self['pore.' + self._phase.name +
'_bcval_Dirichlet'][Dir_pores]
b[Dir_pores] = sp.reshape(Dir_pores_vals, [len(Dir_pores), 1])
except:
pass
try:
individual_Neu_pores = self.pores(self._phase.name +
'_Neumann')
individual_Neu_pores_vals = self[
'pore.' + self._phase.name +
'_bcval_Neumann'][individual_Neu_pores]
b[individual_Neu_pores] = sp.reshape(
individual_Neu_pores_vals, [len(individual_Neu_pores), 1])
except:
pass
try:
self.pores(self._phase.name + '_Neumann_group')
pnum = self._net.num_pores()
b[sp.r_[pnum:(pnum +
len(self._group_Neumann_vals))]] = sp.reshape(
self._group_Neumann_vals[
sp.r_[0:len(self._group_Neumann_vals)]],
[len(self._group_Neumann_vals), 1])
except:
pass
if mode in ['overwrite', 'modify_RHS']:
try:
b = sp.copy(self.b)
except:
pass
# Adding necessary terms such as source terms to the RHS for non-Dirichlet pores
if modified_RHS_pores is not None and RHS_added_data is not None:
if sp.size(modified_RHS_pores) == sp.size(RHS_added_data):
p = sp.in1d(modified_RHS_pores, self._non_Dir_diag)
data = RHS_added_data[p]
b[modified_RHS_pores[p]] = b[
modified_RHS_pores[p]] + data.reshape([len(data), 1])
else:
raise Exception(
'Provided data and pores for modifying RHS matrix should have the same size!'
)
return (b)
def read_from_spplate(in_dir, thid, ra, dec, zqso, plate, mjd, fid, order, log=None, best_obs=False):
pix_data={}
unique_plates = sp.unique(plate)
print("reading {} plates".format(len(unique_plates)))
for p in unique_plates:
wplate = plate==p
plate_mjd = "{}-*".format(p)
mjd_in_plate = sp.unique(mjd[wplate])
spplates = glob.glob(in_dir+"/{}/spPlate-{}.fits".format(p, plate_mjd))
mjds_found = sp.array([spfile.split("-")[-1].replace(".fits",'') for spfile in spplates]).astype(int)
wmissing = ~sp.in1d(mjd_in_plate, mjds_found)
if wmissing.sum()>0:
for m in mjd_in_plate[wmissing]:
print("INFO: can't find spplate {} {}".format(p,m))
if log is not None:
log.write("INFO: can't find spplate {} {}\n".format(p,m))
for spplate in spplates:
h = fitsio.FITS(spplate)
head0 = h[0].read_header()
MJD = head0["MJD"]
t0 = time.time()
wfib = wplate
if best_obs:
## select only the objects which have specified mjd within this plate
wmjd = mjd == MJD
wfib = wplate & wmjd
coeff0 = head0["COEFF0"]
coeff1 = head0["COEFF1"]
flux = h[0].read()
ivar = h[1].read()*(h[2].read()==0)
llam = coeff0 + coeff1*sp.arange(flux.shape[1])
## now convert all those fluxes into forest objects
for (t, r, d, z, p, m, f) in zip(thid[wfib], ra[wfib], dec[wfib], zqso[wfib], plate[wfib], mjd[wfib], fid[wfib]):
index = f-1
d = forest(llam,flux[index],ivar[index], t, r, d, z, p, m, f, order)
if t in pix_data:
pix_data[t] += d
else:
pix_data[t] = d
if log is not None:
log.write("{} read from file {} and mjd {}\n".format(t, spplate, m))
print("INFO: read {} from {} in {} per spec. Progress: {} of {} \n".format(wfib.sum(), os.path.basename(spplate), (time.time()-t0)/(wfib.sum()+1e-3), len(pix_data), len(thid)))
h.close()
data = list(pix_data.values())
return data
def evaluate_trapping(self, p_outlets):
r"""
Finds trapped pores and throats after a full ordinary
percolation simulation has been run.
Parameters
----------
p_outlets : array_like
A list of pores that define the wetting phase outlets.
Disconnection from these outlets results in trapping.
Returns
-------
It creates arrays called ``pore.trapped`` and ``throat.trapped``, but
also adjusts the ``pore.inv_Pc`` and ``throat.inv_Pc`` arrays to set
trapped locations to have infinite invasion pressure.
"""
self['pore.trapped'] = sp.zeros([self.Np, ], dtype=float)
self['throat.trapped'] = sp.zeros([self.Nt, ], dtype=float)
try:
# Get points used in OP
inv_points = sp.unique(self['pore.inv_Pc'])
except:
raise Exception('Orindary percolation has not been run!')
tind = self._net.throats()
conns = self._net.find_connected_pores(tind)
for inv_val in inv_points[0:-1]:
# Find clusters of defender pores
Pinvaded = self['pore.inv_Pc'] <= inv_val
Cstate = sp.sum(Pinvaded[conns], axis=1)
Tinvaded = self['throat.inv_Pc'] <= inv_val
# 0 = all open, 1=1 pore filled,
# 2=2 pores filled 3=2 pores + 1 throat filled
Cstate = Cstate + Tinvaded
clusters = self._net.find_clusters(Cstate == 0)
# Clean up clusters (invaded = -1, defended >=0)
clusters = clusters * (~Pinvaded) - (Pinvaded)
# Identify clusters connected to outlet sites
out_clusters = sp.unique(clusters[p_outlets])
trapped_pores = ~sp.in1d(clusters, out_clusters)
trapped_pores[Pinvaded] = False
if sum(trapped_pores) > 0:
inds = (self['pore.trapped'] == 0) * trapped_pores
self['pore.trapped'][inds] = inv_val
trapped_throats = self._net.find_neighbor_throats(trapped_pores)
trapped_throat_array = np.asarray([False] * len(Cstate))
trapped_throat_array[trapped_throats] = True
inds = (self['throat.trapped'] == 0) * trapped_throat_array
self['throat.trapped'][inds] = inv_val
inds = (self['throat.trapped'] == 0) * (Cstate == 2)
self['throat.trapped'][inds] = inv_val
self['pore.trapped'][self['pore.trapped'] > 0] = sp.inf
self['throat.trapped'][self['throat.trapped'] > 0] = sp.inf
self['pore.inv_Pc'][self['pore.trapped'] > 0] = sp.inf
self['throat.inv_Pc'][self['throat.trapped'] > 0] = sp.inf
def test_map_pores(self):
a = self.geo21['pore._id']
b = self.geo22['pore._id']
assert a.size == self.geo21.Np
assert b.size == self.geo22.Np
assert ~sp.any(sp.in1d(a, b))
Pgeo21 = self.net2.map_pores(pores=self.geo21.Ps, origin=self.geo21)
assert sp.all(Pgeo21 == self.net2.pores(self.geo21.name))
Pgeo22 = self.net2.map_pores(pores=self.geo22.Ps, origin=self.geo22)
assert sp.all(Pgeo22 == self.net2.pores(self.geo22.name))
def test_map_pores(self):
a = self.geo21['pore._id']
b = self.geo22['pore._id']
assert a.size == self.geo21.Np
assert b.size == self.geo22.Np
assert ~sp.any(sp.in1d(a, b))
Pgeo21 = self.net2.map_pores(pores=self.geo21.Ps, origin=self.geo21)
assert sp.all(Pgeo21 == self.net2.pores(self.geo21.name))
Pgeo22 = self.net2.map_pores(pores=self.geo22.Ps, origin=self.geo22)
assert sp.all(Pgeo22 == self.net2.pores(self.geo22.name))
def find_interface_throats(self,labels=[]):
r'''
Finds the throats that join two pore labels.
Parameters
----------
labels : list of strings
The labels of the two pore groups whose interface is sought
Returns
-------
An array of throat numbers that connect the given pore groups
Notes
-----
This method is meant to find interfaces between TWO groups, regions or
clusters of pores (as defined by their label). If the input labels
overlap or are not adjacent, an empty array is returned.
Examples
--------
>>> pn = OpenPNM.Network.TestNet()
>>> pn.set_pore_info(label='domain1',locations=[0,1,2])
>>> pn.set_pore_info(label='domain2',locations=[5,6,7])
>>> pn.find_interface_throats(labels=['domain1','domain2'])
array([1, 4, 7])
'''
Tind = sp.array([],ndmin=1)
if sp.shape(labels)[0] != 2:
self._logger.error('Exactly two labels must be given')
else:
P1 = self.get_pore_indices(labels=labels[0])
P2 = self.get_pore_indices(labels=labels[1])
#Check if labels overlap
if sp.sum(sp.in1d(P1,P2)) > 0:
self._logger.error('Some labels overlap, iterface cannot be found')
else:
T1 = self.find_neighbor_throats(P1)
T2 = self.find_neighbor_throats(P2)
Tmask = sp.in1d(T1,T2)
Tind = T1[Tmask]
return Tind
def on_shifted_sp_curves(self, t):
a = P4Rm()
if a.AllDataDict["model"] == 0:
temp_1 = arange(2, len(a.ParamDict["sp"]) + 1)
temp_2 = temp_1 * t / (len(a.ParamDict["sp"]))
P4Rm.ParamDict["x_sp"] = t - temp_2
shifted_sp = a.ParamDict["sp"][:-1:]
temp_3 = in1d(around(a.ParamDict["depth"], decimals=3), around(a.ParamDict["x_sp"], decimals=3))
temp_4 = a.ParamDict["strain_i"][temp_3]
P4Rm.ParamDict["scale_strain"] = shifted_sp / temp_4
P4Rm.ParamDict["scale_strain"][a.ParamDict["scale_strain"] == 0] = 1.0
P4Rm.ParamDict["strain_shifted"] = asarray(shifted_sp) * 100.0 / a.ParamDict["scale_strain"]
P4Rm.ParamDict["stain_out"] = a.ParamDict["sp"][-1]
elif a.AllDataDict["model"] == 1:
temp_1 = arange(0, len(a.ParamDict["sp"]) + 1 - 3)
temp_2 = temp_1 * t / (len(a.ParamDict["sp"]) - 3)
P4Rm.ParamDict["x_sp"] = t - temp_2
shifted_sp = a.ParamDict["sp"][1:-1:]
temp_3 = in1d(around(a.ParamDict["depth"], decimals=3), around(a.ParamDict["x_sp"], decimals=3))
temp_4 = a.ParamDict["strain_i"][temp_3]
P4Rm.ParamDict["scale_strain"] = shifted_sp / temp_4
P4Rm.ParamDict["scale_strain"][a.ParamDict["scale_strain"] == 0] = 1.0
P4Rm.ParamDict["strain_shifted"] = asarray(shifted_sp) * 100.0 / a.ParamDict["scale_strain"]
temp_5 = array([a.ParamDict["sp"][0], a.ParamDict["sp"][-1]])
P4Rm.ParamDict["stain_out"] = temp_5
elif a.AllDataDict["model"] == 2:
x_sp_temp = []
x_sp_temp.append(t * (1 - a.ParamDict["sp"][1]))
x_sp_temp.append(t * (1 - a.ParamDict["sp"][1] + a.ParamDict["sp"][2] / 2))
x_sp_temp.append(t * (1 - a.ParamDict["sp"][1] - a.ParamDict["sp"][3] / 2))
x_sp_temp.append(t * 0.05)
P4Rm.ParamDict["x_sp"] = x_sp_temp
y_sp_temp = []
y_sp_temp.append(a.ParamDict["sp"][0])
y_sp_temp.append(a.ParamDict["sp"][0] / 2)
y_sp_temp.append(a.ParamDict["sp"][0] / 2 + a.ParamDict["sp"][6] / 2)
y_sp_temp.append(a.ParamDict["sp"][6])
P4Rm.ParamDict["strain_shifted"] = y_sp_temp
def make_unique_by_strain(event_list):
# event_list = make_unique_by_strain(event_list)
rm_idx = []
for i in range(1, event_list.shape[0]):
if i % 1000 == 0:
print '.',
if i % 10000 == 0:
print '%i' % i
old_coords = event_list[i - 1].get_coords(trafo=True)
curr_coords = event_list[i].get_coords(trafo=True)
if old_coords.shape[0] == curr_coords.shape[0] and sp.all(
old_coords == curr_coords):
### assertion that we did everything right
if event_list[i - 1].chr == event_list[i].chr:
assert (event_list[i - 1].strand == event_list[i].strand)
assert (event_list[i].strain.shape[0] == 1)
else:
assert (event_list[i - 1].gene_name != event_list[i].gene_name)
idx = sp.where(event_list[i -
1].strain == event_list[i].strain[0])[0]
if idx.shape[0] > 0:
assert (idx.shape[0] == 1)
assert (sp.all(event_list[i].get_coords(
trafo=True) == event_list[i - 1].get_coords(trafo=True)))
if not event_list[i].gene_name[0] in event_list[i -
1].gene_name:
event_list[
i - 1].gene_name = sp.r_[event_list[i - 1].gene_name,
[event_list[i].gene_name[0]]]
event_list[i] = event_list[i - 1]
else:
event_list[i].strain = sp.r_[[event_list[i - 1].strain[0]],
event_list[i].strain]
assert (sp.all(
sp.sort(event_list[i].strain) == sp.sort(
sp.unique(event_list[i].strain))))
### TODO !!!!!!!!!!!!! make sure that we keep different coordinates if the strains differ ...
if not event_list[i].gene_name[0] in event_list[i -
1].gene_name:
event_list[i].gene_name = sp.r_[
event_list[i - 1].gene_name,
[event_list[i].gene_name[0]]]
rm_idx.append(i - 1)
print 'events dropped: %i' % len(rm_idx)
keep_idx = sp.where(~sp.in1d(sp.arange(event_list.shape[0]), rm_idx))[0]
event_list = event_list[keep_idx]
return event_list
def timereduce(self, timelims=None,timesselected=None):
assert (timelims is not None) or (timesselected is not None), "Need a set of limits or selected set of times"
if timelims is not None:
tkeep = sp.logical_and(self.Time_Vector>=timelims[0],self.Time_Vector<timelims[1])
if timesselected is not None:
tkeep = sp.in1d(self.Time_Vector,timesselected)
# prune the arrays
self.Time_Vector=self.Time_Vector[tkeep]
self.Param_List=self.Param_List[:,tkeep]
self.Velocity=self.Velocity[:,tkeep]
def split_data(RV):
sp.random.seed(0)
n_train = int(4 * RV["Y"].shape[0] / 5.0)
n_test = int(1 * RV["Y"].shape[0] / 10.0)
idxs = sp.random.permutation(RV["Y"].shape[0])
idxs_train = idxs[:n_train]
idxs_test = idxs[n_train:(n_train + n_test)]
idxs_val = idxs[(n_train + n_test):]
Itrain = sp.in1d(sp.arange(RV["Y"].shape[0]), idxs_train)
Itest = sp.in1d(sp.arange(RV["Y"].shape[0]), idxs_test)
Ival = sp.in1d(sp.arange(RV["Y"].shape[0]), idxs_val)
out = {}
for key in RV.keys():
out["%s_train" % key] = RV[key][Itrain]
out["%s_val" % key] = RV[key][Ival]
out["%s_test" % key] = RV[key][Itest]
return out
def test_from_neighbor_throats_max(self):
self.geo.pop('pore.seed', None)
self.geo.models.pop('pore.seed', None)
self.geo.models.pop('throat.seed', None)
self.geo['throat.seed'] = sp.rand(self.net.Nt, )
self.geo.add_model(model=mods.from_neighbor_throats,
propname='pore.seed',
throat_prop='throat.seed',
mode='max')
assert sp.all(sp.in1d(self.geo['pore.seed'], self.geo['throat.seed']))
pmin = sp.amin(self.geo['pore.seed'])
tmin = sp.amin(self.geo['throat.seed'])
assert pmin >= tmin
def argintersect_left(a, b):
"""
find indices in a, whose corresponding values are in b
----------------------------------------------------------------------
Input:
a : array, for which indices are returned that are in the intersect with b
b : array to be intersected with a
----------------------------------------------------------------------
Output:
the indices of elements of a, which are in intersect of a and b
----------------------------------------------------------------------
"""
return sp.arange(a.shape[0])[sp.in1d(a,b)]
def argintersect_left(a, b):
"""
find indices in a, whose corresponding values are in b
----------------------------------------------------------------------
Input:
a : array, for which indices are returned that are in the intersect with b
b : array to be intersected with a
----------------------------------------------------------------------
Output:
the indices of elements of a, which are in intersect of a and b
----------------------------------------------------------------------
"""
return sp.arange(a.shape[0])[sp.in1d(a, b)]
def rate(self, pores='', mode='group'):
r'''
Send a list of pores and receive the net rate
of material moving into them.
Parameters
----------
pores : array_like
The pores where the net rate will be calculated
mode : string, optional
Controls how to return the rate. Options are:
- 'group'(default): It returns the cumulative rate moving into them
- 'single': It calculates the rate for each pore individually.
'''
pores = sp.array(pores, ndmin=1)
R = []
if mode == 'group': iteration = 1
elif mode == 'single': iteration = sp.shape(pores)[0]
for i in sp.r_[0:iteration]:
if mode == 'group': P = pores
elif mode == 'single': P = pores[i]
throats = self._net.find_neighbor_throats(P,
flatten=True,
mode='not_intersection')
p1 = self._net.find_connected_pores(throats)[:, 0]
p2 = self._net.find_connected_pores(throats)[:, 1]
pores1 = sp.copy(p1)
pores2 = sp.copy(p2)
#Changes to pores1 and pores2 to make them as the internal and external pores
pores1[-sp.in1d(p1, P)] = p2[-sp.in1d(p1, P)]
pores2[-sp.in1d(p1, P)] = p1[-sp.in1d(p1, P)]
X1 = self[self._quantity][pores1]
X2 = self[self._quantity][pores2]
g = self['throat.conductance'][throats]
R.append(sp.sum(sp.multiply(g, (X2 - X1))))
return (sp.array(R, ndmin=1))
def get_conf_events(options, gid):
event_info = []
for event_type in options.event_types:
IN = h5py.File(os.path.join(options.outdir, 'merge_graphs_%s_C%i.counts.hdf5' % (event_type, options.confidence)), 'r')
if 'conf_idx' in IN and IN['conf_idx'].shape[0] > 0 and IN['conf_idx'][0] != -1:
conf_idx = IN['conf_idx'][:].astype('int') - 1
k_idx = sp.where(sp.in1d(IN['gene_idx'][:][conf_idx].astype('int') - 1, gid))[0]
if k_idx.shape[0] > 0:
event_info.extend([[event_type, x] for x in conf_idx[k_idx]])
IN.close()
return sp.array(event_info, dtype='str')
def test_neighbor_max(self):
catch = self.geo.pop('pore.seed', None)
catch = self.geo.models.pop('pore.seed', None)
catch = self.geo.models.pop('throat.seed', None)
mod = gm.pore_misc.neighbor
self.geo['throat.seed'] = sp.rand(self.net.Nt,)
self.geo.models.add(model=mod,
propname='pore.seed',
throat_prop='throat.seed',
mode='max')
assert sp.all(sp.in1d(self.geo['pore.seed'], self.geo['throat.seed']))
pmin = sp.amin(self.geo['pore.seed'])
tmin = sp.amin(self.geo['throat.seed'])
assert pmin >= tmin
def test_conduit_conductance_loose(self):
self.phase['pore.occupancy'][[19, 20]] = 0
t1 = self.net.Ts[self.phase['throat.occupancy'] == 0]
t2 = self.net.Ts[~sp.in1d(self.net.Ts, t1)]
self.phys1.models.add(propname='throat.cond_conductance',
throat_conductance='throat.diffusive_conductance',
model=pm.multiphase.conduit_conductance,
mode='loose', factor=0)
self.phys2.models.add(propname='throat.cond_conductance',
throat_conductance='throat.diffusive_conductance',
model=pm.multiphase.conduit_conductance,
mode='loose', factor=0)
assert sp.all(self.phase['throat.cond_conductance'][t1] == 0)
assert sp.all(self.phase['throat.cond_conductance'][t2] != 0)
def _handle_multi_entries(header, data):
cols_of_interest = [[
'tumor_wgs_submitter_specimen_id', 'tumor_wgs_icgc_specimen_id',
'tumor_wgs_submitter_sample_id', 'tumor_wgs_icgc_sample_id',
'tumor_wgs_aliquot_id', 'tumor_wgs_oxog_score', 'tumor_wgs_ContEST',
'tumor_wgs_Stars', 'tumor_wgs_bwa_alignment_gnos_repo',
'tumor_wgs_bwa_alignment_gnos_id',
'is_mar2016_tumor_wgs_bwa_alignment',
'tumor_wgs_bwa_alignment_bam_file_name', 'tumor_wgs_minibam_gnos_repo',
'tumor_wgs_minibam_gnos_id', 'is_mar2016_tumor_wgs_minibam',
'tumor_wgs_minibam_bam_file_name',
'sanger_variant_calling_file_name_prefix',
'dkfz_embl_variant_calling_file_name_prefix',
'broad_variant_calling_file_name_prefix',
'muse_variant_calling_file_name_prefix',
'broad_tar_variant_calling_file_name_prefix',
'tumor_wgs_has_matched_rna_seq'
],
[
'tumor_rna_seq_submitter_specimen_id',
'tumor_rna_seq_icgc_specimen_id',
'tumor_rna_seq_submitter_sample_id',
'tumor_rna_seq_icgc_sample_id',
'tumor_rna_seq_aliquot_id',
'tumor_rna_seq_star_alignment_gnos_repo',
'tumor_rna_seq_star_alignment_gnos_id',
'is_mar2016_tumor_rna_seq_star_alignment',
'tumor_rna_seq_star_alignment_bam_file_name',
'tumor_rna_seq_tophat_alignment_gnos_repo',
'tumor_rna_seq_tophat_alignment_gnos_id',
'is_mar2016_tumor_rna_seq_tophat_alignment',
'tumor_rna_seq_tophat_alignment_bam_file_name'
]]
for cols in cols_of_interest:
c_idx = sp.where(sp.in1d(header, cols))[0]
r_idx = sp.where([',' in x for x in data[:, c_idx[0]]])[0]
for r in r_idx:
data_ = sp.array([x.split(',') for x in data[r, c_idx]])
assert len(data_.shape) > 1
assert data_.shape[1] > 1
for r2 in range(1, data_.shape[1]):
data = sp.r_[data, data[r, :][sp.newaxis, :]]
data[-1, c_idx] = data_[:, r2]
data[r, c_idx] = data_[:, 0]
return data
def _generate_throats(self):
r"""
Generate the throats (connections, numbering and types)
"""
self._logger.info("generate_throats: Define connections between pores")
[Nx, Ny, Nz] = sp.shape(self._template)
Np = Nx*Ny*Nz
ind = np.arange(0, Np)
#Generate throats based on pattern of the adjacency matrix
#This is taken from Cubic
tpore1_1 = ind[(ind % Nx) < (Nx-1)]
tpore2_1 = tpore1_1 + 1
tpore1_2 = ind[(ind % (Nx*Ny)) < (Nx*(Ny-1))]
tpore2_2 = tpore1_2 + Nx
tpore1_3 = ind[(ind % Np) < (Nx*Ny*(Nz-1))]
tpore2_3 = tpore1_3 + Nx*Ny
tpore1 = sp.hstack((tpore1_1, tpore1_2, tpore1_3))
tpore2 = sp.hstack((tpore2_1, tpore2_2, tpore2_3))
connections = sp.vstack((tpore1, tpore2)).T
connections = connections[sp.lexsort((connections[:, 1], connections[:, 0]))]
#Remove throats to non-active pores
img_ind = self.get_pore_data(prop='voxel_index')
temp0 = sp.in1d(connections[:, 0], img_ind)
temp1 = sp.in1d(connections[:, 1], img_ind)
tind = temp0*temp1
connections = connections[tind]
#Need a cleaner way to do this other than voxel_to_pore map...figure out later
self.set_throat_data(prop='connections', data=self._voxel_to_pore_map[connections])
self.set_throat_info(label='all', locations=sp.ones(sp.sum(tind,),dtype=bool))
self.set_throat_data(prop='numbering', data=np.arange(0, sp.sum(tind)))
self._logger.debug("generate_throats: End of method")
def intersect_rows(array1, array2, index = None):
"""Return intersection of rows"""
if (array1.shape[0] == 0):
if index == True:
return (array1, sp.zeros((0,)), sp.zeros((0,)))
else:
return array1
if (array2.shape[0] == 0):
if index == True:
return (array2, sp.zeros((0,)), sp.zeros((0,)))
else:
return array2
array1_v = array1.view([('', array1.dtype)] * array1.shape[1])
array2_v = array2.view([('', array2.dtype)] * array2.shape[1])
array_i = sp.intersect1d(array1_v, array2_v)
if index == True:
a1_i = sp.where(sp.in1d(array1_v, array_i))[0]
a2_i = sp.where(sp.in1d(array2_v, array_i))[0]
return (array_i.view(array1.dtype).reshape(array_i.shape[0], array1.shape[1]), a1_i, a2_i)
else:
return array_i.view(array1.dtype).reshape(array_i.shape[0], array1.shape[1])
def _do_one_inner_iteration(self):
if (self._BCtypes==0).all():
raise Exception('No boundary condition has been applied to this network.')
self._result = sp.zeros(self._net.num_pores())
else:
self._logger.info("Creating Coefficient matrix for the algorithm")
A = self._build_coefficient_matrix()
self._logger.info("Creating RHS matrix for the algorithm")
B = self._build_RHS_matrix()
self._logger.info("Solving AX = B for the sparse matrices")
X = sprslin.spsolve(A,B)
self._Neumann_super_X = X[-sp.in1d(sp.r_[0:len(X)],sp.r_[0:self._net.num_pores()])]
self._result = X[sp.r_[0:self._net.num_pores()]]
return(self._result)
def curate_alt_prime(event_list, CFG):
# event_list = curate_alt_prime(event_list)
if event_list.shape[0] == 0:
return event_list
rm_idx = []
corr_count = 0
for i in range(event_list.shape[0]):
### check if we have introns of zero length
#if sp.any(event_list[i].exons1[:, 1] - event_list[i].exons1[:, 1] < 2) or sp.any(event_list[i].exons2[:, 1] - event_list[i].exons2[:, 1] < 2):
if (event_list[i].exons1[1, 0] - event_list[i].exons1[0, 1] < 1) or (
event_list[i].exons2[1, 0] - event_list[i].exons2[0, 1] < 1):
rm_idx.append(i)
continue
### check if alt exons overlap, otherwise we cannot curate (trim to shortest length)
if (sp.all(event_list[i].exons1[0, :] == event_list[i].exons2[0, :]) and (event_list[i].exons1[1, 1] <= event_list[i].exons2[1, 0] or event_list[i].exons1[1, 0] >= event_list[i].exons2[1, 1])) or \
(sp.all(event_list[i].exons1[1, :] == event_list[i].exons2[1, :]) and (event_list[i].exons1[0, 1] <= event_list[i].exons2[0, 0] or event_list[i].exons1[0, 0] >= event_list[i].exons2[0, 1])):
continue
if sp.all(event_list[i].exons1[0, :] == event_list[i].exons2[0, :]):
if event_list[i].exons1[1, 1] > event_list[i].exons2[1, 1]:
event_list[i].exons1[1, 1] = event_list[i].exons2[1, 1]
corr_count += 1
elif event_list[i].exons1[1, 1] < event_list[i].exons2[1, 1]:
event_list[i].exons2[1, 1] = event_list[i].exons1[1, 1]
corr_count += 1
elif sp.all(event_list[i].exons1[1, :] == event_list[i].exons2[1, :]):
if event_list[i].exons1[0, 0] > event_list[i].exons2[0, 0]:
event_list[i].exons2[0, 0] = event_list[i].exons1[0, 0]
corr_count += 1
elif event_list[i].exons1[0, 0] < event_list[i].exons2[0, 0]:
event_list[i].exons1[0, 0] = event_list[i].exons2[0, 0]
corr_count += 1
### remove events with non-overlapping alt_exons
if len(rm_idx) > 0:
keep_idx = sp.where(
~sp.in1d(sp.arange(event_list.shape[0]), rm_idx))[0]
event_list = event_list[keep_idx]
print 'Corrected %i events' % corr_count
print 'Removed %i events' % len(rm_idx)
return event_list
