# Licensed under the GPLv3 - see LICENSE
import warnings
from collections import namedtuple
import numpy as np
import astropy.units as u
from astropy.utils import lazyproperty
from ..vlbi_base.base import (make_opener, VLBIFileBase, VLBIFileReaderBase,
VLBIStreamBase, VLBIStreamReaderBase,
VLBIStreamWriterBase, HeaderNotFoundError)
from .header import VDIFHeader
from .payload import VDIFPayload
from .frame import VDIFFrame, VDIFFrameSet
from .file_info import VDIFFileReaderInfo
__all__ = ['VDIFFileReader', 'VDIFFileWriter', 'VDIFStreamBase',
'VDIFStreamReader', 'VDIFStreamWriter', 'open']
# Check code on 2015-MAY-30
# 00000000 77 2c db 00 00 00 00 1c 75 02 00 20 fc ff 01 04 # header 0 - 3
# 00000010 10 00 80 03 ed fe ab ac 00 00 40 33 83 15 03 f2 # header 4 - 7
# 00000020 2a 0a 7c 43 8b 69 9d 59 cb 99 6d 9a 99 96 5d 67 # data 0 - 3
# NOTE: thread_id = 1
# 2a = 00 10 10 10 = (lsb first) 1, 1, 1, -3
# 0a = 00 00 10 10 = 1, 1, -3, -3
# 7c = 01 11 11 00 = -3, 3, 3, -1
# m5d evn/Fd/GP052D_FD_No0006.m5a VDIF_5000-512-1-2 100
# Mark5 stream: 0x16cd140
# stream = File-1/1=evn/Fd/GP052D_FD_No0006.m5a
# format = VDIF_5000-512-1-2 = 3
# start mjd/sec = 56824 21367.000000000
# frame duration = 78125.00 ns
# framenum = 0
# sample rate = 256000000 Hz
# offset = 0
# framebytes = 5032 bytes
# datasize = 5000 bytes
# sample granularity = 4
# frame granularity = 1
# gframens = 78125
# payload offset = 32
# read position = 0
# data window size = 1048576 bytes
# 1 1 1 -3 1 1 -3 -3 -3 3 3 -1 -> OK
# fh = vdif.open('evn/Fd/GP052D_FD_No0006.m5a', 'rb')
# fs = fh.read_frameset()
# fs.data.astype(int)[1, :12, 0] # thread id = 1!!
# -> array([ 1, 1, 1, -3, 1, 1, -3, -3, -3, 3, 3, -1]) -> OK
# Also, next frame (thread #3)
# m5d evn/Fd/GP052D_FD_No0006.m5a VDIF_5000-512-1-2 12 5032
# -1 1 -1 1 -3 -1 3 -1 3 -3 1 3
# fs.data.astype(int)[3, :12, 0]
# -> array([-1, 1, -1, 1, -3, -1, 3, -1, 3, -3, 1, 3])
# And first thread #0
# m5d evn/Fd/GP052D_FD_No0006.m5a VDIF_5000-512-1-2 12 20128
# -1 -1 3 -1 1 -1 3 -1 1 3 -1 1
# fs.data.astype(int)[0, :12, 0]
# -> array([-1, -1, 3, -1, 1, -1, 3, -1, 1, 3, -1, 1])
# sanity check that we can read 12 samples with stream reader
# fh.close()
# fh = vdif.open('evn/Fd/GP052D_FD_No0006.m5a', 'rs')
# fh.read(12).astype(int)[:, 0]
# -> array([-1, -1, 3, -1, 1, -1, 3, -1, 1, 3, -1, 1])
[docs]class VDIFFileReader(VLBIFileReaderBase):
"""Simple reader for VDIF files.
Wraps a binary filehandle, providing methods to help interpret the data,
such as `read_frame`, `read_frameset` and `get_frame_rate`.
Parameters
----------
fh_raw : filehandle
Filehandle of the raw binary data file.
"""
info = VDIFFileReaderInfo()
[docs] def read_frame(self, edv=None, verify=True):
"""Read a single frame (header plus payload).
Parameters
----------
edv : int, optional
The expected extended data version for the VDIF Header. If `None`,
use that of the first frame. (Passing it in slightly improves file
integrity checking.)
verify : bool, optional
Whether to do basic checks of frame integrity. Default: `True`.
Returns
-------
frame : `~baseband.vdif.VDIFFrame`
With ``.header`` and ``.data`` properties that return the
:class:`~baseband.vdif.VDIFHeader` and data encoded in the frame,
respectively.
"""
return VDIFFrame.fromfile(self.fh_raw, edv=edv, verify=verify)
[docs] def read_frameset(self, thread_ids=None, edv=None, verify=True):
"""Read a single frame (header plus payload).
Parameters
----------
thread_ids : list, optional
The thread ids that should be read. If `None` (default), read all
threads.
edv : int, optional
The expected extended data version for the VDIF Header. If `None`,
use that of the first frame. (Passing it in slightly improves file
integrity checking.)
verify : bool, optional
Whether to do basic checks of frame integrity. Default: `True`.
Returns
-------
frameset : :class:`~baseband.vdif.VDIFFrameSet`
With ``.headers`` and ``.data`` properties that return a list of
:class:`~baseband.vdif.VDIFHeader` and the data encoded in the
frame set, respectively.
"""
return VDIFFrameSet.fromfile(self.fh_raw, thread_ids, edv=edv,
verify=verify)
[docs] def get_frame_rate(self):
"""Determine the number of frames per second.
This method first tries to determine the frame rate by looking for
the highest frame number in the first second of data. If that fails,
it attempts to extract the sample rate from the header.
Returns
-------
frame_rate : `~astropy.units.Quantity`
Frames per second.
"""
try:
return super().get_frame_rate()
except Exception as exc:
with self.temporary_offset(0):
try:
header = self.read_header()
return (header.sample_rate
/ header.samples_per_frame).to(u.Hz).round()
except Exception:
pass
raise exc
[docs] def get_thread_ids(self, check=2):
"""Determine the number of threads in the VDIF file.
The file is presumed to be positioned at the start of a header.
Usually, it suffices to just seek to the start of the file, but
if not, use `~baseband.vdif.base.VDIFFileReader.find_header`.
Parameters
----------
check : int, optional
Number of extra frames to check. Frame sets are scanned until
the number of thread IDs found no longer increases for ``check``
frames.
Returns
-------
thread_ids : list
Sorted list of all thread ids encountered in the frames scanned.
"""
with self.temporary_offset():
header = header0 = self.read_header()
try:
thread_ids = set()
n_check = 1
while n_check > 0:
frame_nr = header['frame_nr']
n_thread = len(thread_ids)
while header['frame_nr'] == frame_nr:
thread_ids.add(header['thread_id'])
self.seek(header.payload_nbytes, 1)
header = self.read_header(edv=header0.edv)
assert header0.same_stream(header)
if len(thread_ids) > n_thread:
n_check = check
else:
n_check -= 1
except EOFError:
# Hack: let through very short files (like our samples).
if self.seek(0, 2) > ((check+2) * len(thread_ids)
* header0.frame_nbytes):
raise
return sorted(thread_ids)
[docs]class VDIFFileWriter(VLBIFileBase):
"""Simple writer for VDIF files.
Adds `write_frame` and `write_frameset` methods to the basic VLBI
binary file wrapper.
"""
[docs] def write_frame(self, data, header=None, **kwargs):
"""Write a single frame (header plus payload).
Parameters
----------
data : `~numpy.ndarray` or `~baseband.vdif.VDIFFrame`
If an array, a ``header`` should be given, which will be used to
get the information needed to encode the array, and to construct
the VDIF frame.
header : `~baseband.vdif.VDIFHeader`
Can instead give keyword arguments to construct a header. Ignored
if ``data`` is a `~baseband.vdif.VDIFFrame` instance.
**kwargs
If ``header`` is not given, these are used to initialize one.
"""
if not isinstance(data, VDIFFrame):
data = VDIFFrame.fromdata(data, header, **kwargs)
return data.tofile(self.fh_raw)
[docs] def write_frameset(self, data, header=None, **kwargs):
"""Write a single frame set (headers plus payloads).
Parameters
----------
data : `~numpy.ndarray` or :class:`~baseband.vdif.VDIFFrameSet`
If an array, a header should be given, which will be used to
get the information needed to encode the array, and to construct
the VDIF frame set.
header : :class:`~baseband.vdif.VDIFHeader`, list of same
Can instead give keyword arguments to construct a header. Ignored
if ``data`` is a :class:`~baseband.vdif.VDIFFrameSet` instance.
If a list, should have a length matching the number of threads in
``data``; if a single header, ``thread_ids`` corresponding
to the number of threads are generated automatically.
**kwargs
If ``header`` is not given, these are used to initialize one.
"""
if not isinstance(data, VDIFFrameSet):
data = VDIFFrameSet.fromdata(data, header, **kwargs)
return data.tofile(self.fh_raw)
[docs]class VDIFStreamBase(VLBIStreamBase):
"""Base for VDIF streams."""
_sample_shape_maker = namedtuple('SampleShape', 'nthread, nchan')
def __init__(self, fh_raw, header0, sample_rate=None, nthread=1,
squeeze=True, subset=(), fill_value=0., verify=True):
samples_per_frame = header0.samples_per_frame
super().__init__(
fh_raw=fh_raw, header0=header0, sample_rate=sample_rate,
samples_per_frame=samples_per_frame,
unsliced_shape=(nthread, header0.nchan), bps=header0.bps,
complex_data=header0['complex_data'], squeeze=squeeze,
subset=subset, fill_value=fill_value, verify=verify)
def _get_time(self, header):
"""Get time from a header.
This passes on sample rate, since not all VDIF headers can calculate
it.
"""
return header.get_time(frame_rate=self._frame_rate)
def _set_time(self, header, time):
"""Set time in a header.
This passes on sample rate, which not all VDIF headers can calculate.
"""
header.update(time=time, frame_rate=self._frame_rate)
def _get_index(self, header):
# Override to avoid explicit time calculations.
return int(round((header['seconds'] - self.header0['seconds'])
* self._frame_rate.to_value(u.Hz)
+ header['frame_nr'] - self.header0['frame_nr']))
def _set_index(self, header, index):
# Override to avoid explicit time calculations.
dt, frame_nr = divmod(index + self.header0['frame_nr'],
int(round(self._frame_rate.to_value(u.Hz))))
seconds = self.header0['seconds'] + dt
header['seconds'] = seconds
header['frame_nr'] = frame_nr
def __repr__(self):
return ("<{s.__class__.__name__} name={s.name} offset={s.offset}\n"
" sample_rate={s.sample_rate},"
" samples_per_frame={s.samples_per_frame},\n"
" sample_shape={s.sample_shape},\n"
" bps={h.bps}, complex_data={s.complex_data},"
" edv={h.edv}, station={h.station},\n"
" {sub}start_time={s.start_time}>"
.format(s=self, h=self.header0,
sub=('subset={0}, '.format(self.subset)
if self.subset else '')))
[docs]class VDIFStreamReader(VDIFStreamBase, VLBIStreamReaderBase):
"""VLBI VDIF format reader.
Allows access to a VDIF file as a continuous series of samples.
Parameters
----------
fh_raw : filehandle
Filehandle of the raw VDIF stream.
sample_rate : `~astropy.units.Quantity`, optional
Number of complete samples per second, i.e. the rate at which each
channel in each thread is sampled. If `None` (default), will be
inferred from the header or by scanning one second of the file.
squeeze : bool, optional
If `True` (default), remove any dimensions of length unity from
decoded data.
subset : indexing object or tuple of objects, optional
Specific components of the complete sample to decode (after possible
squeezing). If a single indexing object is passed, it selects threads.
If a tuple is passed, the first selects threads and the second selects
channels. If the tuple is empty (default), all components are read.
fill_value : float or complex, optional
Value to use for invalid or missing data. Default: 0.
verify : bool, optional
Whether to do basic checks of frame integrity when reading. The first
frameset of the stream is always checked. Default: `True`.
"""
def __init__(self, fh_raw, sample_rate=None, squeeze=True, subset=(),
fill_value=0., verify='fix'):
fh_raw = VDIFFileReader(fh_raw)
# We read the very first header, hoping this is the right one
# (in some VLBA files not all the headers have the right time).
header0 = fh_raw.read_header()
# Next, we determine how many threads there are, and use those
# to calculate the frameset size. We on purpose do *not* just read
# a frame set, since sometimes the first one is short (see gh-359).
fh_raw.seek(0)
thread_ids = fh_raw.get_thread_ids()
nthread = len(thread_ids)
super().__init__(
fh_raw, header0, sample_rate=sample_rate,
nthread=nthread, squeeze=squeeze, subset=subset,
fill_value=fill_value, verify=verify)
self._raw_offsets.frame_nbytes *= nthread
# Check whether we are reading only some threads. This is somewhat
# messy since normally we apply the whole subset to the whole data,
# but here we need to split it up in the part that selects specific
# threads, which we use to selectively read, and the rest, which we
# do post-decoding.
if self.subset and (nthread > 1 or not self.squeeze):
# Select the thread ids we want using first part of subset.
thread_ids = np.array(thread_ids)[self.subset[0]]
# Use squeese in case subset[0] uses broadcasting, and
# atleast_1d to ensure single threads get upgraded to a list,
# which is needed by the VDIFFrameSet reader.
self._thread_ids = np.atleast_1d(thread_ids.squeeze()).tolist()
# Since we have subset the threads already, we now need to
# determine a new subset that takes this into account.
if thread_ids.shape == ():
# If we indexed with a scalar, we're meant to remove that
# dimension. If we squeeze, this happens automatically, but if
# not, we need to do it explicitly (FrameSet does not do it).
new_subset0 = () if self.squeeze else (0,)
elif len(self._thread_ids) == 1 and self.squeeze:
# If we want a single remaining thread, undo the squeeze.
new_subset0 = (np.newaxis,)
else:
# Just pass on multiple threads or unsqueezed single ones.
new_subset0 = (slice(None),)
self._frameset_subset = new_subset0 + self.subset[1:]
else:
# We either have no subset or we have a single thread that
# will be squeezed away, so the subset is fine as is.
self._frameset_subset = self.subset
self._thread_ids = thread_ids
@lazyproperty
def _last_header(self):
"""Last header of the file."""
# Go to end of file.
maximum = 2 * self._raw_offsets.frame_nbytes
with self.fh_raw.temporary_offset(
-self.header0.frame_nbytes, 2) as fh_raw:
# Find first header with same thread_id going backward.
locations = fh_raw.locate_frames(self.header0, forward=False,
maximum=maximum, check=(-1, 1))
for location in locations:
fh_raw.seek(location)
try:
header = fh_raw.read_header(edv=self.header0.edv)
except Exception:
continue
if header['thread_id'] == self.header0['thread_id']:
return header
raise HeaderNotFoundError(
"corrupt VDIF? No thread_id={0} frame in last {1} bytes."
.format(self.header0['thread_id'], maximum))
def _squeeze_and_subset(self, data):
# Overwrite VLBIStreamReaderBase version, since the threads part of
# subset has already been used.
if self.squeeze:
data = data.reshape(data.shape[:1]
+ tuple(sh for sh in data.shape[1:] if sh > 1))
if self._frameset_subset:
data = data[(slice(None),) + self._frameset_subset]
return data
# Overrides to deal with framesets instead of frames.
def _fh_raw_read_frame(self):
return self.fh_raw.read_frameset(self._thread_ids,
edv=self.header0.edv,
verify=self.verify)
def _bad_frame(self, index, frameset, exc):
# Duplication of base class, but able to deal with missing
# frames inside a frame set.
if (frameset is not None and self._get_index(frameset) == index
and index == self._get_index(self._last_header)):
# If we got an exception because we're trying to read beyond the
# last frame, the frame is almost certainly OK, so keep it.
return frameset
if self.verify != 'fix':
raise exc
# If the frameset does contain the right number of frames, but all
# are invalid, assume not just the data but also the frame number
# and seconds might be wrong, i.e., just proceed with it.
# TODO: make this an option for a specific type of fixing!
if (frameset is not None
and len(frameset.frames) == len(self._thread_ids)
and not any(frame.valid for frame in frameset.frames)):
return frameset
msg = 'problem loading frame set {}.'.format(index)
# Where should we be?
raw_offset = self._seek_frame(index)
# See if we're in the right place. First ensure we have a header.
# Here, it is more important that it is a good one than that we go
# too far, so we insist on two consistent frames after it, as well
# as a good one before to guard against corruption of the start of
# the VDIF header.
self.fh_raw.seek(raw_offset)
try:
header = self.fh_raw.find_header(
self.header0, forward=True, check=(-1, 1, 2),
maximum=3*self.header0.frame_nbytes)
except HeaderNotFoundError:
exc.args += (msg + ' Cannot find header nearby.',)
raise exc
# Don't yet know how to deal with excess data.
header_index = self._get_index(header)
if header_index < index:
exc.args += (msg + ' There appears to be excess data.',)
raise exc
# Go backward until we find previous frame, storing offsets
# as we go. We again increase the maximum since we may need
# to jump over a bad bit. We slightly relax our search pattern.
while header_index >= index:
raw_pos = self.fh_raw.tell()
header1 = header
header1_index = header_index
if raw_pos <= 0:
break
self.fh_raw.seek(-1, 1)
try:
header = self.fh_raw.find_header(
self.header0, forward=False,
maximum=4*self.header0.frame_nbytes,
check=(-1, 1))
except HeaderNotFoundError:
exc.args += (msg + ' Could not find previous index.',)
raise exc
header_index = self._get_index(header)
if header_index < header1_index:
# While we are at it: if we pass an index boundary,
# update the list of known indices.
self._raw_offsets[header1_index] = raw_pos
# Move back to position of last good header (header1).
self.fh_raw.seek(raw_pos)
# Create the header we will use below for constructing the
# frameset. Usually, this is guaranteed to be from this set,
# but we also use it to create a new header for a completely
# messed up frameset below. It is copied to make it mutable
# without any risk of messing up possibly memory mapped data.
header = header1.copy()
if header1_index > index:
# Ouch, whole frame set missing!
msg += ' The frame set seems to be missing altogether.'
# Set up to construct a complete missing frame
# (after the very long else clause).
# TODO: just use the writer's _make_frame??
frames = {}
self._set_index(header, index)
else:
assert header1_index == index, \
'at this point, we should have a good header.'
# This header is the first one of its set.
if raw_pos != raw_offset:
msg += ' Stream off by {0} bytes.'.format(raw_offset
- raw_pos)
# Above, we should have added information about
# this index in our offset table.
assert raw_pos == self._raw_offsets[index]
# Try again to read it, however many threads there are.
# TODO: this somewhat duplicates FrameSet.fromfile; possibly
# move code there.
# TODO: Or keep track of header locations above.
# TODO: remove limitation that threads need to be together.
frames = {}
previous = False
frame_nr = header1['frame_nr']
while True:
raw_pos = self.fh_raw.tell()
try:
frame = self.fh_raw.read_frame(edv=self.header0.edv)
assert header.same_stream(frame.header)
# Check seconds as well, as a sanity check this is
# a real header. (We do allow at this point it to be
# the next frame, hence the second can increase by 1.
assert 0 <= (frame['seconds'] - header['seconds']) <= 1
except EOFError:
# End of file while reading a frame; we're done here.
next_header = None
break
except AssertionError:
# Frame is not OK.
assert previous is not False, \
('first frame should be readable if fully on disk,'
' since we found one correct header.')
# Go back to after previous payload and try finding
# next header. It can be before where we tried above,
# if some bytes in the previous payload were missing.
self.fh_raw.seek(raw_pos - header.payload_nbytes)
try:
next_header = self.fh_raw.find_header(self.header0)
# But sometimes a header is re-found even when
# there isn't really one (e.g., because one of the
# first bytes, defining seconds, is missing).
# Don't ever retry the same one!
if self.fh_raw.tell() == raw_pos:
self.fh_raw.seek(1, 1)
next_header = self.fh_raw.find_header(self.header0)
except HeaderNotFoundError:
# If no header was found, give up. The previous frame
# was likely bad too, so delete it.
if previous is not None:
del frames[previous]
next_header = None
break
# If the next header is not exactly a frame away from
# where we were trying to read, the previous frame was
# likely bad, so discard it.
if self.fh_raw.tell() != raw_pos + header.frame_nbytes:
if previous is not None:
del frames[previous]
previous = None
# Stop if the next header is from a different frame.
if next_header['frame_nr'] != frame_nr:
break
else:
# Successfully read frame. If not of the requested
# set, rewind and break out of the loop.
if frame['frame_nr'] != frame_nr:
next_header = frame.header
self.fh_raw.seek(raw_pos)
break
# Do we have a good frame, giving a new thread?
previous = frame['thread_id']
if previous in frames:
msg += (' Duplicate thread {0} found; discarding.'
.format(previous))
del frames[previous]
else:
# Looks like it, though may still be discarded
# if the next frame is not readable.
frames[previous] = frame
# If the next header is of the next frame, set up the raw
# offset (which likely will be needed, saving some time).
if (next_header is not None
and self._get_index(next_header) == index + 1):
self._raw_offsets[index+1] = self.fh_raw.tell()
# Create invalid frame template,
invalid_payload = VDIFPayload(
np.zeros(header.payload_nbytes // 4, '<u4'), header)
invalid_frame = VDIFFrame(header, invalid_payload, valid=False)
frame_list = []
missing = []
for thread in self._thread_ids:
if thread in frames:
frame_list.append(frames[thread])
else:
missing.append(thread)
invalid_frame.header['thread_id'] = thread
frame_list.append(invalid_frame)
if missing:
if frames == {}:
msg += ' All threads set to invalid.'
else:
msg += (' Thread(s) {0} missing; set to invalid.'
.format(missing))
warnings.warn(msg)
frameset = VDIFFrameSet(frame_list)
return frameset
[docs]class VDIFStreamWriter(VDIFStreamBase, VLBIStreamWriterBase):
"""VLBI VDIF format writer.
Encodes and writes sequences of samples to file.
Parameters
----------
fh_raw : filehandle
Which will write filled sets of frames to storage.
header0 : :class:`~baseband.vdif.VDIFHeader`
Header for the first frame, holding time information, etc. Can instead
give keyword arguments to construct a header (see ``**kwargs``).
sample_rate : `~astropy.units.Quantity`
Number of complete samples per second, i.e. the rate at which each
channel in each thread is sampled. For EDV 1 and 3, can
alternatively set ``sample_rate`` within the header.
nthread : int, optional
Number of threads (e.g., 2 for 2 polarisations). Default: 1.
squeeze : bool, optional
If `True` (default), `write` accepts squeezed arrays as input, and
adds any dimensions of length unity.
**kwargs
If no header is given, an attempt is made to construct one from these.
For a standard header, this would include the following.
--- Header keywords : (see :meth:`~baseband.vdif.VDIFHeader.fromvalues`)
time : `~astropy.time.Time`
Start time of the file. Can instead pass on ``ref_epoch`` and
``seconds``.
nchan : int, optional
Number of channels (default: 1). Note: different numbers of channels
per thread is not supported.
complex_data : bool, optional
Whether data are complex. Default: `False`.
bps : int, optional
Bits per elementary sample, i.e. per real or imaginary component for
complex data. Default: 1.
samples_per_frame : int
Number of complete samples per frame. Can alternatively use
``frame_length``, the number of 8-byte words for header plus payload.
For some EDV, this number is fixed (e.g., ``frame_length=629`` for
``edv=3``, which corresponds to 20000 real 2-bit samples per frame).
station : 2 characters, optional
Station ID. Can also be an unsigned 2-byte integer. Default: 0.
edv : {`False`, 0, 1, 2, 3, 4, 0xab}
Extended Data Version.
"""
def __init__(self, fh_raw, header0=None, sample_rate=None, nthread=1,
squeeze=True, **kwargs):
fh_raw = VDIFFileWriter(fh_raw)
if header0 is None:
if sample_rate is not None:
kwargs['sample_rate'] = sample_rate
header0 = VDIFHeader.fromvalues(**kwargs)
# If header was passed but not sample_rate, extract sample_rate.
if sample_rate is None:
try:
sample_rate = header0.sample_rate
except AttributeError:
raise ValueError("the sample rate must be passed either "
"explicitly, or through the header if it "
"can be stored there.")
super().__init__(fh_raw, header0, sample_rate=sample_rate,
nthread=nthread, squeeze=squeeze)
# Set sample rate in the header, if it's possible, and not set already.
try:
header_sample_rate = self.header0.sample_rate
except AttributeError:
pass
else:
if header_sample_rate == 0:
self.header0.sample_rate = self.sample_rate
assert self.header0.sample_rate == self.sample_rate
self._frame = VDIFFrameSet.fromdata(
np.zeros((self.samples_per_frame,) + self._unsliced_shape,
dtype=np.complex64 if self.complex_data else np.float32),
self.header0)
open = make_opener('VDIF', globals(), doc="""
--- For reading a stream : (see :class:`~baseband.vdif.base.VDIFStreamReader`)
sample_rate : `~astropy.units.Quantity`, optional
Number of complete samples per second, i.e. the rate at which each channel
in each thread is sampled. If `None` (default), will be inferred from the
header or by scanning one second of the file.
squeeze : bool, optional
If `True` (default), remove any dimensions of length unity from
decoded data.
subset : indexing object or tuple of objects, optional
Specific components of the complete sample to decode (after possible
squeezing). If a single indexing object is passed, it selects threads.
If a tuple is passed, the first selects threads and the second selects
channels. If the tuple is empty (default), all components are read.
fill_value : float or complex, optional
Value to use for invalid or missing data. Default: 0.
verify : bool, optional
Whether to do basic checks of frame integrity when reading. The first
frameset of the stream is always checked. Default: `True`.
--- For writing a stream : (see :class:`~baseband.vdif.base.VDIFStreamWriter`)
header0 : `~baseband.vdif.VDIFHeader`
Header for the first frame, holding time information, etc. Can instead
give keyword arguments to construct a header (see ``**kwargs``).
sample_rate : `~astropy.units.Quantity`
Number of complete samples per second, i.e. the rate at which each
channel in each thread is sampled. For EDV 1 and 3, can alternatively set
``sample_rate`` within the header.
nthread : int, optional
Number of threads (e.g., 2 for 2 polarisations). Default: 1.
squeeze : bool, optional
If `True` (default), writer accepts squeezed arrays as input, and adds any
dimensions of length unity.
file_size : int or None, optional
When writing to a sequence of files, the maximum size of one file in bytes.
If `None` (default), the file size is unlimited, and only the first
file will be written to.
**kwargs
If the header is not given, an attempt will be made to construct one
with any further keyword arguments. See
:class:`~baseband.vdif.base.VDIFStreamWriter`.
Notes
-----
One can also pass to ``name`` a list, tuple, or subclass of
`~baseband.helpers.sequentialfile.FileNameSequencer`. For writing to multiple
files, the ``file_size`` keyword must be passed or only the first file will be
written to. One may also pass in a `~baseband.helpers.sequentialfile` object
(opened in 'rb' mode for reading or 'w+b' for writing), though for typical use
cases it is practically identical to passing in a list or template.
""")