def test_export_rectangles(self):
nd = np.arange(0, 80, dtype=float).reshape(8, 10)
rects1 = [[0, 1, 0, 1], [4, 5, 7, 8]]
rects2 = [[4, 5, 0, 10], [0, 8, 4, 5]]
rects3 = [[0, 1, 0, 1], [1, 2, 1, 2], [2, 3, 2, 3],
[3, 5, 3, 6], [3, 6, 3, 7], [3, 7, 3, 8],
[4, 5, 0, 10], [0, 8, 4, 5], [0, 8, 0, 10]]
for rects in [rects1, rects2, rects3]:
for block_size in [3, 4, 10]:
rect_path = new_local_temp_dir()
rect_uri = local_path_uri(rect_path)
bm = BlockMatrix.from_numpy(nd, block_size=block_size)
bm.export_rectangles(rect_uri, rects)
self._assert_rectangles_eq(nd, rect_path, rects)
rect_path_bytes = new_local_temp_dir()
rect_uri_bytes = local_path_uri(rect_path_bytes)
bm.export_rectangles(rect_uri_bytes, rects, binary=True)
self._assert_rectangles_eq(nd, rect_path_bytes, rects, binary=True)
def test_export_rectangles(self):
nd = np.arange(0, 80, dtype=float).reshape(8, 10)
rects1 = [[0, 1, 0, 1], [4, 5, 7, 8]]
rects2 = [[4, 5, 0, 10], [0, 8, 4, 5]]
rects3 = [[0, 1, 0, 1], [1, 2, 1, 2], [2, 3, 2, 3], [3, 5, 3, 6],
[3, 6, 3, 7], [3, 7, 3, 8], [4, 5, 0, 10], [0, 8, 4, 5],
[0, 8, 0, 10]]
for rects in [rects1, rects2, rects3]:
for block_size in [3, 4, 10]:
bm_uri = new_temp_file()
rect_path = new_local_temp_dir()
rect_uri = local_path_uri(rect_path)
(BlockMatrix.from_numpy(
nd,
block_size=block_size).sparsify_rectangles(rects).write(
bm_uri, force_row_major=True))
BlockMatrix.export_rectangles(bm_uri, rect_uri, rects)
for (i, r) in enumerate(rects):
file = rect_path + '/rect-' + str(i) + '_' + '-'.join(
map(str, r))
expected = nd[r[0]:r[1], r[2]:r[3]]
actual = np.loadtxt(file, ndmin=2)
self._assert_eq(expected, actual)
rect_path_bytes = new_local_temp_dir()
rect_uri_bytes = local_path_uri(rect_path_bytes)
BlockMatrix.export_rectangles(bm_uri,
rect_uri_bytes,
rects,
binary=True)
for (i, r) in enumerate(rects):
file = rect_path_bytes + '/rect-' + str(
i) + '_' + '-'.join(map(str, r))
expected = nd[r[0]:r[1], r[2]:r[3]]
actual = np.reshape(np.fromfile(file),
(r[1] - r[0], r[3] - r[2]))
self._assert_eq(expected, actual)
bm_uri = new_temp_file()
rect_uri = new_temp_file()
(BlockMatrix.from_numpy(nd, block_size=5).sparsify_rectangles(
[[0, 1, 0, 1]]).write(bm_uri, force_row_major=True))
with self.assertRaises(FatalError) as e:
BlockMatrix.export_rectangles(bm_uri, rect_uri, [[5, 6, 5, 6]])
self.assertEquals(
e.msg,
'block (1, 1) missing for rectangle 0 with bounds [5, 6, 5, 6]'
)
def test_export_rectangles(self):
nd = np.arange(0, 80, dtype=float).reshape(8, 10)
rects1 = [[0, 1, 0, 1], [4, 5, 7, 8]]
rects2 = [[4, 5, 0, 10], [0, 8, 4, 5]]
rects3 = [[0, 1, 0, 1], [1, 2, 1, 2], [2, 3, 2, 3], [3, 5, 3, 6],
[3, 6, 3, 7], [3, 7, 3, 8], [4, 5, 0, 10], [0, 8, 4, 5],
[0, 8, 0, 10]]
for rects in [rects1, rects2, rects3]:
for block_size in [3, 4, 10]:
rect_path = new_local_temp_dir()
rect_uri = local_path_uri(rect_path)
bm = BlockMatrix.from_numpy(nd, block_size=block_size)
bm.export_rectangles(rect_uri, rects)
self._assert_rectangles_eq(nd, rect_path, rects)
rect_path_bytes = new_local_temp_dir()
rect_uri_bytes = local_path_uri(rect_path_bytes)
bm.export_rectangles(rect_uri_bytes, rects, binary=True)
self._assert_rectangles_eq(nd,
rect_path_bytes,
rects,
binary=True)
def test_export_rectangles(self):
nd = np.arange(0, 80, dtype=float).reshape(8, 10)
rects1 = [[0, 1, 0, 1], [4, 5, 7, 8]]
rects2 = [[4, 5, 0, 10], [0, 8, 4, 5]]
rects3 = [[0, 1, 0, 1], [1, 2, 1, 2], [2, 3, 2, 3],
[3, 5, 3, 6], [3, 6, 3, 7], [3, 7, 3, 8],
[4, 5, 0, 10], [0, 8, 4, 5], [0, 8, 0, 10]]
for rects in [rects1, rects2, rects3]:
for block_size in [3, 4, 10]:
bm_uri = new_temp_file()
rect_path = new_local_temp_dir()
rect_uri = local_path_uri(rect_path)
(BlockMatrix.from_numpy(nd, block_size=block_size)
.sparsify_rectangles(rects)
.write(bm_uri, force_row_major=True))
BlockMatrix.export_rectangles(bm_uri, rect_uri, rects)
for (i, r) in enumerate(rects):
file = rect_path + '/rect-' + str(i) + '_' + '-'.join(map(str, r))
expected = nd[r[0]:r[1], r[2]:r[3]]
actual = np.loadtxt(file, ndmin = 2)
self._assert_eq(expected, actual)
rect_path_bytes = new_local_temp_dir()
rect_uri_bytes = local_path_uri(rect_path_bytes)
BlockMatrix.export_rectangles(bm_uri, rect_uri_bytes, rects, binary=True)
for (i, r) in enumerate(rects):
file = rect_path_bytes + '/rect-' + str(i) + '_' + '-'.join(map(str, r))
expected = nd[r[0]:r[1], r[2]:r[3]]
actual = np.reshape(np.fromfile(file), (r[1] - r[0], r[3] - r[2]))
self._assert_eq(expected, actual)
bm_uri = new_temp_file()
rect_uri = new_temp_file()
(BlockMatrix.from_numpy(nd, block_size=5)
.sparsify_rectangles([[0, 1, 0, 1]])
.write(bm_uri, force_row_major=True))
with self.assertRaises(FatalError) as e:
BlockMatrix.export_rectangles(bm_uri, rect_uri, [[5, 6, 5, 6]])
self.assertEquals(e.msg, 'block (1, 1) missing for rectangle 0 with bounds [5, 6, 5, 6]')
def to_numpy(self):
"""Collects the block matrix into a `NumPy ndarray
<https://docs.scipy.org/doc/numpy/reference/generated/numpy.ndarray.html>`__.
Examples
--------
>>> from hail.linalg import BlockMatrix
>>> bm = BlockMatrix.random(10, 20)
>>> a = bm.to_numpy()
Notes
-----
The number of entries must be less than :math:`2^{31}`.
The resulting ndarray will have the same shape as the block matrix.
Returns
-------
:class:`numpy.ndarray`
"""
path = new_local_temp_file()
uri = local_path_uri(path)
self.tofile(uri)
return np.fromfile(path).reshape((self.n_rows, self.n_cols))
def _jarray_from_ndarray(nd):
if nd.size >= (1 << 31):
raise ValueError(f'size of ndarray must be less than 2^31, found {nd.size}')
nd = _ndarray_as_float64(nd)
path = new_local_temp_file()
uri = local_path_uri(path)
nd.tofile(path)
return Env.hail().utils.richUtils.RichArray.importFromDoubles(Env.hc()._jhc, uri, nd.size)
def test_export_blocks(self):
nd = np.ones(shape=(8, 10))
bm = BlockMatrix.from_numpy(nd, block_size=20)
bm_path = new_local_temp_dir()
bm_uri = local_path_uri(bm_path)
bm.export_blocks(bm_uri, binary=True)
actual = BlockMatrix.rectangles_to_numpy(bm_path, binary=True)
self._assert_eq(nd, actual)
def copy_log(self, path: str) -> None:
log = Env.hc()._log
try:
if self.is_dir(path):
_, tail = os.path.split(log)
path = os.path.join(path, tail)
info(f"copying log to {repr(path)}...")
self.copy(local_path_uri(Env.hc()._log), path)
except Exception as e:
sys.stderr.write(f'Could not copy log: encountered error:\n {e}')
def test_export_blocks(self):
nd = np.ones(shape=(8, 10))
bm = BlockMatrix.from_numpy(nd, block_size=20)
bm_path = new_local_temp_dir()
bm_uri = local_path_uri(bm_path)
bm.export_blocks(bm_uri, binary=True)
actual = BlockMatrix.rectangles_to_numpy(bm_path, binary=True)
self._assert_eq(nd, actual)
def copy_log(self, path: str) -> None:
log = Env.hc()._log
try:
if self.is_dir(path):
_, tail = os.path.split(log)
path = os.path.join(path, tail)
info(f"copying log to {repr(path)}...")
self.copy(local_path_uri(Env.hc()._log), path)
except Exception as e:
sys.stderr.write(f'Could not copy log: encountered error:\n {e}')
def test_rectangles_to_numpy(self):
nd = np.array([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0], [7.0, 8.0, 9.0]])
rects = [[0, 3, 0, 1], [1, 2, 0, 2]]
rect_path = new_local_temp_dir()
rect_uri = local_path_uri(rect_path)
BlockMatrix.from_numpy(nd).export_rectangles(rect_uri, rects)
rect_bytes_path = new_local_temp_dir()
rect_bytes_uri = local_path_uri(rect_bytes_path)
BlockMatrix.from_numpy(nd).export_rectangles(rect_bytes_uri,
rects,
binary=True)
expected = np.array([[1.0, 0.0], [4.0, 5.0], [7.0, 0.0]])
self._assert_eq(expected, BlockMatrix.rectangles_to_numpy(rect_path))
self._assert_eq(
expected,
BlockMatrix.rectangles_to_numpy(rect_bytes_path, binary=True))
def test_rectangles_to_numpy(self):
nd = np.array([[1.0, 2.0, 3.0],
[4.0, 5.0, 6.0],
[7.0, 8.0, 9.0]])
rects = [[0, 3, 0, 1], [1, 2, 0, 2]]
rect_path = new_local_temp_dir()
rect_uri = local_path_uri(rect_path)
BlockMatrix.from_numpy(nd).export_rectangles(rect_uri, rects)
rect_bytes_path = new_local_temp_dir()
rect_bytes_uri = local_path_uri(rect_bytes_path)
BlockMatrix.from_numpy(nd).export_rectangles(rect_bytes_uri, rects, binary=True)
expected = np.array([[1.0, 0.0],
[4.0, 5.0],
[7.0, 0.0]])
self._assert_eq(expected, BlockMatrix.rectangles_to_numpy(rect_path))
self._assert_eq(expected, BlockMatrix.rectangles_to_numpy(rect_bytes_path, binary=True))
def test_export_rectangles_filtered(self):
rect_path = new_local_temp_dir()
rect_uri = local_path_uri(rect_path)
nd = np.array([[1.0, 2.0, 3.0, 4.0], [5.0, 6.0, 7.0, 8.0],
[9.0, 10.0, 11.0, 12.0], [13.0, 14.0, 15.0, 16.0]])
bm = BlockMatrix.from_numpy(nd)
bm = bm[1:3, 1:3]
export_rects = [[0, 1, 0, 2], [1, 2, 0, 2]]
bm.export_rectangles(rect_uri, export_rects)
expected = np.array([[6.0, 7.0], [10.0, 11.0]])
self._assert_rectangles_eq(expected, rect_path, export_rects)
def from_numpy(cls, ndarray, block_size=None):
"""Distributes a `NumPy ndarray
<https://docs.scipy.org/doc/numpy/reference/generated/numpy.ndarray.html>`__
as a block matrix.
Examples
--------
>>> import numpy as np
>>> a = np.random.rand(10, 20)
>>> bm = BlockMatrix.from_numpy(a)
Notes
-----
The ndarray must have two dimensions, each of non-zero size.
The number of entries must be less than :math:`2^{31}`.
Parameters
----------
ndarray: :class:`numpy.ndarray`
ndarray with two dimensions, each of non-zero size.
block_size: :obj:`int`, optional
Block size. Default given by :meth:`default_block_size`.
Returns
-------
:class:`.BlockMatrix`
"""
if not block_size:
block_size = BlockMatrix.default_block_size()
if ndarray.ndim != 2:
raise FatalError(
"from_numpy: ndarray must have two axes, found shape {}".
format(ndarray.shape))
n_rows, n_cols = ndarray.shape
if n_rows == 0 or n_cols == 0:
raise FatalError(
"from_numpy: ndarray dimensions must be non-zero, found shape {}"
.format(ndarray.shape))
if ndarray.dtype != np.float64:
ndarray = ndarray.astype(np.float64)
local_temp_dir = new_local_temp_dir()
path = local_temp_dir + '/binary'
uri = local_path_uri(path)
ndarray.tofile(path)
return cls.fromfile(uri, n_rows, n_cols, block_size)
def test_export_rectangles_sparse(self):
rect_path = new_local_temp_dir()
rect_uri = local_path_uri(rect_path)
nd = np.array([[1.0, 2.0, 3.0, 4.0], [5.0, 6.0, 7.0, 8.0],
[9.0, 10.0, 11.0, 12.0], [13.0, 14.0, 15.0, 16.0]])
bm = BlockMatrix.from_numpy(nd, block_size=2)
sparsify_rects = [[0, 1, 0, 1], [0, 3, 0, 2], [1, 2, 0, 4]]
export_rects = [[0, 1, 0, 1], [0, 3, 0, 2], [1, 2, 0, 4], [2, 4, 2, 4]]
bm.sparsify_rectangles(sparsify_rects).export_rectangles(
rect_uri, export_rects)
expected = np.array([[1.0, 2.0, 3.0, 4.0], [5.0, 6.0, 7.0, 8.0],
[9.0, 10.0, 0.0, 0.0], [13.0, 14.0, 0.0, 0.0]])
self._assert_rectangles_eq(expected, rect_path, export_rects)
def from_numpy(cls, ndarray, block_size=None):
"""Distributes a `NumPy ndarray
<https://docs.scipy.org/doc/numpy/reference/generated/numpy.ndarray.html>`__
as a block matrix.
Examples
--------
>>> import numpy as np
>>> a = np.random.rand(10, 20)
>>> bm = BlockMatrix.from_numpy(a)
Notes
-----
The ndarray must have two dimensions, each of non-zero size.
The number of entries must be less than :math:`2^{31}`.
Parameters
----------
ndarray: :class:`numpy.ndarray`
ndarray with two dimensions, each of non-zero size.
block_size: :obj:`int`, optional
Block size. Default given by :meth:`default_block_size`.
Returns
-------
:class:`.BlockMatrix`
"""
if not block_size:
block_size = BlockMatrix.default_block_size()
if any(i == 0 for i in ndarray.shape):
raise ValueError(
f'from_numpy: ndarray dimensions must be non-zero, found shape {ndarray.shape}'
)
nd = _ndarray_as_2d(ndarray)
nd = _ndarray_as_float64(nd)
n_rows, n_cols = nd.shape
path = new_local_temp_file()
uri = local_path_uri(path)
nd.tofile(path)
return cls.fromfile(uri, n_rows, n_cols, block_size)
def copy_log(path: str) -> None:
"""Attempt to copy the session log to a hadoop-API-compatible location.
Examples
--------
Specify a manual path:
>>> hl.copy_log('gs://my-bucket/analysis-10-jan19.log') # DOCTEST: +SKIP
INFO: copying log to 'gs://my-bucket/analysis-10-jan19.log'...
Copy to a directory:
>>> hl.copy_log('gs://my-bucket/') # DOCTEST: +SKIP
INFO: copying log to 'gs://my-bucket/hail-20180924-2018-devel-46e5fad57524.log'...
Notes
-----
Since Hail cannot currently log directly to distributed file systems, this
function is provided as a utility for offloading logs from ephemeral nodes.
If `path` is a directory, then the log file will be copied using its
base name to the directory (e.g. ``/home/hail.log`` would be copied as
``gs://my-bucket/hail.log`` if `path` is ``gs://my-bucket``.
Parameters
----------
path: :obj:`str`
"""
log = Env.hc()._log
try:
if hadoop_is_dir(path):
_, tail = os.path.split(log)
path = os.path.join(path, tail)
info(f"copying log to {repr(path)}...")
hadoop_copy(local_path_uri(Env.hc()._log), path)
except Exception as e:
sys.stderr.write(f'Could not copy log: encountered error:\n {e}')
def _ndarray_from_jarray(ja):
path = new_local_temp_file()
uri = local_path_uri(path)
Env.hail().utils.richUtils.RichArray.exportToDoubles(Env.hc()._jhc, uri, ja)
return np.fromfile(path)
