11. Python Reference¶

11.1. Subpackages¶

11.2. Submodules¶

11.3. bifrost.DataType module¶

i: signed integer u: unsigned integer f: floating point ci: complex signed integer cf: complex floating pointer

i4: 4-bit signed integer f16: 16-bit floating point ci4: 4+4-bit complex signed integer cf32: 32+32-bit complex floating point

class bifrost.DataType.DataType(t=None)[source]¶

Bases: object

as_BFdtype()[source]¶

as_complex()[source]¶

as_floating_point()[source]¶: Returns the smallest floating-point type that can represent all values that self can.

as_integer(nbit=None)[source]¶

as_nbit(nbit)[source]¶

as_numpy_dtype()[source]¶

as_real()[source]¶

as_vector(veclen)[source]¶

property is_complex¶

property is_floating_point¶

property is_integer¶

property is_real¶

property is_signed¶

property itemsize¶

property itemsize_bits¶

bifrost.DataType.is_vector_structure(dtype)[source]¶

11.4. bifrost.GPUArray module¶

11.5. bifrost.Space module¶

class bifrost.Space.Space(s)[source]¶

Bases: object

as_BFspace()[source]¶

11.6. bifrost.address module¶

class bifrost.address.Address(address, port, family=None)[source]¶

Bases: bifrost.libbifrost.BifrostObject

property address¶

property family¶

property mtu¶

property port¶

11.7. bifrost.affinity module¶

bifrost.affinity.get_core()[source]¶

bifrost.affinity.set_core(core)[source]¶

bifrost.affinity.set_openmp_cores(cores)[source]¶

11.8. bifrost.block module¶

@package block This file defines a generic block class.

Right now the only possible block type is one of a simple transform which works on a span by span basis.

class bifrost.block.CopyBlock(gulp_size=1048576)[source]¶

Bases: bifrost.block.TransformBlock

Copies input ring’s data to the output ring

main(input_rings, output_rings)[source]¶: Iterate through first ring, copying to each output ring

class bifrost.block.DedisperseBlock(ring, core=- 1, gulp_size=4096)[source]¶

Bases: object

This block calculates the dedispersion of sigproc-formatted data in a ring, and tags it in the headers

dedisperse()[source]¶: Dedisperse on input ring, tagging the output ring. This dedispersion algorithm simply adjusts the start time for each channel in the header.

main()[source]¶: Initiate the block’s processing

class bifrost.block.FFTBlock(gulp_size)[source]¶

Bases: bifrost.block.TransformBlock

Performs complex to complex 1D FFT on input ring data

load_settings(input_header)[source]¶: Load in input header and set up block attributes @param[in] input_header Header sent from input ring

main(input_rings, output_rings)[source]¶

@param[in] input_rings First ring in this list will be used for: data
@param[out] output_rings First ring in this list will be used for: data output.

class bifrost.block.FoldBlock(bins, period=0.001, gulp_size=1048576, dispersion_measure=0, core=- 1)[source]¶

Bases: bifrost.block.TransformBlock

This block folds a signal into a histogram

calculate_bin_indices(tstart, tsamp, data_size)[source]¶

Calculate the bin that each time sample should be: added to

@param[in] tstart Time of the first element (s) @param[in] tsamp Difference between the times of

consecutive elements (s)

@param[in] data_size Number of elements @return Which bin each sample is folded into

calculate_delay(frequency, reference_frequency)[source]¶: Calculate the time delay because of frequency dispersion @param[in] frequency The current channel’s frequency(MHz) @param[in] reference_frequency The frequency of the

channel we will hold at zero time delay(MHz)

load_settings(input_header)[source]¶: Load in input header and set up block attributes @param[in] input_header Header sent from input ring

main(input_rings, output_rings)[source]¶

Generate a histogram from the input ring data @param[in] input_rings List with first ring containing

data of interest. Must terminate before histogram is generated.

@param[out] output_rings First ring in this list: will contain the output histogram

class bifrost.block.IFFTBlock(gulp_size)[source]¶

Bases: bifrost.block.TransformBlock

Performs complex to complex 1D IFFT on input ring data

load_settings(input_header)[source]¶: Load in input header and set up block attributes @param[in] input_header Header sent from input ring

main(input_rings, output_rings)[source]¶

@param[in] input_rings First ring in this list will be used for: data input.
@param[out] output_rings First ring in this list will be used for: data output.

class bifrost.block.KurtosisBlock(gulp_size=1048576, core=- 1)[source]¶

Bases: bifrost.block.TransformBlock

This block performs spectral kurtosis and cleaning on sigproc-formatted data in rings

load_settings(input_header)[source]¶: Load in input header and set up block attributes @param[in] input_header Header sent from input ring

main(input_rings, output_rings)[source]¶: Calls a kurtosis algorithm and uses the result to clean the input data of RFI, and move it to the output ring.

class bifrost.block.MultiAddBlock(*args, **kwargs)[source]¶

Bases: bifrost.block.MultiTransformBlock

Block which adds two input rings

main()[source]¶: Iterate through the inputs, and add them to the output

ring_names = {'in_1': 'First input to add. List of floats', 'in_2': 'Second input to add. List of floats', 'out_sum': 'Result of add. List of floats.'}¶

class bifrost.block.MultiTransformBlock[source]¶

Bases: object

Defines functions and attributes for a block with multi input/output

flatten(*args)[source]¶: Flatten a nested tuple/list of tuples/lists

izip(*iterables)[source]¶: Iterate through multpile iterators This differs from itertools in that this izip combines list generators into a single list generator

load_settings()[source]¶: Set by user to interpret input rings

read(*args)[source]¶: Iterate over selection of input rings

write(*args)[source]¶: Iterate over selection of output rings

class bifrost.block.NumpyBlock(function, inputs=1, outputs=1)[source]¶

Bases: bifrost.block.MultiTransformBlock

Perform an arbitrary N ndarray -> M ndarray numpy function Inside of a pipeline. This block will calculate all of the necessary information for Bifrost based on the passed function.

calculate_output_headers(out_arrays)[source]¶: Generate headers based on numpy arrays @param[in] out_arrays The arrays to measure

create_ring_names()[source]¶: Generate dummy ring descriptions

did_header_change(old_header)[source]¶: See if the new headers are different @param[in] old_header The previous headers

load_settings()[source]¶: Generate empty arrays based on input headers.

main()[source]¶: Call self.function on all of the input spans

reshape_inspans(inspans)[source]¶: Fit the input spans to their headers @param[in] inspans The input spans.

class bifrost.block.NumpySourceBlock(generator, outputs=1, grab_headers=False, changing=True)[source]¶

Bases: bifrost.block.MultiTransformBlock

Simulate an incoming stream of data on a ring using an arbitrary generator. This block will calculate all of the necessary information for Bifrost based on the passed function.

calculate_output_settings(arrays)[source]¶: Calculate the outgoing header settings based on the output arrays @param[in] arrays The arrays outputted by self.generator

load_user_headers(headers, arrays)[source]¶: Load in user defined headers @param[in] headers List of dictionaries from self.generator

for each ring’s sequence header

main()[source]¶: Call self.generator and output the arrays into the output

class bifrost.block.Pipeline(blocks)[source]¶

Bases: object

Class which connects blocks linearly, with one ring between each block. Does this by creating one ring for each input/output ‘port’ of each block, and running data through the rings.

main()[source]¶: Start the pipeline, and finish when all threads exit

unique_ring_names()[source]¶: Return a list of unique ring indices

class bifrost.block.SigprocReadBlock(filename, gulp_nframe=4096, core=- 1)[source]¶

Bases: bifrost.block.SourceBlock

This block reads in a sigproc filterbank (.fil) file into a ring buffer

main(output_ring)[source]¶: Read in the sigproc file to output_ring @param[in] output_ring Ring to write to

class bifrost.block.SinkBlock(gulp_size=4096)[source]¶

Bases: object

Defines the structure for a sink block

iterate_ring_read(input_ring)[source]¶: Iterate through one input ring @param[in] input_ring Ring to read through

load_settings(input_header)[source]¶: Load in settings from input ring header

class bifrost.block.SourceBlock(gulp_size=4096)[source]¶

Bases: object

Defines the structure for a source block

iterate_ring_write(output_ring, sequence_name='', sequence_time_tag=0)[source]¶: Iterate over output ring @param[in] output_ring Ring to write to @param[in] sequence_name Name to label sequence @param[in] sequence_time_tag Time tag to label sequence

class bifrost.block.SplitterBlock(sections)[source]¶

Bases: bifrost.block.MultiTransformBlock

Block which splits up a ring into two

load_settings()[source]¶: Set the gulp sizes appropriate to the input ring

main()[source]¶: Split the incoming ring into the outputs rings

ring_names = {'in': 'Input to split. List of floats', 'out_1': 'Gets first share of the ring. List of floats', 'out_2': 'Gets second share of the ring. List of floats'}¶

class bifrost.block.TestingBlock(test_array, complex_numbers=False)[source]¶

Bases: bifrost.block.SourceBlock

Block for debugging purposes. Allows you to pass arbitrary N-dimensional arrays in initialization, which will be outputted into a ring buffer

main(output_ring)[source]¶: Put the test array onto the output ring @param[in] output_ring Holds the flattend test array in a single span

class bifrost.block.TransformBlock(gulp_size=4096)[source]¶

Bases: object

Defines the structure for a transform block

iterate_ring_read(input_ring)[source]¶: Iterate through one input ring

iterate_ring_write(output_ring, sequence_name='', sequence_time_tag=0, sequence_nringlet=1)[source]¶: Iterate through one output ring

load_settings(input_header)[source]¶: Load in input header and set up block attributes @param[in] input_header Header sent from input ring

ring_transfer(input_ring, output_ring)[source]¶: Iterate through two rings span-by-span

class bifrost.block.WaterfallBlock(ring, imagename, core=- 1, gulp_nframe=4096)[source]¶

Bases: object

This block creates a waterfall block based on the data in a ring, and stores it in the headers

generate_waterfall_matrix()[source]¶: Create a matrix for a waterfall image based on the ring’s data

main()[source]¶: Initiate the block’s processing

save_waterfall_plot(waterfall_matrix)[source]¶

Save an image of the waterfall plot using: thread-safe backend for pyplot, and labelling the plot using the header information from the ring
@param[in] waterfall_matrix x axis is frequency and: y axis is time. Values should be power.

class bifrost.block.WriteAsciiBlock(filename, gulp_size=1048576)[source]¶

Bases: bifrost.block.SinkBlock

Copies input ring’s data into ascii format in a text file.

load_settings(input_header)[source]¶: Load in settings from input ring header

main(input_ring)[source]¶

Initiate the writing to filename @param[in] input_rings First ring in this list will be used for

data

@param[out] output_rings This list of rings won’t be used.

class bifrost.block.WriteHeaderBlock(filename)[source]¶

Bases: bifrost.block.SinkBlock

Prints the header of a ring to a file

load_settings(input_header)[source]¶: Load the header from json @param[in] input_header The header from the ring

main(input_ring)[source]¶: Put the header into the file @param[in] input_ring Contains the header in question

bifrost.block.insert_zeros_evenly(input_data, number_zeros)[source]¶

Insert zeros evenly in input_data.: These zeros are distibuted evenly throughout the function, to help for binning of oddly shaped arrays.

@param[in] input_data 1D array to contain zeros. @param[out] number_zeros Number of zeros that need

to be added.

@returns input_data with extra zeros

11.9. bifrost.block_chainer module¶

class bifrost.block_chainer.BlockChainer[source]¶

Bases: object

Convenient tool for constructing linear chains of blocks and views

Examples:

bc = bf.BlockChainer()
bc.blocks.read_sigproc("foo.fil", gulp_nframe=1)
bc.blocks.copy('cuda')
bc.views.split_axis('freq', 2, 'fine_freq')
bc.views.merge_axes('freq', 'fine_freq')
bc.blocks.copy('cuda_host')
bc.custom(my_block)(arg1, arg2, ...)
bc.blocks.write_sigproc()
print(bc.last_block) # The last added block (this can also be set)

property blocks¶

custom(func)[source]¶

print_header(*args, **kwargs)[source]¶

property views¶

11.10. bifrost.core module¶

bifrost.core.cuda_enabled()[source]¶

bifrost.core.debug_enabled()[source]¶

bifrost.core.status_string(status)[source]¶

11.11. bifrost.device module¶

bifrost.device.get_device()[source]¶

bifrost.device.set_device(device)[source]¶

bifrost.device.set_devices_no_spin_cpu()[source]¶

Sets a flag on all GPU devices that tells them not to spin the CPU when synchronizing. This is useful for reducing CPU load in GPU pipelines.

This function must be called _before_ any GPU devices are initialized (i.e., at the start of the process).

bifrost.device.stream_synchronize()[source]¶

11.12. bifrost.dtype module¶

i: signed integer u: unsigned integer f: floating point ci: complex signed integer cu: complex unsigned integer cf: complex floating pointer

i4: 4-bit signed integer f16: 16-bit floating point ci4: 4+4-bit complex signed integer cf32: 32+32-bit complex floating point

bifrost.dtype.from_complex64(f, dtype)[source]¶

bifrost.dtype.name_nbit2numpy(name, nbit)[source]¶

bifrost.dtype.numpy2bifrost(dtype)[source]¶

bifrost.dtype.numpy2string(dtype)[source]¶

bifrost.dtype.split_name_nbit(dtype_str)[source]¶: Splits a dtype string into (name, nbit)

bifrost.dtype.string2numpy(dtype_str)[source]¶

bifrost.dtype.to_complex64(q)[source]¶

11.13. bifrost.fdmt module¶

class bifrost.fdmt.Fdmt[source]¶

Bases: bifrost.libbifrost.BifrostObject

execute(idata, odata, negative_delays=False)[source]¶

execute_workspace(idata, odata, workspace_ptr, workspace_size, negative_delays=False)[source]¶

get_workspace_size(idata, odata)[source]¶

init(nchan, max_delay, f0, df, exponent=- 2.0, space='cuda')[source]¶

11.14. bifrost.fft module¶

class bifrost.fft.Fft[source]¶

Bases: bifrost.libbifrost.BifrostObject

execute(iarray, oarray, inverse=False)[source]¶

execute_workspace(iarray, oarray, workspace_ptr, workspace_size, inverse=False)[source]¶

init(iarray, oarray, axes=None, apply_fftshift=False)[source]¶

11.15. bifrost.fir module¶

class bifrost.fir.Fir[source]¶

Bases: bifrost.libbifrost.BifrostObject

execute(idata, odata)[source]¶

init(coeffs, decim=1, space='cuda')[source]¶

reset_state()[source]¶

set_coeffs(coeffs)[source]¶

11.16. bifrost.guppi_raw module¶

GUPPI Raw format Headers:

Records are 80 chars, padded with spaces Keywords are truncated/padded with spaces to 8 chars “keyword1= <value>” String values are enclosed in single-quotes Numerical values typically not enclosed in quotes, but sometimes may be Final record is always ‘END’ + ‘ ‘*77

Header keywords:

DIRECTIO: If present and non-zero, headers are padded to a 512-byte boundary NBITS: No. bits per real component (e.g., 4 => 4+4-bit complex values);

typical values: 8,4,2 (particularly 8)

BLOCSIZE: No. bytes per binary data block OBSNCHAN (or NCHAN?) NPOL: Single-pol if 1 else dual-pol OBSFREQ: Centre freq of data OBSBW: Bandwidth of data (may be negative to indicate high->low channel

ordering)

BACKEND: ‘GUPPI’ for guppi/BL data [CHAN_BW]

NTIME = BLOCSIZE * 8 // (2 * NPOL * NCHAN * NBITS)

Binary data:: [chan][time][pol][complex]

bifrost.guppi_raw.read_header(f)[source]¶

11.17. bifrost.header_standard module¶

@package header_standard This file enforces a standard header for rings.

Required parameters:

(parameter type definition) nchans int “Number of frequency channels. 1+” nifs int “Number of separate IF channels. 1+” nbits int “Number of bits per value. 1+” fch1 float “Center frequency of first channel given in buffer (MHz). >0” foff float “Bandwidth of each channel (MHz). Negative values used for when

first channel specified has the largest frequency.”

tstart float “Time stamp in MJD of first sample (seconds). >0” tsamp float “Time interval between samples (seconds). >0”

Optional parameters (which some blocks require):

bifrost.header_standard.enforce_header_standard(header_dict)[source]¶: Raise an error if the header dictionary passed does not fit the standard specified above.

11.18. bifrost.libbifrost module¶

class bifrost.libbifrost.BifrostObject(constructor, destructor, *args)[source]¶

Bases: object

Base class for simple objects with create/destroy functions

exception bifrost.libbifrost.EndOfDataStop[source]¶

Bases: RuntimeError

This class is used as a Py3 StopIterator

In Python >3.7, reaching a StopIterator in a generator will raise a RuntimeError (so you can’t do ‘except StopIteration’ to catch it!) See PEP479 https://www.python.org/dev/peps/pep-0479/

11.19. bifrost.linalg module¶

class bifrost.linalg.LinAlg[source]¶

Bases: bifrost.libbifrost.BifrostObject

matmul(alpha, a, b, beta, c)[source]¶

Computes:: c = alpha*a.b + beta*c
or if b is None:: c = alpha*a.a^H + beta*c
or if a is None:: c = alpha*b^H.b + beta*c

where ‘.’ is matrix product and ‘^H’ is Hermitian transpose. Multi-dimensional semantics are the same as numpy.matmul:

The last two dims represent the matrix, and all other dims are used as batch dims to be matched or broadcast between a and b.

11.20. bifrost.map module¶

bifrost.map.clear_map_cache()[source]¶

bifrost.map.list_map_cache()[source]¶

bifrost.map.map(func_string, data, axis_names=None, shape=None, func_name=None, extra_code=None, block_shape=None, block_axes=None)[source]¶

Apply a function to a set of ndarrays.

Parameters

func_string (str) – The function to apply to the arrays, as a string (see below for examples).
data (dict) – Map of string names to ndarrays or scalars.
axis_names (list) – List of string names by which each axis is referenced in func_string.
shape – The shape of the computation. If None, the broadcast shape of all data arrays is used.
func_name (str) – Name of the function, for debugging purposes.
extra_code (str) – Additional code to be included at global scope.
block_shape – The 2D shape of the thread block (y,x) with which the kernel is launched. This is a performance tuning parameter. If NULL, a heuristic is used to select the block shape. Changes to this parameter do _not_ require re-compilation of the kernel.
block_axes – List of axis indices (or names) specifying the 2 computation axes to which the thread block (y,x) is mapped. This is a performance tuning parameter. If NULL, a heuristic is used to select the block axes. Values may be negative for reverse indexing. Changes to this parameter _do_ require re-compilation of the kernel.

Note

Only GPU computation is currently supported.

Examples:

# Add two arrays together
bf.map("c = a + b", {'c': c, 'a': a, 'b': b})

# Compute outer product of two arrays
bf.map("c(i,j) = a(i) * b(j)",
       {'c': c, 'a': a, 'b': b},
       axis_names=('i','j'))

# Split the components of a complex array
bf.map("a = c.real; b = c.imag", {'c': c, 'a': a, 'b': b})

# Raise an array to a scalar power
bf.map("c = pow(a, p)", {'c': c, 'a': a, 'p': 2.0})

# Slice an array with a scalar index
bf.map("c(i) = a(i,k)", {'c': c, 'a': a, 'k': 7}, ['i'], shape=c.shape)

11.21. bifrost.memory module¶

bifrost.memory.alignment()[source]¶

bifrost.memory.memcpy(dst, src)[source]¶

bifrost.memory.memcpy2D(dst, src)[source]¶

bifrost.memory.memset(dst, val=0)[source]¶

bifrost.memory.memset2D(dst, val=0)[source]¶

bifrost.memory.raw_free(ptr, space='auto')[source]¶

bifrost.memory.raw_get_space(ptr)[source]¶

bifrost.memory.raw_malloc(size, space)[source]¶

bifrost.memory.space_accessible(space, from_spaces)[source]¶

11.22. bifrost.ndarray module¶

A np.ndarray subclass that adds support for different spaces and: bifrost-specific metadata.

TODO: Need backend support for broadcasting arrays TODO: Convert dtype of val in __setitem__ TODO: Some calls result in segfault with space=cuda (e.g., __getitem__

returning scalar).

class bifrost.ndarray.BFArrayInfo(space, dtype, native, conjugated, ownbuffer=None)[source]¶: Bases: object

bifrost.ndarray.asarray(arr, space=None)[source]¶

bifrost.ndarray.copy_array(dst, src)[source]¶

bifrost.ndarray.empty(shape, dtype='f32', space=None, **kwargs)[source]¶

bifrost.ndarray.empty_like(arr, space=None)[source]¶

bifrost.ndarray.memset_array(dst, value)[source]¶

class bifrost.ndarray.ndarray(base=None, space=None, shape=None, dtype=None, buffer=None, offset=0, strides=None, native=None, conjugated=None)[source]¶

Bases: numpy.ndarray

as_BFarray()[source]¶

as_GPUArray(*args, **kwargs)[source]¶

as_cupy(*args, **kwargs)[source]¶

byteswap(inplace=False)[source]¶

Swap the bytes of the array elements

Toggle between low-endian and big-endian data representation by returning a byteswapped array, optionally swapped in-place. Arrays of byte-strings are not swapped. The real and imaginary parts of a complex number are swapped individually.

Parameters: inplace (bool, optional) – If True, swap bytes in-place, default is False.
Returns: out – The byteswapped array. If inplace is True, this is a view to self.
Return type: ndarray

Examples

>>> A = np.array([1, 256, 8755], dtype=np.int16)
>>> list(map(hex, A))
['0x1', '0x100', '0x2233']
>>> A.byteswap(inplace=True)
array([  256,     1, 13090], dtype=int16)
>>> list(map(hex, A))
['0x100', '0x1', '0x3322']

Arrays of byte-strings are not swapped

>>> A = np.array([b'ceg', b'fac'])
>>> A.byteswap()
array([b'ceg', b'fac'], dtype='|S3')

A.newbyteorder().byteswap() produces an array with the same values: but different representation in memory

>>> A = np.array([1, 2, 3])
>>> A.view(np.uint8)
array([1, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 3, 0, 0, 0, 0, 0,
       0, 0], dtype=uint8)
>>> A.newbyteorder().byteswap(inplace=True)
array([1, 2, 3])
>>> A.view(np.uint8)
array([0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0,
       0, 3], dtype=uint8)

conj()[source]¶

Complex-conjugate all elements.

Refer to numpy.conjugate for full documentation.

11.23. bifrost.pipeline module¶

class bifrost.pipeline.Block(irings, name=None, type_=None, **kwargs)[source]¶

Bases: bifrost.pipeline.BlockScope

begin_sequences(exit_stack, orings, oheaders, igulp_nframes, istride_nframes)[source]¶

begin_writing(exit_stack, orings)[source]¶

commit_spans(ospans, ostrides_actual, ogulp_overlaps)[source]¶

create_ring(*args, **kwargs)[source]¶

define_output_nframes(input_nframes)[source]¶: Return output nframe for each output, given input_nframes.

define_valid_input_spaces()[source]¶: Return set of valid spaces (or ‘any’) for each input

instance_counts = {}¶

num_outputs()[source]¶

reserve_spans(exit_stack, oseqs, igulp_nframes=[])[source]¶

run()[source]¶

shutdown()[source]¶

class bifrost.pipeline.BlockScope(name=None, gulp_nframe=None, buffer_nframe=None, buffer_factor=None, core=None, gpu=None, share_temp_storage=False, fuse=False)[source]¶

Bases: object

cache_scope_hierarchy()[source]¶

dot_graph(parent_graph=None)[source]¶

get_temp_storage(space)[source]¶

instance_count = 0¶

is_fused_with(other)[source]¶

class bifrost.pipeline.MultiTransformBlock(irings_, guarantee=True, *args, **kwargs)[source]¶

Bases: bifrost.pipeline.Block

define_input_overlap_nframe(iseqs)[source]¶: Return no. input frames that should overlap between successive spans for each input sequence.

define_output_nframes(input_nframes)[source]¶: Return output nframe for each output, given input_nframes.

main(orings)[source]¶

on_data(ispans, ospans)[source]¶: Process data from from ispans to ospans and return the number of frames to commit for each output (or None to commit complete spans).

on_sequence(iseqs)[source]¶: Return: oheaders (one per output)

on_sequence_end(iseqs)[source]¶: Do any necessary cleanup

on_skip(islices, ospans)[source]¶: Handle skipped frames

class bifrost.pipeline.Pipeline(name=None, **kwargs)[source]¶

Bases: bifrost.pipeline.BlockScope

as_default()[source]¶

instance_count = 1¶

run()[source]¶

shutdown()[source]¶

shutdown_on_signals(signals=None)[source]¶

synchronize_block_initializations()[source]¶

exception bifrost.pipeline.PipelineInitError[source]¶: Bases: Exception

class bifrost.pipeline.SinkBlock(iring, *args, **kwargs)[source]¶

Bases: bifrost.pipeline.MultiTransformBlock

define_input_overlap_nframe(iseq)[source]¶: Return no. input frames that should overlap between successive spans.

define_valid_input_spaces()[source]¶: Return set of valid spaces (or ‘any’) for the input

on_data(ispan)[source]¶: Return nothing

on_sequence(iseq)[source]¶: Return islice or None to use simple striding

on_sequence_end(iseq)[source]¶: Do any necessary cleanup

class bifrost.pipeline.SourceBlock(sourcenames, gulp_nframe, space=None, *args, **kwargs)[source]¶

Bases: bifrost.pipeline.Block

create_reader(sourcename)[source]¶: Return an object to use for reading source data

define_output_nframes(_)[source]¶: Return output nframe for each output, given input_nframes.

define_valid_input_spaces()[source]¶: Return set of valid spaces (or ‘any’) for each input

main(orings)[source]¶

on_data(reader, ospans)[source]¶: Process data from from ispans to ospans and return the number of frames to commit for each output.

on_sequence(reader, sourcename)[source]¶: Return header for each output

class bifrost.pipeline.TransformBlock(iring, *args, **kwargs)[source]¶

Bases: bifrost.pipeline.MultiTransformBlock

define_input_overlap_nframe(iseq)[source]¶: Return no. input frames that should overlap between successive spans.

define_output_nframes(input_nframe)[source]¶: Return number of frames that will be produced given input_nframe

define_valid_input_spaces()[source]¶: Return set of valid spaces (or ‘any’) for the input

on_data(ispan, ospan)[source]¶: Return the number of output frames to commit, or None to commit all

on_sequence(iseq)[source]¶: Return oheader

on_sequence_end(iseq)[source]¶: Do any necessary cleanup

on_skip(islice, ospan)[source]¶: Handle skipped frames

bifrost.pipeline.block_scope(*args, **kwargs)[source]¶

bifrost.pipeline.block_view(block, header_transform)[source]¶

View a block with modified output headers

Use this function to adjust the output headers of a ring on-the-fly, effectively producing a new ‘view’ of the block.

Parameters

block (Block) – Input block.
header_transform (function) – A function f(hdr) -> new_hdr.

Returns

A new block that acts as the old block but modifies its sequence headers on-the-fly.

bifrost.pipeline.get_current_block_scope()[source]¶

bifrost.pipeline.get_default_pipeline()[source]¶

bifrost.pipeline.get_ring(block_or_ring)[source]¶

bifrost.pipeline.izip(*iterables)[source]¶

bifrost.pipeline.join_all(threads, timeout)[source]¶

bifrost.pipeline.try_join(thread, timeout=0.0)[source]¶

11.24. bifrost.portaudio module¶

11.25. bifrost.proclog module¶

class bifrost.proclog.ProcLog(name)[source]¶

Bases: bifrost.libbifrost.BifrostObject

update(contents)[source]¶: Updates (replaces) the contents of the log contents: string or dict containing data to write to the log

bifrost.proclog.load_by_filename(filename)[source]¶: Function to read in a ProcLog file and return the contents as a dictionary.

bifrost.proclog.load_by_pid(pid, include_rings=False)[source]¶

Function to read in and parse all ProcLog files associated with a given process ID. The contents of these files are returned as a collection of dictionaries ordered by:

block name

ProcLog name
entry name

11.26. bifrost.psrdada module¶

11.27. bifrost.quantize module¶

bifrost.quantize.quantize(src, dst, scale=1.0)[source]¶

11.28. bifrost.reduce module¶

bifrost.reduce.reduce(idata, odata, op='sum')[source]¶

11.29. bifrost.ring module¶

class bifrost.ring.ReadSequence(ring, which='specific', name='', time_tag=None, other_obj=None, guarantee=True)[source]¶

Bases: bifrost.ring.SequenceBase

acquire(offset, size)[source]¶

close()[source]¶

increment()[source]¶

read(span_size, stride=None, begin=0)[source]¶

class bifrost.ring.ReadSpan(sequence, offset, size)[source]¶

Bases: bifrost.ring.SpanBase

release()[source]¶

class bifrost.ring.Ring(space='system', name=None, core=None)[source]¶

Bases: bifrost.libbifrost.BifrostObject

begin_writing()[source]¶

property core¶

end_writing()[source]¶

property name¶

open_earliest_sequence(guarantee=True)[source]¶

open_latest_sequence(guarantee=True)[source]¶

open_sequence(name, guarantee=True)[source]¶

open_sequence_at(time_tag, guarantee=True)[source]¶

read(whence='earliest', guarantee=True)[source]¶

resize(contiguous_span, total_span=None, nringlet=1, buffer_factor=4)[source]¶

property space¶

writing_ended()[source]¶

class bifrost.ring.RingWriter(ring)[source]¶

Bases: object

begin_sequence(name='', time_tag=- 1, header='', nringlet=1)[source]¶

class bifrost.ring.SequenceBase(ring)[source]¶

Bases: object

Python object for a ring’s sequence (data unit)

property header¶

property header_size¶

property name¶

property nringlet¶

property ring¶

property time_tag¶

class bifrost.ring.SpanBase(ring, writeable)[source]¶

Bases: object

property data¶

data_view(dtype=<class 'numpy.uint8'>, shape=-1)[source]¶

property nringlet¶

property offset¶

property ring¶

property size¶

property stride¶

class bifrost.ring.WriteSequence(ring, name='', time_tag=- 1, header='', nringlet=1)[source]¶

Bases: bifrost.ring.SequenceBase

end()[source]¶

reserve(size, nonblocking=False)[source]¶

class bifrost.ring.WriteSpan(ring, size, nonblocking=False)[source]¶

Bases: bifrost.ring.SpanBase

close()[source]¶

commit(size)[source]¶

11.30. bifrost.ring2 module¶

class bifrost.ring2.ReadSequence(ring, which='specific', name='', time_tag=None, other_obj=None, guarantee=True, header_transform=None)[source]¶

Bases: bifrost.ring2.SequenceBase

acquire(frame_offset, nframe)[source]¶

close()[source]¶

property header¶

increment()[source]¶

read(nframe, stride=None, begin=0)[source]¶

resize(gulp_nframe, buf_nframe=None, buffer_factor=None)[source]¶

class bifrost.ring2.ReadSpan(sequence, frame_offset, nframe)[source]¶

Bases: bifrost.ring2.SpanBase

property nframe_overwritten¶

release()[source]¶

class bifrost.ring2.Ring(space='system', name=None, owner=None, core=None)[source]¶

Bases: bifrost.libbifrost.BifrostObject

begin_writing()[source]¶

property core¶

end_writing()[source]¶

instance_count = 0¶

property name¶

open_earliest_sequence(guarantee=True)[source]¶

open_latest_sequence(guarantee=True)[source]¶

open_sequence(name, guarantee=True)[source]¶

open_sequence_at(time_tag, guarantee=True)[source]¶

read(whence='earliest', guarantee=True)[source]¶

resize(contiguous_bytes, total_bytes=None, nringlet=1)[source]¶

view()[source]¶

class bifrost.ring2.RingWriter(ring)[source]¶

Bases: object

begin_sequence(header, gulp_nframe, buf_nframe)[source]¶

class bifrost.ring2.SequenceBase(ring)[source]¶

Bases: object

Python object for a ring’s sequence (data unit)

property header¶

property header_size¶

property name¶

property nringlet¶

property ring¶

property tensor¶

property time_tag¶

class bifrost.ring2.SpanBase(ring, sequence, writeable)[source]¶

Bases: object

property data¶

property dtype¶

property frame_nbyte¶

property frame_offset¶

property nframe¶

property ring¶

property sequence¶

property shape¶

property strides¶

property tensor¶

class bifrost.ring2.WriteSequence(ring, header, gulp_nframe, buf_nframe)[source]¶

Bases: bifrost.ring2.SequenceBase

end()[source]¶

reserve(nframe, nonblocking=False)[source]¶

class bifrost.ring2.WriteSpan(ring, sequence, nframe, nonblocking=False)[source]¶

Bases: bifrost.ring2.SpanBase

close()[source]¶

commit(nframe)[source]¶

bifrost.ring2.accumulate(vals, op='+', init=None, reverse=False)[source]¶

bifrost.ring2.compose_unary_funcs(f, g)[source]¶

bifrost.ring2.ring_view(ring, header_transform)[source]¶

bifrost.ring2.split_shape(shape)[source]¶: Splits a shape into its ringlet shape and frame shape E.g., (2,3,-1,4,5) -> (2,3), (4,5)

11.31. bifrost.romein module¶

class bifrost.romein.Romein[source]¶

Bases: bifrost.libbifrost.BifrostObject

execute(idata, odata)[source]¶

init(positions, kernels, ngrid, polmajor=True)[source]¶

set_kernels(kernels)[source]¶

set_positions(positions)[source]¶

11.32. bifrost.sigproc module¶

telescope_id: 0 (FAKE) machine_id: 0 (FAKE) data_type: 2 # Time-series data rawdatafile: <delete> source_name: <observer-specified> barycentric: 0 pulsarcentric: <delete> az_start: <delete> or <observer-specified> za_start: <delete> or <observer-specified> src_raj: <observer-specified> or <delete> src_dej: <observer-specified> or <delete> tstart: MJD of first sample tsamp: (secs) E.g., 0.5/(2400*24kHz=57.6MHz)=8.68055556ns nbits: 8 nsamples: No. time samples in file (“rarely used any more”) fch1: 58.776 MHz (center frequency) foff: MHz nchans: 1 nifs: 2 (pols) refdm: 0.0 [pc/cm^3] period: <delete> data: [time][pol][nbit] (General case: [time][if/pol][chan][nbit])

class bifrost.sigproc.SigprocFile[source]¶

Bases: bifrost.sigproc.SigprocSettings

Reads from or writes to a sigproc filterbank file

append_data(input_data)[source]¶: append data to local data and file

clear()[source]¶: Erases file contents

close()[source]¶: closes file object

get_nframe()[source]¶: calculate the number of frames from the data

open(filename, mode)[source]¶: open the filename, and read the header and data from it

read_data(start=None, end=None)[source]¶: read data from file and store it locally

read_header()[source]¶: reads in a header from the file and sets local settings

write_to(filename)[source]¶: writes data and header to a different file

class bifrost.sigproc.SigprocSettings[source]¶

Bases: object

defines, reads, writes sigproc settings

interpret_header()[source]¶: redefine variables from header dictionary

bifrost.sigproc.pack(data, nbit)[source]¶: downgrade data from 8bits to nbits (per value)

bifrost.sigproc.seek_to_data(file_object)[source]¶: Go the the location in the file where the data begins

bifrost.sigproc.unpack(data, nbit)[source]¶: upgrade data from nbits to 8bits

11.33. bifrost.sigproc2 module¶

telescope_id: 0 (FAKE) machine_id: 0 (FAKE) data_type: 2 # Time-series data rawdatafile: <delete> source_name: <observer-specified> barycentric: 0 pulsarcentric: <delete> az_start: <delete> or <observer-specified> za_start: <delete> or <observer-specified> src_raj: <observer-specified> or <delete> src_dej: <observer-specified> or <delete> tstart: MJD of first sample tsamp: (secs) E.g., 0.5/(2400*24kHz=57.6MHz)=8.68055556ns nbits: 8 nsamples: No. time samples in file (“rarely used any more”) fch1: 58.776 MHz (center frequency) foff: MHz nchans: 1 nifs: 2 (pols) refdm: 0.0 [pc/cm^3] period: <delete> data: [time][pol][nbit] (General case: [time][if/pol][chan][nbit])

class bifrost.sigproc2.SigprocFile(filename=None)[source]¶

Bases: object

bandwidth()[source]¶

cfreq()[source]¶

close()[source]¶

duration()[source]¶

nframe()[source]¶

open(filename)[source]¶

read(nframe_or_start, end=None)[source]¶

readinto(buf)[source]¶: Fills buf with raw bytes straight from the file

seek(offset, whence=0)[source]¶

bifrost.sigproc2.id2machine(id_)[source]¶

bifrost.sigproc2.id2telescope(id_)[source]¶

bifrost.sigproc2.machine2id(name)[source]¶

bifrost.sigproc2.telescope2id(name)[source]¶

bifrost.sigproc2.unpack(data, nbit)[source]¶

bifrost.sigproc2.write_header(hdr, f)[source]¶

11.34. bifrost.temp_storage module¶

class bifrost.temp_storage.TempStorage(space)[source]¶

Bases: object

allocate(size)[source]¶

class bifrost.temp_storage.TempStorageAllocation(parent, size)[source]¶

Bases: object

release()[source]¶

11.35. bifrost.transpose module¶

bifrost.transpose.transpose(dst, src, axes=None)[source]¶

11.36. bifrost.udp_capture module¶

class bifrost.udp_capture.UDPCapture(fmt, sock, ring, nsrc, src0, max_payload_size, buffer_ntime, slot_ntime, sequence_callback, core=None)[source]¶