def __init__(self,
file_name=None,
mode='a',
data_fields=None,
default_field=None,
default_value=0.0,
regions=None,
edges=None,
_table_name_regions='region_data',
_table_name_edges='edges',
tmpdir=None):
RegionMatrixTable.__init__(self,
file_name=file_name,
additional_edge_fields=data_fields,
mode=mode,
tmpdir=tmpdir,
default_score_field=default_field,
default_value=default_value,
_table_name_regions=_table_name_regions,
_table_name_edges=_table_name_edges)
ArchitecturalFeature.__init__(self)
for edge_table in self._edge_table_iter():
if len(edge_table) > 0:
self._calculated = True
break
if regions is not None:
self.add_regions(regions)
# process data
if edges is not None:
self.add_edges(edges)
def as_data_frame(self, key, weight_column=None):
"""
See :class:`~RegionMatrixTable`
"""
return RegionMatrixTable.as_data_frame(self,
key,
weight_column=weight_column)
def _get_matrix(self,
row_ranges=None,
col_ranges=None,
weight_column=None,
default_value=0.0):
return RegionMatrixTable._get_matrix(self,
row_ranges=row_ranges,
col_ranges=col_ranges,
weight_column=weight_column)
def as_matrix(self,
key=slice(0, None, None),
values_from=None,
mask_missing=False,
impute_missing=False,
default_value=0.0):
"""
See :class:`~RegionMatrixTable`
"""
return RegionMatrixTable.as_matrix(self,
key=key,
values_from=values_from,
mask_missing=mask_missing,
impute_missing=impute_missing)
def setup_method(self, method):
self.rmt = RegionMatrixTable(
additional_edge_fields={
'weight': tables.Int32Col(pos=0),
'foo': tables.Int32Col(pos=1),
'bar': tables.Float32Col(pos=2),
'baz': tables.StringCol(50, pos=3)
})
for i in range(10):
if i < 5:
chromosome = 'chr1'
start = i * 1000
end = (i + 1) * 1000
elif i < 8:
chromosome = 'chr2'
start = (i - 5) * 1000
end = (i + 1 - 5) * 1000
else:
chromosome = 'chr3'
start = (i - 8) * 1000
end = (i + 1 - 8) * 1000
node = GenomicRegion(chromosome=chromosome, start=start, end=end)
self.rmt.add_region(node)
self.rmt.flush()
for i in range(10):
for j in range(i, 10):
edge = Edge(source=i,
sink=j,
weight=i * j,
foo=i,
bar=j,
baz='x' + str(i * j))
self.rmt.add_edge(edge)
self.rmt.flush()
class TestRegionMatrixTable:
def setup_method(self, method):
self.rmt = RegionMatrixTable(
additional_edge_fields={
'weight': tables.Int32Col(pos=0),
'foo': tables.Int32Col(pos=1),
'bar': tables.Float32Col(pos=2),
'baz': tables.StringCol(50, pos=3)
})
for i in range(10):
if i < 5:
chromosome = 'chr1'
start = i * 1000
end = (i + 1) * 1000
elif i < 8:
chromosome = 'chr2'
start = (i - 5) * 1000
end = (i + 1 - 5) * 1000
else:
chromosome = 'chr3'
start = (i - 8) * 1000
end = (i + 1 - 8) * 1000
node = GenomicRegion(chromosome=chromosome, start=start, end=end)
self.rmt.add_region(node)
self.rmt.flush()
for i in range(10):
for j in range(i, 10):
edge = Edge(source=i,
sink=j,
weight=i * j,
foo=i,
bar=j,
baz='x' + str(i * j))
self.rmt.add_edge(edge)
self.rmt.flush()
def teardown_method(self, method):
self.rmt.close()
def test_matrix(self):
m = self.rmt.matrix()
for row_region in m.row_regions:
i = row_region.ix
for col_region in m.col_regions:
j = col_region.ix
if np.ma.is_masked(m[i, j]):
assert np.ma.is_masked(m[j, i])
else:
assert m[i, j] == m[j, i] == i * j
m = self.rmt.matrix(score_field='foo')
for row_region in m.row_regions:
i = row_region.ix
for col_region in m.col_regions:
j = col_region.ix
if np.ma.is_masked(m[i, j]):
assert np.ma.is_masked(m[j, i])
else:
assert m[i, j] == m[j, i] == min(i, j)
m = self.rmt.matrix(score_field='bar')
for row_region in m.row_regions:
i = row_region.ix
for col_region in m.col_regions:
j = col_region.ix
if np.ma.is_masked(m[i, j]):
assert np.ma.is_masked(m[j, i])
else:
assert m[i, j] == m[j, i] == max(i, j)
def test_matrix_subset(self):
m = self.rmt.matrix(key=('chr2', 'chr2'), score_field='bar')
for i, row_region in enumerate(m.row_regions):
for j, col_region in enumerate(m.col_regions):
assert (np.ma.is_masked(m[i, j]) and np.ma.is_masked(m[j, i])) or \
(m[i, j] == m[j, i] == max(row_region.ix, col_region.ix))
m = self.rmt.matrix(key=('chr2', 'chr3'),
score_field='bar',
mask=False)
for i, row_region in enumerate(m.row_regions):
for j, col_region in enumerate(m.col_regions):
assert m[i, j] == max(row_region.ix, col_region.ix)
m = self.rmt.matrix(key=('chr3', 'chr2'),
score_field='bar',
mask=False)
for i, row_region in enumerate(m.row_regions):
for j, col_region in enumerate(m.col_regions):
assert m[i, j] == max(row_region.ix, col_region.ix)
def _get_nodes_from_key(self, key, as_index=False):
return RegionMatrixTable._get_nodes_from_key(self,
key,
as_index=as_index)
def close(self, **kwargs):
RegionMatrixTable.close(self, **kwargs)
def _edges_iter(self):
return RegionMatrixTable._edges_iter(self)
def edges_sorted(self, sortby, *args, **kwargs):
"""
See :class:`~RegionMatrixTable`
"""
return RegionMatrixTable.edges_sorted(self, sortby, *args, **kwargs)
def _nodes_iter(self):
return RegionMatrixTable._nodes_iter(self)
def get_edge(self, ix, lazy=False, **kwargs):
"""
See :class:`~RegionMatrixTable`
"""
return RegionMatrixTable.get_edge(self, ix, lazy=lazy)
def get_node(self, key):
"""
See :class:`~RegionMatrixTable`
"""
return RegionMatrixTable.get_node(self, key)
def _getitem_nodes(self, key, as_index=False):
return RegionMatrixTable._getitem_nodes(self, key, as_index=as_index)