# -*- coding: utf-8 -*-
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
from functools import wraps
from africanus.rime.phase import (phase_delay as np_phase_delay,
PHASE_DELAY_DOCS)
from africanus.rime.parangles import parallactic_angles as np_parangles
from africanus.rime.feeds import feed_rotation as np_feed_rotation
from africanus.rime.feeds import FEED_ROTATION_DOCS
from africanus.rime.transform import transform_sources as np_transform_sources
from africanus.rime.beam_cubes import beam_cube_dde as np_beam_cude_dde
from africanus.rime.predict import PREDICT_DOCS, predict_checks
from africanus.rime.predict import predict_vis as np_predict_vis
from africanus.rime.zernike import zernike_dde as np_zernike_dde
from africanus.util.docs import mod_docs
from africanus.util.requirements import requires_optional
from africanus.util.type_inference import infer_complex_dtype
import numpy as np
try:
import dask.array as da
from dask.highlevelgraph import HighLevelGraph
from dask.blockwise import blockwise
except ImportError as e:
da_import_error = e
else:
da_import_error = None
@wraps(np_phase_delay)
def _phase_delay_wrap(lm, uvw, frequency):
return np_phase_delay(lm[0], uvw[0], frequency)
[docs]@requires_optional('dask.array', da_import_error)
def phase_delay(lm, uvw, frequency):
""" Dask wrapper for phase_delay function """
return da.core.blockwise(_phase_delay_wrap, ("source", "row", "chan"),
lm, ("source", "(l,m)"),
uvw, ("row", "(u,v,w)"),
frequency, ("chan",),
dtype=infer_complex_dtype(lm, uvw, frequency))
@wraps(np_parangles)
def _parangle_wrapper(t, ap, fc, **kw):
return np_parangles(t, ap[0], fc[0], **kw)
[docs]@requires_optional('dask.array', da_import_error)
def parallactic_angles(times, antenna_positions, field_centre, **kwargs):
return da.core.blockwise(_parangle_wrapper, ("time", "ant"),
times, ("time",),
antenna_positions, ("ant", "xyz"),
field_centre, ("fc",),
dtype=times.dtype,
**kwargs)
[docs]@requires_optional('dask.array', da_import_error)
def feed_rotation(parallactic_angles, feed_type):
pa_dims = tuple("pa-%d" % i for i in range(parallactic_angles.ndim))
corr_dims = ('corr-1', 'corr-2')
if parallactic_angles.dtype == np.float32:
dtype = np.complex64
elif parallactic_angles.dtype == np.float64:
dtype = np.complex128
else:
raise ValueError("parallactic_angles have "
"non-floating point dtype")
return da.core.blockwise(np_feed_rotation, pa_dims + corr_dims,
parallactic_angles, pa_dims,
feed_type=feed_type,
new_axes={'corr-1': 2, 'corr-2': 2},
dtype=dtype)
@wraps(np_transform_sources)
def _xform_wrap(lm, parallactic_angles, pointing_errors,
antenna_scaling, frequency, dtype_):
return np_transform_sources(lm[0], parallactic_angles,
pointing_errors[0], antenna_scaling,
frequency, dtype=dtype_)
@wraps(np_beam_cude_dde)
def _beam_wrapper(beam, coords, l_grid, m_grid, freq_grid,
spline_order=1, mode='nearest'):
return np_beam_cude_dde(beam[0][0][0], coords[0],
l_grid[0], m_grid[0], freq_grid[0],
spline_order=spline_order, mode=mode)
[docs]@requires_optional('dask.array', da_import_error)
def beam_cube_dde(beam, coords, l_grid, m_grid, freq_grid,
spline_order=1, mode='nearest'):
coord_shapes = coords.shape[1:]
corr_shapes = beam.shape[3:]
corr_dims = tuple("corr-%d" % i for i in range(len(corr_shapes)))
coord_dims = tuple("coord-%d" % i for i in range(len(coord_shapes)))
beam_dims = ("beam_lw", "beam_mh", "beam_nud") + corr_dims
return da.core.blockwise(_beam_wrapper, coord_dims + corr_dims,
beam, beam_dims,
coords, ("coords",) + coord_dims,
l_grid, ("beam_lw",),
m_grid, ("beam_mh",),
freq_grid, ("beam_nud",),
spline_order=spline_order,
mode=mode,
dtype=beam.dtype)
@wraps(np_zernike_dde)
def _zernike_wrapper(coords, coeffs, noll_index):
# coords loses "three" dim
# coeffs loses "poly" dim
# noll_index loses "poly" dim
return np_zernike_dde(coords[0], coeffs[0], noll_index[0])
[docs]@requires_optional('dask.array', da_import_error)
def zernike_dde(coords, coeffs, noll_index):
ncorrs = len(coeffs.shape[2:-1])
corr_dims = tuple("corr-%d" % i for i in range(ncorrs))
return da.core.blockwise(_zernike_wrapper,
("source", "time", "ant", "chan") + corr_dims,
coords,
("three", "source", "time", "ant", "chan"),
coeffs,
("ant", "chan") + corr_dims + ("poly",),
noll_index,
("ant", "chan") + corr_dims + ("poly",),
dtype=coeffs.dtype)
@wraps(np_predict_vis)
def _predict_coh_wrapper(time_index, antenna1, antenna2,
dde1_jones, source_coh, dde2_jones,
die1_jones, base_vis, die2_jones):
return (np_predict_vis(time_index, antenna1, antenna2,
# dde1_jones loses the 'ant' dim
dde1_jones[0] if dde1_jones else None,
# source_coh loses the 'source' dim
source_coh,
# dde2_jones loses the 'source' and 'ant' dims
dde2_jones[0] if dde2_jones else None,
# die1_jones loses the 'ant' dim
die1_jones[0] if die1_jones else None,
base_vis,
# die2_jones loses the 'ant' dim
die2_jones[0] if die2_jones else None)
# Introduce an extra dimension (source dim reduced to 1)
[None, ...])
@wraps(np_predict_vis)
def _predict_dies_wrapper(time_index, antenna1, antenna2,
dde1_jones, source_coh, dde2_jones,
die1_jones, base_vis, die2_jones):
return np_predict_vis(time_index, antenna1, antenna2,
# dde1_jones loses the 'source' and 'ant' dims
dde1_jones[0][0] if dde1_jones else None,
# source_coh loses the 'source' dim
source_coh[0] if source_coh else None,
# dde2_jones loses the 'source' and 'ant' dims
dde2_jones[0][0] if dde2_jones else None,
# die1_jones loses the 'ant' dim
die1_jones[0] if die1_jones else None,
base_vis,
# die2_jones loses the 'ant' dim
die2_jones[0] if die2_jones else None)
[docs]@requires_optional('dask.array', da_import_error)
def predict_vis(time_index, antenna1, antenna2,
dde1_jones=None, source_coh=None, dde2_jones=None,
die1_jones=None, base_vis=None, die2_jones=None):
tup = predict_checks(time_index, antenna1, antenna2,
dde1_jones, source_coh, dde2_jones,
die1_jones, base_vis, die2_jones)
(have_ddes1, have_coh, have_ddes2, have_dies1, have_bvis, have_dies2) = tup
have_ddes = have_ddes1 and have_ddes2
if have_ddes:
if dde1_jones.shape[2] != dde1_jones.chunks[2][0]:
raise ValueError("Subdivision of antenna dimension into "
"multiple chunks is not supported.")
if dde2_jones.shape[2] != dde2_jones.chunks[2][0]:
raise ValueError("Subdivision of antenna dimension into "
"multiple chunks is not supported.")
if dde1_jones.chunks != dde2_jones.chunks:
raise ValueError("dde1_jones.chunks != dde2_jones.chunks")
if len(dde1_jones.chunks[1]) != len(time_index.chunks[0]):
raise ValueError("Number of row chunks (%s) does not equal "
"number of time chunks (%s)." %
(time_index.chunks[0], dde1_jones.chunks[1]))
have_dies = have_dies1 and have_dies2
if have_dies:
if die1_jones.shape[1] != die1_jones.chunks[1][0]:
raise ValueError("Subdivision of antenna dimension into "
"multiple chunks is not supported.")
if die2_jones.shape[1] != die2_jones.chunks[1][0]:
raise ValueError("Subdivision of antenna dimension into "
"multiple chunks is not supported.")
if die1_jones.chunks != die2_jones.chunks:
raise ValueError("die1_jones.chunks != die2_jones.chunks")
if len(die1_jones.chunks[0]) != len(time_index.chunks[0]):
raise ValueError("Number of row chunks (%s) does not equal "
"number of time chunks (%s)." %
(time_index.chunks[0], die1_jones.chunks[1]))
# Generate strings for the correlation dimensions
if have_ddes:
cdims = tuple("corr-%d" % i for i in range(len(dde1_jones.shape[4:])))
elif have_coh:
cdims = tuple("corr-%d" % i for i in range(len(source_coh.shape[3:])))
elif have_dies:
cdims = tuple("corr-%d" % i for i in range(len(die1_jones.shape[3:])))
else:
raise ValueError("Missing both antenna and baseline jones terms")
# Infer the output dtype
dtype_arrays = [dde1_jones, source_coh, dde2_jones, die1_jones, die2_jones]
out_dtype = np.result_type(*(np.dtype(a.dtype.name)
for a in dtype_arrays if a is not None))
# In the case of predict_vis, the "row" and "time" dimensions
# are intimately related -- a contiguous series of rows
# are related to a contiguous series of timesteps.
# This means that the number of chunks of these
# two dimensions must match even though the chunk sizes may not.
# blockwise insists on matching chunk sizes.
# For this reason, we use the lower level blockwise and
# substitute "row" for "time" in arrays such as dde1_jones
# and die1_jones.
token = da.core.tokenize(time_index, antenna1, antenna2,
dde1_jones, source_coh, dde2_jones, base_vis)
ajones_dims = ("src", "row", "ant", "chan") + cdims
gjones_dims = ("row", "ant", "chan") + cdims
# Setup
# 1. Optional blockwise arguments
# 2. Optional numblocks kwarg
# 3. HighLevelGraph dependencies
bw_args = [time_index.name, ("row",),
antenna1.name, ("row",),
antenna2.name, ("row",)]
numblocks = {
time_index.name: time_index.numblocks,
antenna1.name: antenna1.numblocks,
antenna2.name: antenna2.numblocks
}
# Dependencies
deps = [time_index, antenna1, antenna2]
# Handle presence/absence of dde1_jones
if have_ddes:
bw_args.extend([dde1_jones.name, ajones_dims])
numblocks[dde1_jones.name] = dde1_jones.numblocks
deps.append(dde1_jones)
other_chunks = dde1_jones.chunks[3:]
src_chunks = dde1_jones.chunks[0]
else:
bw_args.extend([None, None])
# Handle presence/absence of source_coh
if have_coh:
bw_args.extend([source_coh.name, ("src", "row", "chan") + cdims])
numblocks[source_coh.name] = source_coh.numblocks
other_chunks = source_coh.chunks[2:]
src_chunks = source_coh.chunks[0]
deps.append(source_coh)
else:
bw_args.extend([None, None])
# Handle presence/absence of dde2_jones
if have_ddes:
bw_args.extend([dde2_jones.name, ajones_dims])
numblocks[dde2_jones.name] = dde2_jones.numblocks
other_chunks = dde1_jones.chunks[3:]
deps.append(dde2_jones)
other_chunks = dde2_jones.chunks[3:]
src_chunks = dde1_jones.chunks[0]
else:
bw_args.extend([None, None])
# die1_jones, base_vis and die2_jones absent for this part of the graph
bw_args.extend([None, None, None, None, None, None])
assert len(bw_args) // 2 == 9, len(bw_args) // 2
name = "-".join(("predict_vis", token))
layer = blockwise(_predict_coh_wrapper,
name, ("src", "row", "chan") + cdims,
*bw_args, numblocks=numblocks)
graph = HighLevelGraph.from_collections(name, layer, deps)
# We can infer output chunk sizes from source_coh
chunks = ((1,)*len(src_chunks), time_index.chunks[0],) + other_chunks
sum_coherencies = da.Array(graph, name, chunks, dtype=out_dtype)
sum_coherencies = sum_coherencies.sum(axis=0)
if have_bvis:
sum_coherencies += base_vis
if not have_dies:
return sum_coherencies
# Now apply any Direction Independent Effect Terms
# Setup
# 1. Optional blockwise arguments
# 2. Optional numblocks kwarg
# 3. HighLevelGraph dependencies
bw_args = [time_index.name, ("row",),
antenna1.name, ("row",),
antenna2.name, ("row",)]
numblocks = {
time_index.name: time_index.numblocks,
antenna1.name: antenna1.numblocks,
antenna2.name: antenna2.numblocks
}
deps = [time_index, antenna1, antenna2]
# dde1_jones, source_coh and dde2_jones not present
bw_args.extend([None, None, None, None, None, None])
bw_args.extend([die1_jones.name, gjones_dims])
bw_args.extend([sum_coherencies.name, ("row", "chan") + cdims])
bw_args.extend([die2_jones.name, gjones_dims])
numblocks[die1_jones.name] = die1_jones.numblocks
numblocks[sum_coherencies.name] = sum_coherencies.numblocks
numblocks[die2_jones.name] = die2_jones.numblocks
deps.extend([die1_jones, sum_coherencies, die2_jones])
assert len(bw_args) // 2 == 9
token = da.core.tokenize(time_index, antenna1, antenna2,
die1_jones, sum_coherencies, die2_jones)
name = '-'.join(("predict_vis", token))
layer = blockwise(_predict_dies_wrapper,
name, ("row", "chan") + cdims,
*bw_args, numblocks=numblocks)
graph = HighLevelGraph.from_collections(name, layer, deps)
chunks = (time_index.chunks[0],) + other_chunks
return da.Array(graph, name, chunks, dtype=out_dtype)
try:
phase_delay.__doc__ = PHASE_DELAY_DOCS.substitute(
array_type=":class:`dask.array.Array`")
parallactic_angles.__doc__ = mod_docs(np_parangles.__doc__,
[(":class:`numpy.ndarray`",
":class:`dask.array.Array`")])
feed_rotation.__doc__ = FEED_ROTATION_DOCS.substitute(
array_type=":class:`numpy.ndarray`")
transform_sources.__doc__ = mod_docs(np_transform_sources.__doc__,
[(":class:`numpy.ndarray`",
":class:`dask.array.Array`")])
beam_cube_dde.__doc__ = mod_docs(np_beam_cude_dde.__doc__,
[(":class:`numpy.ndarray`",
":class:`dask.array.Array`")])
zernike_dde.__doc__ = mod_docs(np_zernike_dde.__doc__,
[(":class:`numpy.ndarray`",
":class:`dask.array.Array`")])
EXTRA_DASK_NOTES = """
* The ``ant`` dimension should only contain a single chunk equal
to the number of antenna. Since each ``row`` can contain
any antenna, random access must be preserved along this dimension.
* The chunks in the ``row`` and ``time`` dimension **must** align.
This subtle point **must be understood otherwise
invalid results will be produced** by the chunking scheme.
In the example below
we have four unique time indices :code:`[0,1,2,3]`, and
four unique antenna :code:`[0,1,2,3]` indexing :code:`10` rows.
.. code-block:: python
# Row indices into the time/antenna indexed arrays
time_idx = np.asarray([0,0,1,1,2,2,2,2,3,3])
ant1 = np.asarray( [0,0,0,0,1,1,1,2,2,3]
ant2 = np.asarray( [0,1,2,3,1,2,3,2,3,3])
A reasonable chunking scheme for the
``row`` and ``time`` dimension would be :code:`(4,4,2)`
and :code:`(2,1,1)` respectively.
Another way of explaining this is that the first
four rows contain two unique timesteps, the second four
rows contain one unique timestep and the last two rows
contain one unique timestep.
Some rules of thumb:
1. The number chunks in ``row`` and ``time`` must match
although the individual chunk sizes need not.
2. Unique timesteps should not be split across row chunks.
3. For a Measurement Set whose rows are ordered on the
``TIME`` column, the following is a good way of obtaining
the row chunking strategy:
.. code-block:: python
import numpy as np
import pyrap.tables as pt
ms = pt.table("data.ms")
times = ms.getcol("TIME")
unique_times, chunks = np.unique(times, return_counts=True)
4. Use :func:`~africanus.util.shapes.aggregate_chunks`
to aggregate multiple ``row`` and ``time``
chunks into chunks large enough such that functions operating
on the resulting data can drop the GIL and spend time
processing the data. Expanding the previous example:
.. code-block:: python
# Aggregate row
utimes = unique_times.size
# Single chunk for each unique time
time_chunks = (1,)*utimes
# Aggregate row chunks into chunks <= 10000
aggregate_chunks((chunks, time_chunks), (10000, utimes))
"""
predict_vis.__doc__ = PREDICT_DOCS.substitute(
array_type=":class:`dask.array.Array`",
extra_notes=EXTRA_DASK_NOTES)
except AttributeError:
pass