TimeSeries¶

class gwpy.timeseries.TimeSeries(data, unit=
None
, t0=None
, dt=None
, sample_rate=None
, times=None
, channel=None
, name=None
, **kwargs)[source]¶ A timedomain data array.
 Parameters:¶
 valuearraylike
input data array
 unit
Unit
, optional physical unit of these data
 t0
LIGOTimeGPS
,float
,str
, optional GPS epoch associated with these data, any input parsable by
to_gps
is fine dt
float
,Quantity
, optional time between successive samples (seconds), can also be given inversely via
sample_rate
 sample_rate
float
,Quantity
, optional the rate of samples per second (Hertz), can also be given inversely via
dt
 times
arraylike
the complete array of GPS times accompanying the data for this series. This argument takes precedence over
t0
anddt
so should be given in place of these if relevant, not alongside name
str
, optional descriptive title for this array
 channel
Channel
,str
, optional source data stream for these data
 dtype
dtype
, optional input data type
 copy
bool
, optional choose to copy the input data to new memory
 subok
bool
, optional allow passing of subclasses by the array generator
Notes
The necessary metadata to reconstruct timing information are recorded in the
epoch
andsample_rate
attributes. This timestamps can be returned via thetimes
property.All comparison operations performed on a
TimeSeries
will return aStateTimeSeries
 a boolean array with metadata copied from the startingTimeSeries
.Examples
>>> from gwpy.timeseries import TimeSeries
To create an array of random numbers, sampled at 100 Hz, in units of ‘metres’:
>>> from numpy import random >>> series = TimeSeries(random.random(1000), sample_rate=100, unit='m')
which can then be simply visualised via
>>> plot = series.plot() >>> plot.show()
(
png
)Attributes Summary
View of the transposed array.
Base object if memory is from some other object.
Returns a copy of the current
Quantity
instance with CGS units.Instrumental channel associated with these data
An object to simplify the interaction of the array with the ctypes module.
Python buffer object pointing to the start of the array's data.
Xaxis sample separation
Datatype of the array's elements.
Duration of this series in seconds
Xaxis sample separation
GPS epoch for these data.
A list of equivalencies that will be applied by default during unit conversions.
Information about the memory layout of the array.
A 1D iterator over the Quantity array.
The imaginary part of the array.
Container for meta information like name, description, format.
True if the
value
of this quantity is a scalar, or False if it is an arraylike object.Length of one array element in bytes.
View of the matrix transposed array.
Name for this data set
Total bytes consumed by the elements of the array.
Number of array dimensions.
The real part of the array.
Data rate for this
TimeSeries
in samples per second (Hertz).Tuple of array dimensions.
Returns a copy of the current
Quantity
instance with SI units.Number of elements in the array.
Xaxis [low, high) segment encompassed by these data
Tuple of bytes to step in each dimension when traversing an array.
Xaxis coordinate of the first data point
Positions of the data on the xaxis
The physical unit of these data
The numerical value of this instance.
Xaxis coordinate of the first data point
Positions of the data on the xaxis
Xaxis [low, high) segment encompassed by these data
Unit of xaxis index
Methods Summary
abs
(x, /[, out, where, casting, order, ...])Calculate the absolute value elementwise.
all
([axis, out, keepdims, where])Returns True if all elements evaluate to True.
any
([axis, out, keepdims, where])Returns True if any of the elements of
a
evaluate to True.append
(other[, inplace, pad, gap, resize])Connect another series onto the end of the current one.
argmax
([axis, out, keepdims])Return indices of the maximum values along the given axis.
argmin
([axis, out, keepdims])Return indices of the minimum values along the given axis.
argpartition
(kth[, axis, kind, order])Returns the indices that would partition this array.
argsort
([axis, kind, order])Returns the indices that would sort this array.
asd
([fftlength, overlap, window, method])Calculate the ASD
FrequencySeries
of thisTimeSeries
astype
(dtype[, order, casting, subok, copy])Copy of the array, cast to a specified type.
auto_coherence
(dt[, fftlength, overlap, window])Calculate the frequencycoherence between this
TimeSeries
and a timeshifted copy of itself.average_fft
([fftlength, overlap, window])Compute the averaged onedimensional DFT of this
TimeSeries
.bandpass
(flow, fhigh[, gpass, gstop, fstop, ...])Filter this
TimeSeries
with a bandpass filter.byteswap
([inplace])Swap the bytes of the array elements
choose
(choices[, out, mode])Use an index array to construct a new array from a set of choices.
clip
([min, max, out])Return an array whose values are limited to
[min, max]
.coherence
(other[, fftlength, overlap, window])Calculate the frequencycoherence between this
TimeSeries
and another.coherence_spectrogram
(other, stride[, ...])Calculate the coherence spectrogram between this
TimeSeries
and other.compress
(condition[, axis, out])Return selected slices of this array along given axis.
conj
()Complexconjugate all elements.
Return the complex conjugate, elementwise.
convolve
(fir[, window])Convolve this
TimeSeries
with an FIR filter using thecopy
([order])Return a copy of the array.
correlate
(mfilter[, window, detrend, ...])Crosscorrelate this
TimeSeries
with another signalcrop
([start, end, copy])Crop this series to the given xaxis extent.
csd
(other[, fftlength, overlap, window])Calculate the CSD
FrequencySeries
for twoTimeSeries
csd_spectrogram
(other, stride[, fftlength, ...])Calculate the cross spectral density spectrogram of this
cumprod
([axis, dtype, out])Return the cumulative product of the elements along the given axis.
cumsum
([axis, dtype, out])Return the cumulative sum of the elements along the given axis.
decompose
([bases])Generates a new
Quantity
with the units decomposed.demodulate
(f[, stride, exp, deg])Compute the average magnitude and phase of this
TimeSeries
once per stride at a given frequencydetrend
([detrend])Remove the trend from this
TimeSeries
diagonal
([offset, axis1, axis2])Return specified diagonals.
diff
([n, axis])Calculate the nth order discrete difference along given axis.
dot
(b[, out])dump
(file)Not implemented, use
.value.dump()
instead.dumps
()Returns the pickle of the array as a string.
ediff1d
([to_end, to_begin])fetch
(channel, start, end[, host, port, ...])Fetch data from NDS
fetch_open_data
(ifo, start, end[, ...])Fetch openaccess data from the LIGO Open Science Center
fft
([nfft])Compute the onedimensional discrete Fourier transform of this
TimeSeries
.fftgram
(fftlength[, overlap, window])Calculate the Fouriergram of this
TimeSeries
.fill
(value)Fill the array with a scalar value.
filter
(*filt, **kwargs)Filter this
TimeSeries
with an IIR or FIR filterfind
(channel, start, end[, frametype, pad, ...])Find and read data from frames for a channel
find_gates
([tzero, whiten, threshold, ...])Identify points that should be gates using a provided threshold and clustered within a provided time window.
flatten
([order])Return a copy of the array collapsed into one dimension.
from_lal
(lalts[, copy])Generate a new TimeSeries from a LAL TimeSeries of any type.
from_nds2_buffer
(buffer_[, scaled, copy])Construct a new series from an
nds2.buffer
objectfrom_pycbc
(pycbcseries[, copy])Convert a
pycbc.types.timeseries.TimeSeries
into aTimeSeries
gate
([tzero, tpad, whiten, threshold, ...])Removes high amplitude peaks from data using inverse Planck window.
get
(channel, start, end[, pad, scaled, ...])Get data for this channel from frames or NDS
getfield
(dtype[, offset])Returns a field of the given array as a certain type.
heterodyne
(phase[, stride, singlesided])Compute the average magnitude and phase of this
TimeSeries
once per stride after heterodyning with a given phase serieshighpass
(frequency[, gpass, gstop, fstop, ...])Filter this
TimeSeries
with a highpass filter.inject
(other)Add two compatible
Series
along their shared xaxis values.insert
(obj, values[, axis])Insert values along the given axis before the given indices and return a new
Quantity
object.is_compatible
(other)Check whether this series and other have compatible metadata
is_contiguous
(other[, tol])Check whether other is contiguous with self.
item
(*args)Copy an element of an array to a scalar Quantity and return it.
lowpass
(frequency[, gpass, gstop, fstop, ...])Filter this
TimeSeries
with a Butterworth lowpass filter.mask
([deadtime, flag, query_open_data, ...])Mask away portions of this
TimeSeries
that fall within a given list of time segmentsmax
([axis, out, keepdims, initial, where])Return the maximum along a given axis.
mean
([axis, dtype, out, keepdims, where])Returns the average of the array elements along given axis.
median
([axis])Compute the median along the specified axis.
min
([axis, out, keepdims, initial, where])Return the minimum along a given axis.
nansum
([axis, out, keepdims, initial, where])nonzero
()Return the indices of the elements that are nonzero.
notch
(frequency[, type, filtfilt])Notch out a frequency in this
TimeSeries
.override_unit
(unit[, parse_strict])Forcefully reset the unit of these data
pad
(pad_width, **kwargs)Pad this series to a new size
partition
(kth[, axis, kind, order])Partially sorts the elements in the array in such a way that the value of the element in kth position is in the position it would be in a sorted array.
plot
([method, figsize, xscale])Plot the data for this timeseries
prepend
(other[, inplace, pad, gap, resize])Connect another series onto the start of the current one.
prod
([axis, dtype, out, keepdims, initial, ...])Return the product of the array elements over the given axis
psd
([fftlength, overlap, window, method])Calculate the PSD
FrequencySeries
for thisTimeSeries
put
(indices, values[, mode])Set
a.flat[n] = values[n]
for alln
in indices.q_gram
([qrange, frange, mismatch, snrthresh])Scan a
TimeSeries
using the multiQ transform and return anEventTable
of the most significant tilesq_transform
([qrange, frange, gps, search, ...])Scan a
TimeSeries
using the multiQ transform and return an interpolated highresolution spectrogramravel
([order])Return a flattened array.
rayleigh_spectrogram
(stride[, fftlength, ...])Calculate the Rayleigh statistic spectrogram of this
TimeSeries
rayleigh_spectrum
([fftlength, overlap, window])Calculate the Rayleigh
FrequencySeries
for thisTimeSeries
.read
(source, *args, **kwargs)Read data into a
TimeSeries
repeat
(repeats[, axis])Repeat elements of an array.
resample
(rate[, window, ftype, n])Resample this Series to a new rate
reshape
(shape, /, *[, order, copy])Returns an array containing the same data with a new shape.
resize
(new_shape[, refcheck])Change shape and size of array inplace.
rms
([stride])Calculate the rootmeansquare value of this
TimeSeries
once per stride.round
([decimals, out])Return
a
with each element rounded to the given number of decimals.searchsorted
(v[, side, sorter])Find indices where elements of v should be inserted in a to maintain order.
setfield
(val, dtype[, offset])Put a value into a specified place in a field defined by a datatype.
setflags
([write, align, uic])Set array flags WRITEABLE, ALIGNED, WRITEBACKIFCOPY, respectively.
shift
(delta)Shift this
Series
forward on the Xaxis bydelta
sort
([axis, kind, order])Sort an array inplace.
spectral_variance
(stride[, fftlength, ...])Calculate the
SpectralVariance
of thisTimeSeries
.spectrogram
(stride[, fftlength, overlap, ...])Calculate the average power spectrogram of this
TimeSeries
using the specified average spectrum method.spectrogram2
(fftlength[, overlap, window])Calculate the nonaveraged power
Spectrogram
of thisTimeSeries
squeeze
([axis])Remove axes of length one from
a
.std
([axis, dtype, out, ddof, keepdims, where])Returns the standard deviation of the array elements along given axis.
step
(**kwargs)Create a step plot of this series
sum
([axis, dtype, out, keepdims, initial, where])Return the sum of the array elements over the given axis.
swapaxes
(axis1, axis2)Return a view of the array with
axis1
andaxis2
interchanged.take
(indices[, axis, out, mode])Return an array formed from the elements of
a
at the given indices.taper
([side, duration, nsamples])Taper the ends of this
TimeSeries
smoothly to zero.to
(unit[, equivalencies, copy])Return a new
Quantity
object with the specified unit.to_lal
()Convert this
TimeSeries
into a LAL TimeSeries.to_pycbc
([copy])Convert this
TimeSeries
into a PyCBCTimeSeries
to_string
([unit, precision, format, subfmt])Generate a string representation of the quantity and its unit.
to_value
([unit, equivalencies])The numerical value, possibly in a different unit.
tobytes
([order])Not implemented, use
.value.tobytes()
instead.tofile
(fid[, sep, format])Not implemented, use
.value.tofile()
instead.tolist
()Return the array as an
a.ndim
levels deep nested list of Python scalars.tostring
([order])Construct Python bytes containing the raw data bytes in the array.
trace
([offset, axis1, axis2, dtype, out])Return the sum along diagonals of the array.
transfer_function
(other[, fftlength, ...])Calculate the transfer function between this
TimeSeries
and another.transpose
(*axes)Returns a view of the array with axes transposed.
update
(other[, inplace])Update this series by appending new data from an other and dropping the same amount of data off the start.
value_at
(x)Return the value of this
Series
at the givenxindex
valuevar
([axis, dtype, out, ddof, keepdims, where])Returns the variance of the array elements, along given axis.
view
([dtype][, type])New view of array with the same data.
whiten
([fftlength, overlap, method, window, ...])Whiten this
TimeSeries
using inverse spectrum truncationwrite
(target, *args, **kwargs)Write this
TimeSeries
to a filezip
()zpk
(zeros, poles, gain[, analog, unit])Filter this
TimeSeries
by applying a zeropolegain filterAttributes Documentation
 T¶
View of the transposed array.
Same as
self.transpose()
.See also
Examples
>>> import numpy as np >>> a = np.array([[1, 2], [3, 4]]) >>> a array([[1, 2], [3, 4]]) >>> a.T array([[1, 3], [2, 4]])
>>> a = np.array([1, 2, 3, 4]) >>> a array([1, 2, 3, 4]) >>> a.T array([1, 2, 3, 4])
 base¶
Base object if memory is from some other object.
Examples
The base of an array that owns its memory is None:
>>> import numpy as np >>> x = np.array([1,2,3,4]) >>> x.base is None True
Slicing creates a view, whose memory is shared with x:
>>> y = x[2:] >>> y.base is x True
 cgs¶
Returns a copy of the current
Quantity
instance with CGS units. The value of the resulting object will be scaled.
 ctypes¶
An object to simplify the interaction of the array with the ctypes module.
This attribute creates an object that makes it easier to use arrays when calling shared libraries with the ctypes module. The returned object has, among others, data, shape, and strides attributes (see Notes below) which themselves return ctypes objects that can be used as arguments to a shared library.
See also
Notes
Below are the public attributes of this object which were documented in “Guide to NumPy” (we have omitted undocumented public attributes, as well as documented private attributes):
 _ctypes.data
A pointer to the memory area of the array as a Python integer. This memory area may contain data that is not aligned, or not in correct byteorder. The memory area may not even be writeable. The array flags and datatype of this array should be respected when passing this attribute to arbitrary Ccode to avoid trouble that can include Python crashing. User Beware! The value of this attribute is exactly the same as:
self._array_interface_['data'][0]
.Note that unlike
data_as
, a reference won’t be kept to the array: code likectypes.c_void_p((a + b).ctypes.data)
will result in a pointer to a deallocated array, and should be spelt(a + b).ctypes.data_as(ctypes.c_void_p)
 _ctypes.shape
(c_intp*self.ndim): A ctypes array of length self.ndim where the basetype is the Cinteger corresponding to
dtype('p')
on this platform (seec_intp
). This basetype could bectypes.c_int
,ctypes.c_long
, orctypes.c_longlong
depending on the platform. The ctypes array contains the shape of the underlying array.
 _ctypes.strides
(c_intp*self.ndim): A ctypes array of length self.ndim where the basetype is the same as for the shape attribute. This ctypes array contains the strides information from the underlying array. This strides information is important for showing how many bytes must be jumped to get to the next element in the array.
 _ctypes.data_as(obj)
Return the data pointer cast to a particular ctypes object. For example, calling
self._as_parameter_
is equivalent toself.data_as(ctypes.c_void_p)
. Perhaps you want to use the data as a pointer to a ctypes array of floatingpoint data:self.data_as(ctypes.POINTER(ctypes.c_double))
.The returned pointer will keep a reference to the array.
 _ctypes.shape_as(obj)
Return the shape tuple as an array of some other ctypes type. For example:
self.shape_as(ctypes.c_short)
.
 _ctypes.strides_as(obj)
Return the strides tuple as an array of some other ctypes type. For example:
self.strides_as(ctypes.c_longlong)
.
If the ctypes module is not available, then the ctypes attribute of array objects still returns something useful, but ctypes objects are not returned and errors may be raised instead. In particular, the object will still have the
as_parameter
attribute which will return an integer equal to the data attribute.Examples
>>> import numpy as np >>> import ctypes >>> x = np.array([[0, 1], [2, 3]], dtype=np.int32) >>> x array([[0, 1], [2, 3]], dtype=int32) >>> x.ctypes.data 31962608 # may vary >>> x.ctypes.data_as(ctypes.POINTER(ctypes.c_uint32)) <__main__.LP_c_uint object at 0x7ff2fc1fc200> # may vary >>> x.ctypes.data_as(ctypes.POINTER(ctypes.c_uint32)).contents c_uint(0) >>> x.ctypes.data_as(ctypes.POINTER(ctypes.c_uint64)).contents c_ulong(4294967296) >>> x.ctypes.shape <numpy._core._internal.c_long_Array_2 object at 0x7ff2fc1fce60> # may vary >>> x.ctypes.strides <numpy._core._internal.c_long_Array_2 object at 0x7ff2fc1ff320> # may vary
 data¶
Python buffer object pointing to the start of the array’s data.
 device¶
 dtype¶
Datatype of the array’s elements.
Warning
Setting
arr.dtype
is discouraged and may be deprecated in the future. Setting will replace thedtype
without modifying the memory (see alsondarray.view
andndarray.astype
).See also
ndarray.astype
Cast the values contained in the array to a new datatype.
ndarray.view
Create a view of the same data but a different datatype.
numpy.dtype
Examples
>>> x array([[0, 1], [2, 3]]) >>> x.dtype dtype('int32') >>> type(x.dtype) <type 'numpy.dtype'>
 equivalencies¶
A list of equivalencies that will be applied by default during unit conversions.
 flags¶
Information about the memory layout of the array.
 Attributes:¶
 C_CONTIGUOUS (C)
The data is in a single, Cstyle contiguous segment.
 F_CONTIGUOUS (F)
The data is in a single, Fortranstyle contiguous segment.
 OWNDATA (O)
The array owns the memory it uses or borrows it from another object.
 WRITEABLE (W)
The data area can be written to. Setting this to False locks the data, making it readonly. A view (slice, etc.) inherits WRITEABLE from its base array at creation time, but a view of a writeable array may be subsequently locked while the base array remains writeable. (The opposite is not true, in that a view of a locked array may not be made writeable. However, currently, locking a base object does not lock any views that already reference it, so under that circumstance it is possible to alter the contents of a locked array via a previously created writeable view onto it.) Attempting to change a nonwriteable array raises a RuntimeError exception.
 ALIGNED (A)
The data and all elements are aligned appropriately for the hardware.
 WRITEBACKIFCOPY (X)
This array is a copy of some other array. The CAPI function PyArray_ResolveWritebackIfCopy must be called before deallocating to the base array will be updated with the contents of this array.
 FNC
F_CONTIGUOUS and not C_CONTIGUOUS.
 FORC
F_CONTIGUOUS or C_CONTIGUOUS (onesegment test).
 BEHAVED (B)
ALIGNED and WRITEABLE.
 CARRAY (CA)
BEHAVED and C_CONTIGUOUS.
 FARRAY (FA)
BEHAVED and F_CONTIGUOUS and not C_CONTIGUOUS.
Notes
The
flags
object can be accessed dictionarylike (as ina.flags['WRITEABLE']
), or by using lowercased attribute names (as ina.flags.writeable
). Short flag names are only supported in dictionary access.Only the WRITEBACKIFCOPY, WRITEABLE, and ALIGNED flags can be changed by the user, via direct assignment to the attribute or dictionary entry, or by calling
ndarray.setflags
.The array flags cannot be set arbitrarily:
WRITEBACKIFCOPY can only be set
False
.ALIGNED can only be set
True
if the data is truly aligned.WRITEABLE can only be set
True
if the array owns its own memory or the ultimate owner of the memory exposes a writeable buffer interface or is a string.
Arrays can be both Cstyle and Fortranstyle contiguous simultaneously. This is clear for 1dimensional arrays, but can also be true for higher dimensional arrays.
Even for contiguous arrays a stride for a given dimension
arr.strides[dim]
may be arbitrary ifarr.shape[dim] == 1
or the array has no elements. It does not generally hold thatself.strides[1] == self.itemsize
for Cstyle contiguous arrays orself.strides[0] == self.itemsize
for Fortranstyle contiguous arrays is true.
 flat¶
A 1D iterator over the Quantity array.
This returns a
QuantityIterator
instance, which behaves the same as theflatiter
instance returned byflat
, and is similar to, but not a subclass of, Python’s builtin iterator object.
 imag¶
The imaginary part of the array.
Examples
>>> import numpy as np >>> x = np.sqrt([1+0j, 0+1j]) >>> x.imag array([ 0. , 0.70710678]) >>> x.imag.dtype dtype('float64')
 info¶
Container for meta information like name, description, format. This is required when the object is used as a mixin column within a table, but can be used as a general way to store meta information.
 isscalar¶
True if the
value
of this quantity is a scalar, or False if it is an arraylike object.Note
This is subtly different from
numpy.isscalar
in thatnumpy.isscalar
returns False for a zerodimensional array (e.g.np.array(1)
), while this is True for quantities, since quantities cannot represent true numpy scalars.
 itemset¶
 itemsize¶
Length of one array element in bytes.
Examples
>>> import numpy as np >>> x = np.array([1,2,3], dtype=np.float64) >>> x.itemsize 8 >>> x = np.array([1,2,3], dtype=np.complex128) >>> x.itemsize 16
 mT¶
View of the matrix transposed array.
The matrix transpose is the transpose of the last two dimensions, even if the array is of higher dimension.
Added in version 2.0.
 Raises:¶
 ValueError
If the array is of dimension less than 2.
Examples
>>> import numpy as np >>> a = np.array([[1, 2], [3, 4]]) >>> a array([[1, 2], [3, 4]]) >>> a.mT array([[1, 3], [2, 4]])
>>> a = np.arange(8).reshape((2, 2, 2)) >>> a array([[[0, 1], [2, 3]], [[4, 5], [6, 7]]]) >>> a.mT array([[[0, 2], [1, 3]], [[4, 6], [5, 7]]])
 nbytes¶
Total bytes consumed by the elements of the array.
See also
sys.getsizeof
Memory consumed by the object itself without parents in case view. This does include memory consumed by nonelement attributes.
Notes
Does not include memory consumed by nonelement attributes of the array object.
Examples
>>> import numpy as np >>> x = np.zeros((3,5,2), dtype=np.complex128) >>> x.nbytes 480 >>> np.prod(x.shape) * x.itemsize 480
 ndim¶
Number of array dimensions.
Examples
>>> import numpy as np >>> x = np.array([1, 2, 3]) >>> x.ndim 1 >>> y = np.zeros((2, 3, 4)) >>> y.ndim 3
 newbyteorder¶
 ptp¶
 real¶
The real part of the array.
See also
numpy.real
equivalent function
Examples
>>> import numpy as np >>> x = np.sqrt([1+0j, 0+1j]) >>> x.real array([ 1. , 0.70710678]) >>> x.real.dtype dtype('float64')
 sample_rate¶
Data rate for this
TimeSeries
in samples per second (Hertz).This attribute is stored internally by the
dx
attribute
 shape¶
Tuple of array dimensions.
The shape property is usually used to get the current shape of an array, but may also be used to reshape the array inplace by assigning a tuple of array dimensions to it. As with
numpy.reshape
, one of the new shape dimensions can be 1, in which case its value is inferred from the size of the array and the remaining dimensions. Reshaping an array inplace will fail if a copy is required.Warning
Setting
arr.shape
is discouraged and may be deprecated in the future. Usingndarray.reshape
is the preferred approach.See also
numpy.shape
Equivalent getter function.
numpy.reshape
Function similar to setting
shape
.ndarray.reshape
Method similar to setting
shape
.
Examples
>>> import numpy as np >>> x = np.array([1, 2, 3, 4]) >>> x.shape (4,) >>> y = np.zeros((2, 3, 4)) >>> y.shape (2, 3, 4) >>> y.shape = (3, 8) >>> y array([[ 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0.]]) >>> y.shape = (3, 6) Traceback (most recent call last): File "<stdin>", line 1, in <module> ValueError: total size of new array must be unchanged >>> np.zeros((4,2))[::2].shape = (1,) Traceback (most recent call last): File "<stdin>", line 1, in <module> AttributeError: Incompatible shape for inplace modification. Use `.reshape()` to make a copy with the desired shape.
 si¶
Returns a copy of the current
Quantity
instance with SI units. The value of the resulting object will be scaled.
 size¶
Number of elements in the array.
Equal to
np.prod(a.shape)
, i.e., the product of the array’s dimensions.Notes
a.size
returns a standard arbitrary precision Python integer. This may not be the case with other methods of obtaining the same value (like the suggestednp.prod(a.shape)
, which returns an instance ofnp.int_
), and may be relevant if the value is used further in calculations that may overflow a fixed size integer type.Examples
>>> import numpy as np >>> x = np.zeros((3, 5, 2), dtype=np.complex128) >>> x.size 30 >>> np.prod(x.shape) 30
 strides¶
Tuple of bytes to step in each dimension when traversing an array.
The byte offset of element
(i[0], i[1], ..., i[n])
in an arraya
is:offset = sum(np.array(i) * a.strides)
A more detailed explanation of strides can be found in The Ndimensional array (ndarray).
Warning
Setting
arr.strides
is discouraged and may be deprecated in the future.numpy.lib.stride_tricks.as_strided
should be preferred to create a new view of the same data in a safer way.See also
Notes
Imagine an array of 32bit integers (each 4 bytes):
x = np.array([[0, 1, 2, 3, 4], [5, 6, 7, 8, 9]], dtype=np.int32)
This array is stored in memory as 40 bytes, one after the other (known as a contiguous block of memory). The strides of an array tell us how many bytes we have to skip in memory to move to the next position along a certain axis. For example, we have to skip 4 bytes (1 value) to move to the next column, but 20 bytes (5 values) to get to the same position in the next row. As such, the strides for the array
x
will be(20, 4)
.Examples
>>> import numpy as np >>> y = np.reshape(np.arange(2*3*4), (2,3,4)) >>> y array([[[ 0, 1, 2, 3], [ 4, 5, 6, 7], [ 8, 9, 10, 11]], [[12, 13, 14, 15], [16, 17, 18, 19], [20, 21, 22, 23]]]) >>> y.strides (48, 16, 4) >>> y[1,1,1] 17 >>> offset=sum(y.strides * np.array((1,1,1))) >>> offset/y.itemsize 17
>>> x = np.reshape(np.arange(5*6*7*8), (5,6,7,8)).transpose(2,3,1,0) >>> x.strides (32, 4, 224, 1344) >>> i = np.array([3,5,2,2]) >>> offset = sum(i * x.strides) >>> x[3,5,2,2] 813 >>> offset / x.itemsize 813
 value¶
The numerical value of this instance.
See also
to_value
Get the numerical value in a given unit.
Methods Documentation
 abs(x, /, out=None, *, where=True, casting='same_kind', order='K', dtype=None, subok=True[, signature])[source]¶
Calculate the absolute value elementwise.
np.abs
is a shorthand for this function. Parameters:¶
 xarray_like
Input array.
 outndarray, None, or tuple of ndarray and None, optional
A location into which the result is stored. If provided, it must have a shape that the inputs broadcast to. If not provided or None, a freshlyallocated array is returned. A tuple (possible only as a keyword argument) must have length equal to the number of outputs.
 wherearray_like, optional
This condition is broadcast over the input. At locations where the condition is True, the
out
array will be set to the ufunc result. Elsewhere, theout
array will retain its original value. Note that if an uninitializedout
array is created via the defaultout=None
, locations within it where the condition is False will remain uninitialized. **kwargs
For other keywordonly arguments, see the ufunc docs.
 Returns:¶
 absolutendarray
An ndarray containing the absolute value of each element in
. This is a scalar ifx
. For complex input,a + ib
, the absolute value isx
is a scalar.
Examples
>>> import numpy as np >>> x = np.array([1.2, 1.2]) >>> np.absolute(x) array([ 1.2, 1.2]) >>> np.absolute(1.2 + 1j) 1.5620499351813308
Plot the function over
[10, 10]
:>>> import matplotlib.pyplot as plt
>>> x = np.linspace(start=10, stop=10, num=101) >>> plt.plot(x, np.absolute(x)) >>> plt.show()
(
png
)Plot the function over the complex plane:
>>> xx = x + 1j * x[:, np.newaxis] >>> plt.imshow(np.abs(xx), extent=[10, 10, 10, 10], cmap='gray') >>> plt.show()
(
png
)The
abs
function can be used as a shorthand fornp.absolute
on ndarrays.>>> x = np.array([1.2, 1.2]) >>> abs(x) array([1.2, 1.2])

all(axis=
None
, out=None
, keepdims=False
, *, where=True
)¶ Returns True if all elements evaluate to True.
Refer to
numpy.all
for full documentation.See also
numpy.all
equivalent function

any(axis=
None
, out=None
, keepdims=False
, *, where=True
)¶ Returns True if any of the elements of
a
evaluate to True.Refer to
numpy.any
for full documentation.See also
numpy.any
equivalent function

append(other, inplace=
True
, pad=None
, gap=None
, resize=True
)[source]¶ Connect another series onto the end of the current one.
 Parameters:¶
 other
Series
another series of the same type to connect to this one
 inplace
bool
, optional perform operation inplace, modifying current series, otherwise copy data and return new series, default:
True
Warning
inplace
append bypasses the reference check innumpy.ndarray.resize
, so be carefully to only use this for arrays that haven’t been sharing their memory! pad
float
, optional value with which to pad discontiguous series, by default gaps will result in a
ValueError
. gap
str
, optional action to perform if there’s a gap between the other series and this one. One of
'raise'
 raise aValueError
'ignore'
 remove gap and join data'pad'
 pad gap with zeros
If
pad
is given and is notNone
, the default is'pad'
, otherwise'raise'
. Ifgap='pad'
is given, the default forpad
is0
. resize
bool
, optional resize this array to accommodate new data, otherwise shift the old data to the left (potentially falling off the start) and put the new data in at the end, default:
True
.
 other
 Returns:¶
 series
Series
a new series containing joined data sets
 series

argmax(axis=
None
, out=None
, *, keepdims=False
)¶ Return indices of the maximum values along the given axis.
Refer to
numpy.argmax
for full documentation.See also
numpy.argmax
equivalent function

argmin(axis=
None
, out=None
, *, keepdims=False
)¶ Return indices of the minimum values along the given axis.
Refer to
numpy.argmin
for detailed documentation.See also
numpy.argmin
equivalent function

argpartition(kth, axis=
1
, kind='introselect'
, order=None
)¶ Returns the indices that would partition this array.
Refer to
numpy.argpartition
for full documentation.Added in version 1.8.0.
See also
numpy.argpartition
equivalent function

argsort(axis=
1
, kind=None
, order=None
)¶ Returns the indices that would sort this array.
Refer to
numpy.argsort
for full documentation.See also
numpy.argsort
equivalent function

asd(fftlength=
None
, overlap=None
, window='hann'
, method='median'
, **kwargs)[source]¶ Calculate the ASD
FrequencySeries
of thisTimeSeries
 Parameters:¶
 fftlength
float
number of seconds in single FFT, defaults to a single FFT covering the full duration
 overlap
float
, optional number of seconds of overlap between FFTs, defaults to the recommended overlap for the given window (if given), or 0
 window
str
,numpy.ndarray
, optional window function to apply to timeseries prior to FFT, see
scipy.signal.get_window()
for details on acceptable formats method
str
, optional FFTaveraging method (default:
'median'
), see Notes for more details
 fftlength
 Returns:¶
 asd
FrequencySeries
a data series containing the ASD
 asd
See also
Notes
The accepted
method
arguments are:'bartlett'
: a mean average of nonoverlapping periodograms'median'
: a median average of overlapping periodograms'welch'
: a mean average of overlapping periodograms

astype(dtype, order=
'K'
, casting='unsafe'
, subok=True
, copy=True
)¶ Copy of the array, cast to a specified type.
 Parameters:¶
 dtypestr or dtype
Typecode or datatype to which the array is cast.
 order{‘C’, ‘F’, ‘A’, ‘K’}, optional
Controls the memory layout order of the result. ‘C’ means C order, ‘F’ means Fortran order, ‘A’ means ‘F’ order if all the arrays are Fortran contiguous, ‘C’ order otherwise, and ‘K’ means as close to the order the array elements appear in memory as possible. Default is ‘K’.
 casting{‘no’, ‘equiv’, ‘safe’, ‘same_kind’, ‘unsafe’}, optional
Controls what kind of data casting may occur. Defaults to ‘unsafe’ for backwards compatibility.
‘no’ means the data types should not be cast at all.
‘equiv’ means only byteorder changes are allowed.
‘safe’ means only casts which can preserve values are allowed.
‘same_kind’ means only safe casts or casts within a kind, like float64 to float32, are allowed.
‘unsafe’ means any data conversions may be done.
 subokbool, optional
If True, then subclasses will be passedthrough (default), otherwise the returned array will be forced to be a baseclass array.
 copybool, optional
By default, astype always returns a newly allocated array. If this is set to false, and the
dtype
,order
, andsubok
requirements are satisfied, the input array is returned instead of a copy.
 Returns:¶
 Raises:¶
 ComplexWarning
When casting from complex to float or int. To avoid this, one should use
a.real.astype(t)
.
Notes
Changed in version 1.17.0: Casting between a simple data type and a structured one is possible only for “unsafe” casting. Casting to multiple fields is allowed, but casting from multiple fields is not.
Changed in version 1.9.0: Casting from numeric to string types in ‘safe’ casting mode requires that the string dtype length is long enough to store the max integer/float value converted.
Examples
>>> import numpy as np >>> x = np.array([1, 2, 2.5]) >>> x array([1. , 2. , 2.5])
>>> x.astype(int) array([1, 2, 2])

auto_coherence(dt, fftlength=
None
, overlap=None
, window='hann'
, **kwargs)[source]¶ Calculate the frequencycoherence between this
TimeSeries
and a timeshifted copy of itself.The standard
TimeSeries.coherence()
is calculated between the inputTimeSeries
and acropped
copy of itself. Since the cropped version will be shorter, the input series will be shortened to match. Parameters:¶
 dt
float
duration (in seconds) of timeshift
 fftlength
float
, optional number of seconds in single FFT, defaults to a single FFT covering the full duration
 overlap
float
, optional number of seconds of overlap between FFTs, defaults to the recommended overlap for the given window (if given), or 0
 window
str
,numpy.ndarray
, optional window function to apply to timeseries prior to FFT, see
scipy.signal.get_window()
for details on acceptable formats **kwargs
any other keyword arguments accepted by
matplotlib.mlab.cohere()
exceptNFFT
,window
, andnoverlap
which are superceded by the above keyword arguments
 dt
 Returns:¶
 coherence
FrequencySeries
the coherence
FrequencySeries
of thisTimeSeries
with the other
 coherence
See also
matplotlib.mlab.cohere
for details of the coherence calculator
Notes
The
TimeSeries.auto_coherence()
will perform best whendt
is approximatelyfftlength / 2
.

average_fft(fftlength=
None
, overlap=0
, window=None
)[source]¶ Compute the averaged onedimensional DFT of this
TimeSeries
.This method computes a number of FFTs of duration
fftlength
andoverlap
(both given in seconds), and returns the mean average. This method is analogous to the Welch average method for power spectra. Parameters:¶
 fftlength
float
number of seconds in single FFT, default, use whole
TimeSeries
 overlap
float
, optional number of seconds of overlap between FFTs, defaults to the recommended overlap for the given window (if given), or 0
 window
str
,numpy.ndarray
, optional window function to apply to timeseries prior to FFT, see
scipy.signal.get_window()
for details on acceptable formats
 fftlength
 Returns:¶
 outcomplexvalued
FrequencySeries
the transformed output, with populated frequencies array metadata
 outcomplexvalued
See also
TimeSeries.fft
The FFT method used.

bandpass(flow, fhigh, gpass=
2
, gstop=30
, fstop=None
, type='iir'
, filtfilt=True
, **kwargs)[source]¶ Filter this
TimeSeries
with a bandpass filter. Parameters:¶
 flow
float
lower corner frequency of pass band
 fhigh
float
upper corner frequency of pass band
 gpass
float
the maximum loss in the passband (dB).
 gstop
float
the minimum attenuation in the stopband (dB).
 fstop
tuple
offloat
, optional (low, high)
edgefrequencies of stop band type
str
the filter type, either
'iir'
or'fir'
 **kwargs
other keyword arguments are passed to
gwpy.signal.filter_design.bandpass()
 flow
 Returns:¶
 bpseries
TimeSeries
a bandpassed version of the input
TimeSeries
 bpseries
See also
gwpy.signal.filter_design.bandpass
for details on the filter design
TimeSeries.filter
for details on how the filter is applied

byteswap(inplace=
False
)¶ Swap the bytes of the array elements
Toggle between lowendian and bigendian data representation by returning a byteswapped array, optionally swapped inplace. Arrays of bytestrings are not swapped. The real and imaginary parts of a complex number are swapped individually.
 Parameters:¶
 inplacebool, optional
If
True
, swap bytes inplace, default isFalse
.
 Returns:¶
 outndarray
The byteswapped array. If
inplace
isTrue
, this is a view to self.
Examples
>>> import numpy as np >>> 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 bytestrings are not swapped
>>> A = np.array([b'ceg', b'fac']) >>> A.byteswap() array([b'ceg', b'fac'], dtype='S3')
A.view(A.dtype.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.view(A.dtype.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)

choose(choices, out=
None
, mode='raise'
)¶ Use an index array to construct a new array from a set of choices.
Refer to
numpy.choose
for full documentation.See also
numpy.choose
equivalent function

clip(min=
None
, max=None
, out=None
, **kwargs)¶ Return an array whose values are limited to
[min, max]
. One of max or min must be given.Refer to
numpy.clip
for full documentation.See also
numpy.clip
equivalent function

coherence(other, fftlength=
None
, overlap=None
, window='hann'
, **kwargs)[source]¶ Calculate the frequencycoherence between this
TimeSeries
and another. Parameters:¶
 other
TimeSeries
TimeSeries
signal to calculate coherence with fftlength
float
, optional number of seconds in single FFT, defaults to a single FFT covering the full duration
 overlap
float
, optional number of seconds of overlap between FFTs, defaults to the recommended overlap for the given window (if given), or 0
 window
str
,numpy.ndarray
, optional window function to apply to timeseries prior to FFT, see
scipy.signal.get_window()
for details on acceptable formats **kwargs
any other keyword arguments accepted by
matplotlib.mlab.cohere()
exceptNFFT
,window
, andnoverlap
which are superceded by the above keyword arguments
 other
 Returns:¶
 coherence
FrequencySeries
the coherence
FrequencySeries
of thisTimeSeries
with the other
 coherence
See also
scipy.signal.coherence
for details of the coherence calculator
Notes
If
self
andother
have differenceTimeSeries.sample_rate
values, the higher sampledTimeSeries
will be downsampled to match the lower.

coherence_spectrogram(other, stride, fftlength=
None
, overlap=None
, window='hann'
, nproc=1
)[source]¶ Calculate the coherence spectrogram between this
TimeSeries
and other. Parameters:¶
 other
TimeSeries
the second
TimeSeries
in this CSD calculation stride
float
number of seconds in single PSD (column of spectrogram)
 fftlength
float
number of seconds in single FFT
 overlap
float
, optional number of seconds of overlap between FFTs, defaults to the recommended overlap for the given window (if given), or 0
 window
str
,numpy.ndarray
, optional window function to apply to timeseries prior to FFT, see
scipy.signal.get_window()
for details on acceptable formats nproc
int
number of parallel processes to use when calculating individual coherence spectra.
 other
 Returns:¶
 spectrogram
Spectrogram
timefrequency coherence spectrogram as generated from the input timeseries.
 spectrogram

compress(condition, axis=
None
, out=None
)¶ Return selected slices of this array along given axis.
Refer to
numpy.compress
for full documentation.See also
numpy.compress
equivalent function
 conj()¶
Complexconjugate all elements.
Refer to
numpy.conjugate
for full documentation.See also
numpy.conjugate
equivalent function
 conjugate()¶
Return the complex conjugate, elementwise.
Refer to
numpy.conjugate
for full documentation.See also
numpy.conjugate
equivalent function

convolve(fir, window=
'hann'
)[source]¶  Convolve this
TimeSeries
with an FIR filter using the overlapsave method
 Parameters:¶
 fir
numpy.ndarray
the time domain filter to convolve with
 window
str
, optional window function to apply to boundaries, default:
'hann'
seescipy.signal.get_window()
for details on acceptable formats
 fir
 Returns:¶
 out
TimeSeries
the result of the convolution
 out
See also
scipy.signal.fftconvolve
for details on the convolution scheme used here
TimeSeries.filter
for an alternative method designed for short filters
Notes
The output
TimeSeries
is the same length and has the same timestamps as the input.Due to filter settlein, a segment half the length of
fir
will be corrupted at the left and right boundaries. To prevent spectral leakage these segments will be windowed before convolving. Convolve this

copy(order=
'C'
)[source]¶ Return a copy of the array.
 Parameters:¶
 order{‘C’, ‘F’, ‘A’, ‘K’}, optional
Controls the memory layout of the copy. ‘C’ means Corder, ‘F’ means Forder, ‘A’ means ‘F’ if
a
is Fortran contiguous, ‘C’ otherwise. ‘K’ means match the layout ofa
as closely as possible. (Note that this function andnumpy.copy()
are very similar but have different default values for their order= arguments, and this function always passes subclasses through.)
See also
numpy.copy
Similar function with different default behavior
numpy.copyto
Notes
This function is the preferred method for creating an array copy. The function
numpy.copy()
is similar, but it defaults to using order ‘K’, and will not pass subclasses through by default.Examples
>>> import numpy as np >>> x = np.array([[1,2,3],[4,5,6]], order='F')
>>> y = x.copy()
>>> x.fill(0)
>>> x array([[0, 0, 0], [0, 0, 0]])
>>> y array([[1, 2, 3], [4, 5, 6]])
>>> y.flags['C_CONTIGUOUS'] True
For arrays containing Python objects (e.g. dtype=object), the copy is a shallow one. The new array will contain the same object which may lead to surprises if that object can be modified (is mutable):
>>> a = np.array([1, 'm', [2, 3, 4]], dtype=object) >>> b = a.copy() >>> b[2][0] = 10 >>> a array([1, 'm', list([10, 3, 4])], dtype=object)
To ensure all elements within an
object
array are copied, usecopy.deepcopy
:>>> import copy >>> a = np.array([1, 'm', [2, 3, 4]], dtype=object) >>> c = copy.deepcopy(a) >>> c[2][0] = 10 >>> c array([1, 'm', list([10, 3, 4])], dtype=object) >>> a array([1, 'm', list([2, 3, 4])], dtype=object)

correlate(mfilter, window=
'hann'
, detrend='linear'
, whiten=False
, wduration=2
, highpass=None
, **asd_kw)[source]¶ Crosscorrelate this
TimeSeries
with another signal Parameters:¶
 mfilter
TimeSeries
the time domain signal to correlate with
 window
str
, optional window function to apply to timeseries prior to FFT, default:
'hann'
seescipy.signal.get_window()
for details on acceptable formats detrend
str
, optional type of detrending to do before FFT (see
detrend
for more details), default:'linear'
 whiten
bool
, optional boolean switch to enable (
True
) or disable (False
) data whitening, default:False
 wduration
float
, optional duration (in seconds) of the timedomain FIR whitening filter, only used if
whiten=True
, defaults to 2 seconds highpass
float
, optional highpass corner frequency (in Hz) of the FIR whitening filter, only used if
whiten=True
, default:None
 **asd_kw
keyword arguments to pass to
TimeSeries.asd
to generate an ASD, only used ifwhiten=True
 mfilter
 Returns:¶
 snr
TimeSeries
the correlated signaltonoise ratio (SNR) timeseries
 snr
See also
TimeSeries.asd
for details on the ASD calculation
TimeSeries.convolve
for details on convolution with the overlapsave method
Notes
The
window
argument is used in ASD estimation, whitening, and preventing spectral leakage in the output. It is not used to condition the matchedfilter, which should be windowed before passing to this method.Due to filter settlein, a segment half the length of
mfilter
will be corrupted at the beginning and end of the output. Seeconvolve
for more details.The input and matchedfilter will be detrended, and the output will be normalised so that the SNR measures number of standard deviations from the expected mean.

crop(start=
None
, end=None
, copy=False
)[source]¶ Crop this series to the given xaxis extent.
Notes
If either
start
orend
are outside of the originalSeries
span, warnings will be printed and the limits will be restricted to thexspan
.

csd(other, fftlength=
None
, overlap=None
, window='hann'
, **kwargs)[source]¶ Calculate the CSD
FrequencySeries
for twoTimeSeries
 Parameters:¶
 other
TimeSeries
the second
TimeSeries
in this CSD calculation fftlength
float
number of seconds in single FFT, defaults to a single FFT covering the full duration
 overlap
float
, optional number of seconds of overlap between FFTs, defaults to the recommended overlap for the given window (if given), or 0
 window
str
,numpy.ndarray
, optional window function to apply to timeseries prior to FFT, see
scipy.signal.get_window()
for details on acceptable formats
 other
 Returns:¶
 csd
FrequencySeries
a data series containing the CSD.
 csd

csd_spectrogram(other, stride, fftlength=
None
, overlap=0
, window='hann'
, nproc=1
, **kwargs)[source]¶  Calculate the cross spectral density spectrogram of this
TimeSeries
with ‘other’.
 Parameters:¶
 other
TimeSeries
second timeseries for cross spectral density calculation
 stride
float
number of seconds in single PSD (column of spectrogram).
 fftlength
float
number of seconds in single FFT.
 overlap
float
, optional number of seconds of overlap between FFTs, defaults to the recommended overlap for the given window (if given), or 0
 window
str
,numpy.ndarray
, optional window function to apply to timeseries prior to FFT, see
scipy.signal.get_window()
for details on acceptable formats nproc
int
maximum number of independent frame reading processes, default is set to singleprocess file reading.
 other
 Returns:¶
 spectrogram
Spectrogram
timefrequency cross spectrogram as generated from the two input timeseries.
 spectrogram

cumprod(axis=
None
, dtype=None
, out=None
)¶ Return the cumulative product of the elements along the given axis.
Refer to
numpy.cumprod
for full documentation.See also
numpy.cumprod
equivalent function

cumsum(axis=
None
, dtype=None
, out=None
)¶ Return the cumulative sum of the elements along the given axis.
Refer to
numpy.cumsum
for full documentation.See also
numpy.cumsum
equivalent function

decompose(bases=
[]
)¶ Generates a new
Quantity
with the units decomposed. Decomposed units have only irreducible units in them (seeastropy.units.UnitBase.decompose
). Parameters:¶
 basessequence of
UnitBase
, optional The bases to decompose into. When not provided, decomposes down to any irreducible units. When provided, the decomposed result will only contain the given units. This will raises a
UnitsError
if it’s not possible to do so.
 basessequence of
 Returns:¶
 newq
Quantity
A new object equal to this quantity with units decomposed.
 newq

demodulate(f, stride=
1
, exp=False
, deg=True
)[source]¶ Compute the average magnitude and phase of this
TimeSeries
once per stride at a given frequency Parameters:¶
 f
float
frequency (Hz) at which to demodulate the signal
 stride
float
, optional stride (seconds) between calculations, defaults to 1 second
 exp
bool
, optional return the magnitude and phase trends as one
TimeSeries
object representing a complex exponential, default: False deg
bool
, optional if
exp=False
, calculates the phase in degrees
 f
 Returns:¶
 mag, phase
TimeSeries
if
exp=False
, returns a pair ofTimeSeries
objects representing magnitude and phase trends withdt=stride
 out
TimeSeries
if
exp=True
, returns a singleTimeSeries
with magnitude and phase trends represented asmag * exp(1j*phase)
withdt=stride
 mag, phase
See also
TimeSeries.heterodyne
for the underlying heterodyne detection method
Examples
Demodulation is useful when trying to examine steady sinusoidal signals we know to be contained within data. For instance, we can download some data from GWOSC to look at trends of the amplitude and phase of LIGO Livingston’s calibration line at 331.3 Hz:
>>> from gwpy.timeseries import TimeSeries >>> data = TimeSeries.fetch_open_data('L1', 1131350417, 1131357617)
We can demodulate the
TimeSeries
at 331.3 Hz with a stride of one minute:>>> amp, phase = data.demodulate(331.3, stride=60)
We can then plot these trends to visualize fluctuations in the amplitude of the calibration line:
>>> from gwpy.plot import Plot >>> plot = Plot(amp) >>> ax = plot.gca() >>> ax.set_ylabel('Strain Amplitude at 331.3 Hz') >>> plot.show()
(
png
)

detrend(detrend=
'constant'
)[source]¶ Remove the trend from this
TimeSeries
This method just wraps
scipy.signal.detrend()
to return an object of the same type as the input. Parameters:¶
 detrend
str
, optional the type of detrending.
 detrend
 Returns:¶
 detrended
TimeSeries
the detrended input series
 detrended
See also
scipy.signal.detrend
for details on the options for the
detrend
argument, and how the operation is done

diagonal(offset=
0
, axis1=0
, axis2=1
)¶ Return specified diagonals. In NumPy 1.9 the returned array is a readonly view instead of a copy as in previous NumPy versions. In a future version the readonly restriction will be removed.
Refer to
numpy.diagonal()
for full documentation.See also
numpy.diagonal
equivalent function

diff(n=
1
, axis=1
)[source]¶ Calculate the nth order discrete difference along given axis.
The first order difference is given by
out[n] = a[n+1]  a[n]
along the given axis, higher order differences are calculated by usingdiff
recursively. Parameters:¶
 nint, optional
The number of times values are differenced.
 axisint, optional
The axis along which the difference is taken, default is the last axis.
 Returns:¶
 diff
Series
The
n
order differences. The shape of the output is the same as the input, except alongaxis
where the dimension is smaller byn
.
 diff
See also
numpy.diff
for documentation on the underlying method
 dumps()[source]¶
Returns the pickle of the array as a string. pickle.loads will convert the string back to an array.
 Parameters:¶
 None

classmethod fetch(channel, start, end, host=
None
, port=None
, verbose=False
, connection=None
, verify=False
, pad=None
, allow_tape=None
, scaled=None
, type=None
, dtype=None
)[source]¶ Fetch data from NDS
 Parameters:¶
 channel
str
,Channel
the data channel for which to query
 start
LIGOTimeGPS
,float
,str
GPS start time of required data, any input parseable by
to_gps
is fine end
LIGOTimeGPS
,float
,str
GPS end time of required data, any input parseable by
to_gps
is fine host
str
, optional URL of NDS server to use, if blank will try any server (in a relatively sensible order) to get the data
 port
int
, optional port number for NDS server query, must be given with
host
 verify
bool
, optional, default:False
check channels exist in database before asking for data
 scaled
bool
, optional apply slope and bias calibration to ADC data, for nonADC data this option has no effect
 connection
nds2.connection
, optional open NDS connection to use
 verbose
bool
, optional print verbose output about NDS progress, useful for debugging; if
verbose
is specified as a string, this defines the prefix for the progress meter type
int
, optional NDS2 channel type integer or string name to match
 dtype
type
,numpy.dtype
,str
, optional NDS2 data type to match
 channel

classmethod fetch_open_data(ifo, start, end, sample_rate=
4096
, version=None
, format='hdf5'
, host='https://gwosc.org'
, verbose=False
, cache=None
, **kwargs)[source]¶ Fetch openaccess data from the LIGO Open Science Center
 Parameters:¶
 ifo
str
the twocharacter prefix of the IFO in which you are interested, e.g.
'L1'
 start
LIGOTimeGPS
,float
,str
, optional GPS start time of required data, defaults to start of data found; any input parseable by
to_gps
is fine end
LIGOTimeGPS
,float
,str
, optional GPS end time of required data, defaults to end of data found; any input parseable by
to_gps
is fine sample_rate
float
, optional, the sample rate of desired data; most data are stored by GWOSC at 4096 Hz, however there may be eventrelated data releases with a 16384 Hz rate, default:
4096
 version
int
, optional version of files to download, defaults to highest discovered version
 format
str
, optional the data format to download and parse, default:
'h5py'
'hdf5'
'gwf'
 requiresLDAStools.frameCPP
 host
str
, optional HTTP host name of GWOSC server to access
 verbose
bool
, optional, default:False
print verbose output while fetching data
 cache
bool
, optional save/read a local copy of the remote URL, default:
False
; useful if the same remote data are to be accessed multiple times. SetGWPY_CACHE=1
in the environment to autocache. **kwargs
any other keyword arguments are passed to the
TimeSeries.read
method that parses the file that was downloaded
 ifo
Notes
StateVector
data are not available intxt.gz
format.Examples
>>> from gwpy.timeseries import (TimeSeries, StateVector) >>> print(TimeSeries.fetch_open_data('H1', 1126259446, 1126259478)) TimeSeries([ 2.17704028e19, 2.08763900e19, 2.39681183e19, ..., 3.55365541e20, 6.33533516e20, 7.58121195e20] unit: Unit(dimensionless), t0: 1126259446.0 s, dt: 0.000244140625 s, name: Strain, channel: None) >>> print(StateVector.fetch_open_data('H1', 1126259446, 1126259478)) StateVector([127,127,127,127,127,127,127,127,127,127,127,127, 127,127,127,127,127,127,127,127,127,127,127,127, 127,127,127,127,127,127,127,127] unit: Unit(dimensionless), t0: 1126259446.0 s, dt: 1.0 s, name: Data quality, channel: None, bits: Bits(0: data present 1: passes cbc CAT1 test 2: passes cbc CAT2 test 3: passes cbc CAT3 test 4: passes burst CAT1 test 5: passes burst CAT2 test 6: passes burst CAT3 test, channel=None, epoch=1126259446.0))
For the
StateVector
, the naming of the bits will beformat
dependent, because they are recorded differently by GWOSC in different formats.

fft(nfft=
None
)[source]¶ Compute the onedimensional discrete Fourier transform of this
TimeSeries
. Parameters:¶
 nfft
int
, optional length of the desired Fourier transform, input will be cropped or padded to match the desired length. If nfft is not given, the length of the
TimeSeries
will be used
 nfft
 Returns:¶
 out
FrequencySeries
the normalised, complexvalued FFT
FrequencySeries
.
 out
See also
numpy.fft.rfft
The FFT implementation used in this method.
Notes
This method, in constrast to the
numpy.fft.rfft()
method it calls, applies the necessary normalisation such that the amplitude of the outputFrequencySeries
is correct.

fftgram(fftlength, overlap=
None
, window='hann'
, **kwargs)[source]¶ Calculate the Fouriergram of this
TimeSeries
.At every
stride
, a single, complex FFT is calculated. Parameters:¶
 fftlength
float
number of seconds in single FFT.
 overlap
float
, optional number of seconds of overlap between FFTs, defaults to the recommended overlap for the given window (if given), or 0
 window
str
,numpy.ndarray
, optional window function to apply to timeseries prior to FFT, see
scipy.signal.get_window()
for details on acceptable
 fftlength
 Returns:¶
 a Fouriergram
 fill(value)¶
Fill the array with a scalar value.
 Parameters:¶
 valuescalar
All elements of
a
will be assigned this value.
Examples
>>> import numpy as np >>> a = np.array([1, 2]) >>> a.fill(0) >>> a array([0, 0]) >>> a = np.empty(2) >>> a.fill(1) >>> a array([1., 1.])
Fill expects a scalar value and always behaves the same as assigning to a single array element. The following is a rare example where this distinction is important:
>>> a = np.array([None, None], dtype=object) >>> a[0] = np.array(3) >>> a array([array(3), None], dtype=object) >>> a.fill(np.array(3)) >>> a array([array(3), array(3)], dtype=object)
Where other forms of assignments will unpack the array being assigned:
>>> a[...] = np.array(3) >>> a array([3, 3], dtype=object)
 filter(*filt, **kwargs)[source]¶
Filter this
TimeSeries
with an IIR or FIR filter Parameters:¶
 *filtfilter arguments
1, 2, 3, or 4 arguments defining the filter to be applied,
 filtfilt
bool
, optional filter forward and backwards to preserve phase, default:
False
 analog
bool
, optional if
True
, filter coefficients will be converted from Hz to Zdomain digital representation, default:False
 inplace
bool
, optional if
True
, this array will be overwritten with the filtered version, default:False
 unit: `str`
 If zpk, the frequency response units this filter was designed for,
either Hz or rad/s. Default: ‘Hz’ if analog. Rad/s if digital.
 **kwargs
other keyword arguments are passed to the filter method
 Returns:¶
 result
TimeSeries
the filtered version of the input
TimeSeries
 result
 Raises:¶
 ValueError
if
filt
arguments cannot be interpreted properly
See also
scipy.signal.sosfilt
for details on filtering with secondorder sections
scipy.signal.sosfiltfilt
for details on forwardbackward filtering with secondorder sections
scipy.signal.lfilter
for details on filtering (without SOS)
scipy.signal.filtfilt
for details on forwardbackward filtering (without SOS)
Notes
IIR filters are converted into cascading secondorder sections before being applied to this
TimeSeries
.FIR filters are passed directly to
scipy.signal.lfilter()
orscipy.signal.filtfilt()
without any conversions.Examples
We can design an arbitrarily complicated filter using
gwpy.signal.filter_design
>>> from gwpy.signal import filter_design >>> bp = filter_design.bandpass(50, 250, 4096.) >>> notches = [filter_design.notch(f, 4096.) for f in (60, 120, 180)] >>> zpk = filter_design.concatenate_zpks(bp, *notches)
And then can download some data from GWOSC to apply it using
TimeSeries.filter
:>>> from gwpy.timeseries import TimeSeries >>> data = TimeSeries.fetch_open_data('H1', 1126259446, 1126259478) >>> filtered = data.filter(zpk, filtfilt=True)
We can plot the original signal, and the filtered version, cutting off either end of the filtered data to remove filteredge artefacts
>>> from gwpy.plot import Plot >>> plot = Plot(data, filtered[128:128], separate=True) >>> plot.show()
(
png
)

classmethod find(channel, start, end, frametype=
None
, pad=None
, scaled=None
, nproc=1
, verbose=False
, **readargs)[source]¶ Find and read data from frames for a channel
 Parameters:¶
 channel
str
,Channel
the name of the channel to read, or a
Channel
object. start
LIGOTimeGPS
,float
,str
GPS start time of required data, any input parseable by
to_gps
is fine end
LIGOTimeGPS
,float
,str
GPS end time of required data, any input parseable by
to_gps
is fine frametype
str
, optional name of frametype in which this channel is stored, will search for containing frame types if necessary
 nproc
int
, optional, default:1
number of parallel processes to use, serial process by default.
 pad
float
, optional value with which to fill gaps in the source data, by default gaps will result in a
ValueError
. allow_tape
bool
, optional, default:True
allow reading from frame files on (slow) magnetic tape
 verbose
bool
, optional print verbose output about read progress, if
verbose
is specified as a string, this defines the prefix for the progress meter **readargs
any other keyword arguments to be passed to
read()
 channel

find_gates(tzero=
1.0
, whiten=True
, threshold=50.0
, cluster_window=0.5
, **whiten_kwargs)[source]¶ Identify points that should be gates using a provided threshold and clustered within a provided time window.
 Parameters:¶
 tzero
int
, optional halfwidth time duration (seconds) in which the timeseries is set to zero
 whiten
bool
, optional if True, data will be whitened before gating points are discovered, use of this option is highly recommended
 threshold
float
, optional amplitude threshold, if the data exceeds this value a gating window will be placed
 cluster_window
float
, optional time duration (seconds) over which gating points will be clustered
 **whiten_kwargs
other keyword arguments that will be passed to the
TimeSeries.whiten
method if it is being used when discovering gating points
 tzero
 Returns:¶
 out
SegmentList
a list of segments that should be gated based on the provided parameters
 out
See also
TimeSeries.gate
for a method that applies the identified gates

flatten(order=
'C'
)[source]¶ Return a copy of the array collapsed into one dimension.
Any index information is removed as part of the flattening, and the result is returned as a
Quantity
array. Parameters:¶
 order{‘C’, ‘F’, ‘A’, ‘K’}, optional
‘C’ means to flatten in rowmajor (Cstyle) order. ‘F’ means to flatten in columnmajor (Fortran style) order. ‘A’ means to flatten in columnmajor order if
a
is Fortran contiguous in memory, rowmajor order otherwise. ‘K’ means to flattena
in the order the elements occur in memory. The default is ‘C’.
 Returns:¶
 y
Quantity
A copy of the input array, flattened to one dimension.
 y
Examples
>>> a = Array([[1,2], [3,4]], unit='m', name='Test') >>> a.flatten() <Quantity [1., 2., 3., 4.] m>

classmethod from_lal(lalts, copy=
True
)[source]¶ Generate a new TimeSeries from a LAL TimeSeries of any type.

classmethod from_nds2_buffer(buffer_, scaled=
None
, copy=True
, **metadata)[source]¶ Construct a new series from an
nds2.buffer
objectRequires:
NDS2
 Parameters:¶
 buffer_
nds2.buffer
the input NDS2client buffer to read
 scaled
bool
, optional apply slope and bias calibration to ADC data, for nonADC data this option has no effect
 copy
bool
, optional if
True
, copy the contained data array to new to a new array **metadata
any other metadata keyword arguments to pass to the
TimeSeries
constructor
 buffer_
 Returns:¶
 timeseries
TimeSeries
a new
TimeSeries
containing the data from thends2.buffer
, and the appropriate metadata
 timeseries

classmethod from_pycbc(pycbcseries, copy=
True
)[source]¶ Convert a
pycbc.types.timeseries.TimeSeries
into aTimeSeries
 Parameters:¶
 pycbcseries
pycbc.types.timeseries.TimeSeries
the input PyCBC
TimeSeries
array copy
bool
, optional, default:True
if
True
, copy these data to a new array
 pycbcseries
 Returns:¶
 timeseries
TimeSeries
a GWpy version of the input timeseries
 timeseries

gate(tzero=
1.0
, tpad=0.5
, whiten=True
, threshold=50.0
, cluster_window=0.5
, **whiten_kwargs)[source]¶ Removes high amplitude peaks from data using inverse Planck window.
Points will be discovered automatically using a provided threshold and clustered within a provided time window.
 Parameters:¶
 tzero
int
, optional halfwidth time duration (seconds) in which the timeseries is set to zero
 tpad
int
, optional halfwidth time duration (seconds) in which the Planck window is tapered
 whiten
bool
, optional if True, data will be whitened before gating points are discovered, use of this option is highly recommended
 threshold
float
, optional amplitude threshold, if the data exceeds this value a gating window will be placed
 cluster_window
float
, optional time duration (seconds) over which gating points will be clustered
 **whiten_kwargs
other keyword arguments that will be passed to the
TimeSeries.whiten
method if it is being used when discovering gating points
 tzero
 Returns:¶
 out
TimeSeries
a copy of the original
TimeSeries
that has had gating windows applied
 out
See also
TimeSeries.mask
for the method that masks out unwanted data
TimeSeries.find_gates
for the method that identifies gating points
TimeSeries.whiten
for the whitening filter used to identify gating points
Examples
Read data into a
TimeSeries
>>> from gwpy.timeseries import TimeSeries >>> data = TimeSeries.fetch_open_data('H1', 1135148571, 1135148771)
Apply gating using custom arguments
>>> gated = data.gate(tzero=1.0, tpad=1.0, threshold=10.0, fftlength=4, overlap=2, method='median')
Plot the original data and the gated data, whiten both for visualization purposes
>>> overlay = data.whiten(4,2,method='median').plot(dpi=150, label='Ungated', color='dodgerblue', zorder=2) >>> ax = overlay.gca() >>> ax.plot(gated.whiten(4,2,method='median'), label='Gated', color='orange', zorder=3) >>> ax.set_xlim(1135148661, 1135148681) >>> ax.legend() >>> overlay.show()

classmethod get(channel, start, end, pad=
None
, scaled=None
, dtype=None
, verbose=False
, allow_tape=None
, **kwargs)[source]¶ Get data for this channel from frames or NDS
This method dynamically accesses either frames on disk, or a remote NDS2 server to find and return data for the given interval
 Parameters:¶
 channel
str
,Channel
the name of the channel to read, or a
Channel
object. start
LIGOTimeGPS
,float
,str
GPS start time of required data, any input parseable by
to_gps
is fine end
LIGOTimeGPS
,float
,str
GPS end time of required data, any input parseable by
to_gps
is fine pad
float
, optional value with which to fill gaps in the source data, by default gaps will result in a
ValueError
. scaled
bool
, optional apply slope and bias calibration to ADC data, for nonADC data this option has no effect
 nproc
int
, optional, default:1
number of parallel processes to use, serial process by default.
 allow_tape
bool
, optional, default:None
allow the use of frames that are held on tape, default is
None
to attempt to allow theTimeSeries.fetch
method to intelligently select a server that doesn’t use tapes for data storage (doesn’t always work), but to eventually allow retrieving data from tape if required verbose
bool
, optional print verbose output about data access progress, if
verbose
is specified as a string, this defines the prefix for the progress meter **kwargs
other keyword arguments to pass to either
find()
(for direct GWF file access) orfetch()
for remote NDS2 access
 channel
See also
TimeSeries.fetch
for grabbing data from a remote NDS2 server
TimeSeries.find
for discovering and reading data from local GWF files

getfield(dtype, offset=
0
)¶ Returns a field of the given array as a certain type.
A field is a view of the array data with a given datatype. The values in the view are determined by the given type and the offset into the current array in bytes. The offset needs to be such that the view dtype fits in the array dtype; for example an array of dtype complex128 has 16byte elements. If taking a view with a 32bit integer (4 bytes), the offset needs to be between 0 and 12 bytes.
 Parameters:¶
 dtypestr or dtype
The data type of the view. The dtype size of the view can not be larger than that of the array itself.
 offsetint
Number of bytes to skip before beginning the element view.
Examples
>>> import numpy as np >>> x = np.diag([1.+1.j]*2) >>> x[1, 1] = 2 + 4.j >>> x array([[1.+1.j, 0.+0.j], [0.+0.j, 2.+4.j]]) >>> x.getfield(np.float64) array([[1., 0.], [0., 2.]])
By choosing an offset of 8 bytes we can select the complex part of the array for our view:
>>> x.getfield(np.float64, offset=8) array([[1., 0.], [0., 4.]])

heterodyne(phase, stride=
1
, singlesided=False
)[source]¶ Compute the average magnitude and phase of this
TimeSeries
once per stride after heterodyning with a given phase series Parameters:¶
 phase
array_like
an array of phase measurements (radians) with which to heterodyne the signal
 stride
float
, optional stride (seconds) between calculations, defaults to 1 second
 singlesided
bool
, optional Boolean switch to return singlesided output (i.e., to multiply by 2 so that the signal is distributed across positive frequencies only), default: False
 phase
 Returns:¶
 out
TimeSeries
magnitude and phase trends, represented as
mag * exp(1j*phase)
withdt=stride
 out
See also
TimeSeries.demodulate
for a method to heterodyne at a fixed frequency
Notes
This is similar to the
demodulate()
method, but is more general in that it accepts a varying phase evolution, rather than a fixed frequency.Unlike
demodulate()
, the complex output is doublesided by default, so is not multiplied by 2.Examples
Heterodyning can be useful in analysing quasimonochromatic signals with a known phase evolution, such as continuouswave signals from rapidly rotating neutron stars. These sources radiate at a frequency that slowly decreases over time, and is Doppler modulated due to the Earth’s rotational and orbital motion.
To see an example of heterodyning in action, we can simulate a signal whose phase evolution is described by the frequency and its first derivative with respect to time. We can download some O1 era LIGOLivingston data from GWOSC, inject the simulated signal, and recover its amplitude.
>>> from gwpy.timeseries import TimeSeries >>> data = TimeSeries.fetch_open_data('L1', 1131350417, 1131354017)
We now need to set the signal parameters, generate the expected phase evolution, and create the signal:
>>> import numpy >>> f0 = 123.456789 # signal frequency (Hz) >>> fdot = 9.87654321e7 # signal frequency derivative (Hz/s) >>> fepoch = 1131350417 # phase epoch >>> amp = 1.5e22 # signal amplitude >>> phase0 = 0.4 # signal phase at the phase epoch >>> times = data.times.value  fepoch >>> phase = 2 * numpy.pi * (f0 * times + 0.5 * fdot * times**2) >>> signal = TimeSeries(amp * numpy.cos(phase + phase0), >>> sample_rate=data.sample_rate, t0=data.t0) >>> data = data.inject(signal)
To recover the signal, we can bandpass the injected data around the signal frequency, then heterodyne using our phase model with a stride of 60 seconds:
>>> filtdata = data.bandpass(f0  0.5, f0 + 0.5) >>> het = filtdata.heterodyne(phase, stride=60, singlesided=True)
We can then plot signal amplitude over time (cropping the first two minutes to remove the filter response):
>>> plot = het.crop(het.x0.value + 180).abs().plot() >>> ax = plot.gca() >>> ax.set_ylabel("Strain amplitude") >>> plot.show()
(
png
)

highpass(frequency, gpass=
2
, gstop=30
, fstop=None
, type='iir'
, filtfilt=True
, **kwargs)[source]¶ Filter this
TimeSeries
with a highpass filter. Parameters:¶
 frequency
float
highpass corner frequency
 gpass
float
the maximum loss in the passband (dB).
 gstop
float
the minimum attenuation in the stopband (dB).
 fstop
float
stopband edge frequency, defaults to
frequency * 1.5
 type
str
the filter type, either
'iir'
or'fir'
 **kwargs
other keyword arguments are passed to
gwpy.signal.filter_design.highpass()
 frequency
 Returns:¶
 hpseries
TimeSeries
a highpassed version of the input
TimeSeries
 hpseries
See also
gwpy.signal.filter_design.highpass
for details on the filter design
TimeSeries.filter
for details on how the filter is applied
 inject(other)[source]¶
Add two compatible
Series
along their shared xaxis values. Parameters:¶
 other
Series
a
Series
whose xindex intersects withself.xindex
 other
 Returns:¶
 out
Series
the sum of
self
andother
along their shared xaxis values
 out
 Raises:¶
 ValueError
if
self
andother
have incompatible units or xindex intervals
Notes
If
other.xindex
andself.xindex
do not intersect, this method will return a copy ofself
. If the series have uniformly offset indices, this method will raise a warning.If
self.xindex
is an array of timestamps, and ifother.xspan
is not a subset ofself.xspan
, thenother
will be cropped before being adding toself
.Users who wish to taper or window their
Series
should do so before passing it to this method. SeeTimeSeries.taper()
andplanck()
for more information.

insert(obj, values, axis=
None
)¶ Insert values along the given axis before the given indices and return a new
Quantity
object.This is a thin wrapper around the
numpy.insert
function. Parameters:¶
 objint, slice or sequence of int
Object that defines the index or indices before which
values
is inserted. valuesarraylike
Values to insert. If the type of
values
is different from that of quantity,values
is converted to the matching type.values
should be shaped so that it can be broadcast appropriately The unit ofvalues
must be consistent with this quantity. axisint, optional
Axis along which to insert
values
. Ifaxis
is None then the quantity array is flattened before insertion.
 Returns:¶
 out
Quantity
A copy of quantity with
values
inserted. Note that the insertion does not occur inplace: a new quantity array is returned.
 out
Examples
>>> import astropy.units as u >>> q = [1, 2] * u.m >>> q.insert(0, 50 * u.cm) <Quantity [ 0.5, 1., 2.] m>
>>> q = [[1, 2], [3, 4]] * u.m >>> q.insert(1, [10, 20] * u.m, axis=0) <Quantity [[ 1., 2.], [ 10., 20.], [ 3., 4.]] m>
>>> q.insert(1, 10 * u.m, axis=1) <Quantity [[ 1., 10., 2.], [ 3., 10., 4.]] m>
 is_compatible(other)[source]¶
Check whether this series and other have compatible metadata
This method tests that the
sample size
, and theunit
match.

is_contiguous(other, tol=
3.814697265625e06
)[source]¶ Check whether other is contiguous with self.
 Parameters:¶
 other
Series
,numpy.ndarray
another series of the same type to test for contiguity
 tol
float
, optional the numerical tolerance of the test
 other
 Returns:¶
 1
if
other
is contiguous with this series, i.e. would attach seamlessly onto the end
 1
if
other
is anticontiguous with this seires, i.e. would attach seamlessly onto the start
 0
if
other
is completely discontiguous with thie series
Notes
if a raw
numpy.ndarray
is passed as other, with no metadata, then the contiguity check will always pass
 item(*args)¶
Copy an element of an array to a scalar Quantity and return it.
Like
item()
except that it always returns aQuantity
, not a Python scalar.

lowpass(frequency, gpass=
2
, gstop=30
, fstop=None
, type='iir'
, filtfilt=True
, **kwargs)[source]¶ Filter this
TimeSeries
with a Butterworth lowpass filter. Parameters:¶
 frequency
float
lowpass corner frequency
 gpass
float
the maximum loss in the passband (dB).
 gstop
float
the minimum attenuation in the stopband (dB).
 fstop
float
stopband edge frequency, defaults to
frequency * 1.5
 type
str
the filter type, either
'iir'
or'fir'
 **kwargs
other keyword arguments are passed to
gwpy.signal.filter_design.lowpass()
 frequency
 Returns:¶
 lpseries
TimeSeries
a lowpassed version of the input
TimeSeries
 lpseries
See also
gwpy.signal.filter_design.lowpass
for details on the filter design
TimeSeries.filter
for details on how the filter is applied

mask(deadtime=
None
, flag=None
, query_open_data=False
, const=nan
, tpad=0.5
, **kwargs)[source]¶ Mask away portions of this
TimeSeries
that fall within a given list of time segments Parameters:¶
 deadtime
SegmentList
, optional a list of time segments defining the deadtime (i.e., masked portions) of the output, will supersede
flag
if given flag
str
, optional the name of a dataquality flag for which to query, required if
deadtime
is not given query_open_data
bool
, optional if
True
, will query for publicly released dataquality segments through the Gravitationalwave Open Science Center (GWOSC), default:False
 const
float
, optional constant value with which to mask deadtime data, default:
nan
 tpad
float
, optional length of time (in seconds) over which to taper off data at mask segment boundaries, default: 0.5 seconds
 **kwargs
dict
, optional additional keyword arguments to
query
orfetch_open_data
, see “Notes” below
 deadtime
 Returns:¶
 out
TimeSeries
the masked version of this
TimeSeries
 out
See also
gwpy.segments.DataQualityFlag.query
for the method to query segments of a given dataquality flag
gwpy.segments.DataQualityFlag.fetch_open_data
for the method to query dataquality flags from the GWOSC database
gwpy.signal.window.planck
for the generic Plancktaper window
Notes
If
tpad
is nonzero, the Plancktaper window is used to smoothly ramp data down to zero over a timescaletpad
approaching every segment boundary indeadtime
. However, this does not apply to the left or right bounds of the originalTimeSeries
.The
deadtime
segment list will always be coalesced and restricted to the limits ofself.span
. In particular, when querying a dataquality flag, this means thestart
andend
arguments toquery
will effectively be reset and therefore need not be given.If
flag
is interpreted positively, i.e. ifflag
being active corresponds to a “good” state, then its complement inself.span
will be used to define the deadtime for masking.
 max(axis=None, out=None, keepdims=False, initial=<no value>, where=True)¶
Return the maximum along a given axis.
Refer to
numpy.amax
for full documentation.See also
numpy.amax
equivalent function

mean(axis=
None
, dtype=None
, out=None
, keepdims=False
, *, where=True
)¶ Returns the average of the array elements along given axis.
Refer to
numpy.mean
for full documentation.See also
numpy.mean
equivalent function

median(axis=
None
, **kwargs)[source]¶ Compute the median along the specified axis.
Returns the median of the array elements.
 Parameters:¶
 aarray_like
Input array or object that can be converted to an array.
 axis{int, sequence of int, None}, optional
Axis or axes along which the medians are computed. The default, axis=None, will compute the median along a flattened version of the array.
Added in version 1.9.0.
If a sequence of axes, the array is first flattened along the given axes, then the median is computed along the resulting flattened axis.
 outndarray, optional
Alternative output array in which to place the result. It must have the same shape and buffer length as the expected output, but the type (of the output) will be cast if necessary.
 overwrite_inputbool, optional
If True, then allow use of memory of input array
a
for calculations. The input array will be modified by the call tomedian
. This will save memory when you do not need to preserve the contents of the input array. Treat the input as undefined, but it will probably be fully or partially sorted. Default is False. Ifoverwrite_input
isTrue
anda
is not already anndarray
, an error will be raised. keepdimsbool, optional
If this is set to True, the axes which are reduced are left in the result as dimensions with size one. With this option, the result will broadcast correctly against the original
arr
.Added in version 1.9.0.
 Returns:¶
 medianndarray
A new array holding the result. If the input contains integers or floats smaller than
float64
, then the output datatype isnp.float64
. Otherwise, the datatype of the output is the same as that of the input. Ifout
is specified, that array is returned instead.
See also
mean
,percentile
Notes
Given a vector
V
of lengthN
, the median ofV
is the middle value of a sorted copy ofV
,V_sorted
 i e.,V_sorted[(N1)/2]
, whenN
is odd, and the average of the two middle values ofV_sorted
whenN
is even.Examples
>>> import numpy as np >>> a = np.array([[10, 7, 4], [3, 2, 1]]) >>> a array([[10, 7, 4], [ 3, 2, 1]]) >>> np.median(a) np.float64(3.5) >>> np.median(a, axis=0) array([6.5, 4.5, 2.5]) >>> np.median(a, axis=1) array([7., 2.]) >>> np.median(a, axis=(0, 1)) np.float64(3.5) >>> m = np.median(a, axis=0) >>> out = np.zeros_like(m) >>> np.median(a, axis=0, out=m) array([6.5, 4.5, 2.5]) >>> m array([6.5, 4.5, 2.5]) >>> b = a.copy() >>> np.median(b, axis=1, overwrite_input=True) array([7., 2.]) >>> assert not np.all(a==b) >>> b = a.copy() >>> np.median(b, axis=None, overwrite_input=True) np.float64(3.5) >>> assert not np.all(a==b)
 min(axis=None, out=None, keepdims=False, initial=<no value>, where=True)¶
Return the minimum along a given axis.
Refer to
numpy.amin
for full documentation.See also
numpy.amin
equivalent function

nansum(axis=
None
, out=None
, keepdims=False
, *, initial=None
, where=True
)¶ Deprecated since version 5.3: The nansum method is deprecated and may be removed in a future version. Use np.nansum instead.
 nonzero()¶
Return the indices of the elements that are nonzero.
Refer to
numpy.nonzero
for full documentation.See also
numpy.nonzero
equivalent function

notch(frequency, type=
'iir'
, filtfilt=True
, **kwargs)[source]¶ Notch out a frequency in this
TimeSeries
. Parameters:¶
 frequency
float
,Quantity
frequency (default in Hertz) at which to apply the notch
 type
str
, optional type of filter to apply, currently only ‘iir’ is supported
 **kwargs
other keyword arguments to pass to
scipy.signal.iirdesign
 frequency
 Returns:¶
 notched
TimeSeries
a notchfiltered copy of the input
TimeSeries
 notched
See also
TimeSeries.filter
for details on the filtering method
scipy.signal.iirdesign
for details on the IIR filter design method

override_unit(unit, parse_strict=
'raise'
)[source]¶ Forcefully reset the unit of these data
Use of this method is discouraged in favour of
to()
, which performs accurate conversions from one unit to another. The method should really only be used when the original unit of the array is plain wrong.
 pad(pad_width, **kwargs)[source]¶
Pad this series to a new size
See also
numpy.pad
for details on the underlying functionality

partition(kth, axis=
1
, kind='introselect'
, order=None
)¶ Partially sorts the elements in the array in such a way that the value of the element in kth position is in the position it would be in a sorted array. In the output array, all elements smaller than the kth element are located to the left of this element and all equal or greater are located to its right. The ordering of the elements in the two partitions on the either side of the kth element in the output array is undefined.
Added in version 1.8.0.
 Parameters:¶
 kthint or sequence of ints
Element index to partition by. The kth element value will be in its final sorted position and all smaller elements will be moved before it and all equal or greater elements behind it. The order of all elements in the partitions is undefined. If provided with a sequence of kth it will partition all elements indexed by kth of them into their sorted position at once.
Deprecated since version 1.22.0: Passing booleans as index is deprecated.
 axisint, optional
Axis along which to sort. Default is 1, which means sort along the last axis.
 kind{‘introselect’}, optional
Selection algorithm. Default is ‘introselect’.
 orderstr or list of str, optional
When
a
is an array with fields defined, this argument specifies which fields to compare first, second, etc. A single field can be specified as a string, and not all fields need to be specified, but unspecified fields will still be used, in the order in which they come up in the dtype, to break ties.
See also
numpy.partition
Return a partitioned copy of an array.
argpartition
Indirect partition.
sort
Full sort.
Notes
See
np.partition
for notes on the different algorithms.Examples
>>> import numpy as np >>> a = np.array([3, 4, 2, 1]) >>> a.partition(3) >>> a array([2, 1, 3, 4]) # may vary
>>> a.partition((1, 3)) >>> a array([1, 2, 3, 4])

plot(method=
'plot'
, figsize=(12, 4)
, xscale='autogps'
, **kwargs)[source]¶ Plot the data for this timeseries
See also
matplotlib.pyplot.figure
for documentation of keyword arguments used to create the figure
matplotlib.figure.Figure.add_subplot
for documentation of keyword arguments used to create the axes
matplotlib.axes.Axes.plot
for documentation of keyword arguments used in rendering the data

prepend(other, inplace=
True
, pad=None
, gap=None
, resize=True
)[source]¶ Connect another series onto the start of the current one.
 Parameters:¶
 other
Series
another series of the same type as this one
 inplace
bool
, optional perform operation inplace, modifying current series, otherwise copy data and return new series, default:
True
Warning
inplace
prepend bypasses the reference check innumpy.ndarray.resize
, so be carefully to only use this for arrays that haven’t been sharing their memory! pad
float
, optional value with which to pad discontiguous series, by default gaps will result in a
ValueError
. gap
str
, optional action to perform if there’s a gap between the other series and this one. One of
'raise'
 raise aValueError
'ignore'
 remove gap and join data'pad'
 pad gap with zeros
If
pad
is given and is notNone
, the default is'pad'
, otherwise'raise'
. resize
bool
, optional resize this array to accommodate new data, otherwise shift the old data to the left (potentially falling off the start) and put the new data in at the end, default:
True
.
 other
 Returns:¶
 series
TimeSeries
timeseries containing joined data sets
 series

prod(axis=
None
, dtype=None
, out=None
, keepdims=False
, initial=1
, where=True
)¶ Return the product of the array elements over the given axis
Refer to
numpy.prod
for full documentation.See also
numpy.prod
equivalent function

psd(fftlength=
None
, overlap=None
, window='hann'
, method='median'
, **kwargs)[source]¶ Calculate the PSD
FrequencySeries
for thisTimeSeries
 Parameters:¶
 fftlength
float
number of seconds in single FFT, defaults to a single FFT covering the full duration
 overlap
float
, optional number of seconds of overlap between FFTs, defaults to the recommended overlap for the given window (if given), or 0
 window
str
,numpy.ndarray
, optional window function to apply to timeseries prior to FFT, see
scipy.signal.get_window()
for details on acceptable formats method
str
, optional FFTaveraging method (default:
'median'
), see Notes for more details **kwargs
other keyword arguments are passed to the underlying PSDgeneration method
 fftlength
 Returns:¶
 psd
FrequencySeries
a data series containing the PSD.
 psd
Notes
The accepted
method
arguments are:'bartlett'
: a mean average of nonoverlapping periodograms'median'
: a median average of overlapping periodograms'welch'
: a mean average of overlapping periodograms

put(indices, values, mode=
'raise'
)¶ Set
a.flat[n] = values[n]
for alln
in indices.Refer to
numpy.put
for full documentation.See also
numpy.put
equivalent function

q_gram(qrange=
(4, 64)
, frange=(0, inf)
, mismatch=0.2
, snrthresh=5.5
, **kwargs)[source]¶ Scan a
TimeSeries
using the multiQ transform and return anEventTable
of the most significant tiles Parameters:¶
 qrange
tuple
offloat
, optional (low, high)
range of Qs to scan frange
tuple
offloat
, optional (low, high)
range of frequencies to scan mismatch
float
, optional maximum allowed fractional mismatch between neighbouring tiles
 snrthresh
float
, optional lower inclusive threshold on individual tile SNR to keep in the table
 **kwargs
other keyword arguments to be passed to
QTiling.transform()
, including'epoch'
and'search'
 qrange
 Returns:¶
 qgram
EventTable
a table of timefrequency tiles on the most significant
QPlane
 qgram
See also
TimeSeries.q_transform
for a method to interpolate the raw Qtransform over a regularly gridded spectrogram
gwpy.signal.qtransform
for code and documentation on how the Qtransform is implemented
gwpy.table.EventTable.tile
to render this
EventTable
as a collection of polygons
Notes
Only tiles with signal energy greater than or equal to
snrthresh ** 2 / 2
will be stored in the outputEventTable
. The table columns are'time'
,'duration'
,'frequency'
,'bandwidth'
, and'energy'
.

q_transform(qrange=
(4, 64)
, frange=(0, inf)
, gps=None
, search=0.5
, tres='<default>'
, fres='<default>'
, logf=False
, norm='median'
, mismatch=0.2
, outseg=None
, whiten=True
, fduration=2
, highpass=None
, **asd_kw)[source]¶ Scan a
TimeSeries
using the multiQ transform and return an interpolated highresolution spectrogramBy default, this method returns a highresolution spectrogram in both time and frequency, which can result in a large memory footprint. If you know that you only need a subset of the output for, say, a figure, consider using
outseg
and the other keyword arguments to restrict the size of the returned data. Parameters:¶
 qrange
tuple
offloat
, optional (low, high)
range of Qs to scan frange
tuple
offloat
, optional (log, high)
range of frequencies to scan gps
float
, optional central time of interest for determine loudest Qplane
 search
float
, optional window around
gps
in which to find peak energies, only used ifgps
is given tres
float
, optional desired time resolution (seconds) of output
Spectrogram
, default isabs(outseg) / 1000.
 fres
float
,int
,None
, optional desired frequency resolution (Hertz) of output
Spectrogram
, or, iflogf=True
, the number of frequency samples; giveNone
to skip this step and return the original resolution, default is 0.5 Hz or 500 frequency samples logf
bool
, optional boolean switch to enable (
True
) or disable (False
) use of logsampled frequencies in the outputSpectrogram
, ifTrue
thenfres
is interpreted as a number of frequency samples, default:False
 norm
bool
,str
, optional whether to normalize the returned Qtransform output, or how, default:
True
('median'
), other options:False
,'mean'
 mismatch
float
maximum allowed fractional mismatch between neighbouring tiles
 outseg
Segment
, optional GPS
[start, stop)
segment for outputSpectrogram
, default is the full duration of the input whiten
bool
,FrequencySeries
, optional boolean switch to enable (
True
) or disable (False
) data whitening, or an ASDFrequencySeries
with which to whiten the data fduration
float
, optional duration (in seconds) of the timedomain FIR whitening filter, only used if
whiten
is notFalse
, defaults to 2 seconds highpass
float
, optional highpass corner frequency (in Hz) of the FIR whitening filter, used only if
whiten
is notFalse
, default:None
 **asd_kw
keyword arguments to pass to
TimeSeries.asd
to generate an ASD to use when whitening the data
 qrange
 Returns:¶
 out
Spectrogram
output
Spectrogram
of normalised Q energy
 out
See also
TimeSeries.asd
for documentation on acceptable
**asd_kw
TimeSeries.whiten
for documentation on how the whitening is done
gwpy.signal.qtransform
for code and documentation on how the Qtransform is implemented
Notes
This method will return a
Spectrogram
of dtypefloat32
ifnorm
is given, andfloat64
otherwise.To optimize plot rendering with
pcolormesh
, the outputSpectrogram
can be given a logsampled frequency axis by passinglogf=True
at runtime. Thefres
argument is then the number of points on the frequency axis. Note, this is incompatible withimshow
.It is also highly recommended to use the
outseg
keyword argument when only a small window around a given GPS time is of interest. This will speed up this method a little, but can greatly speed up rendering the resultingSpectrogram
usingpcolormesh
.If you aren’t going to use
pcolormesh
in the end, don’t worry.Examples
>>> from numpy.random import normal >>> from scipy.signal import gausspulse >>> from gwpy.timeseries import TimeSeries
Generate a
TimeSeries
containing Gaussian noise sampled at 4096 Hz, centred on GPS time 0, with a sineGaussian pulse (‘glitch’) at 500 Hz:>>> noise = TimeSeries(normal(loc=1, size=4096*4), sample_rate=4096, epoch=2) >>> glitch = TimeSeries(gausspulse(noise.times.value, fc=500) * 4, sample_rate=4096) >>> data = noise + glitch
Compute and plot the Qtransform of these data:
>>> q = data.q_transform() >>> plot = q.plot() >>> ax = plot.gca() >>> ax.set_xlim(.2, .2) >>> ax.set_epoch(0) >>> plot.show()
(
png
)
 ravel([order])¶
Return a flattened array.
Refer to
numpy.ravel
for full documentation.See also
numpy.ravel
equivalent function
ndarray.flat
a flat iterator on the array.

rayleigh_spectrogram(stride, fftlength=
None
, overlap=0
, window='hann'
, nproc=1
, **kwargs)[source]¶ Calculate the Rayleigh statistic spectrogram of this
TimeSeries
 Parameters:¶
 stride
float
number of seconds in single PSD (column of spectrogram).
 fftlength
float
number of seconds in single FFT.
 overlap
float
, optional number of seconds of overlap between FFTs, passing
None
will choose based on the window method, default:0
 window
str
,numpy.ndarray
, optional window function to apply to timeseries prior to FFT, see
scipy.signal.get_window()
for details on acceptable formats nproc
int
, optional maximum number of independent frame reading processes, default default:
1
 stride
 Returns:¶
 spectrogram
Spectrogram
timefrequency Rayleigh spectrogram as generated from the input timeseries.
 spectrogram
See also
TimeSeries.rayleigh
for details of the statistic calculation

rayleigh_spectrum(fftlength=
None
, overlap=0
, window='hann'
)[source]¶ Calculate the Rayleigh
FrequencySeries
for thisTimeSeries
.The Rayleigh statistic is calculated as the ratio of the standard deviation and the mean of a number of periodograms.
 Parameters:¶
 fftlength
float
number of seconds in single FFT, defaults to a single FFT covering the full duration
 overlap
float
, optional number of seconds of overlap between FFTs, passing
None
will choose based on the window method, default:0
 window
str
,numpy.ndarray
, optional window function to apply to timeseries prior to FFT, see
scipy.signal.get_window()
for details on acceptable formats
 fftlength
 Returns:¶
 psd
FrequencySeries
a data series containing the PSD.
 psd
 classmethod read(source, *args, **kwargs)[source]¶
Read data into a
TimeSeries
Arguments and keywords depend on the output format, see the online documentation for full details for each format, the parameters below are common to most formats.
 Parameters:¶
 source
str
,list
Source of data, any of the following:
 name
str
,Channel
the name of the channel to read, or a
Channel
object. start
LIGOTimeGPS
,float
,str
, optional GPS start time of required data, defaults to start of data found; any input parseable by
to_gps
is fine end
LIGOTimeGPS
,float
,str
, optional GPS end time of required data, defaults to end of data found; any input parseable by
to_gps
is fine format
str
, optional source format identifier. If not given, the format will be detected if possible. See below for list of acceptable formats.
 nproc
int
, optional number of parallel processes to use, serial process by default.
 pad
float
, optional value with which to fill gaps in the source data, by default gaps will result in a
ValueError
.
 source
 Raises:¶
 IndexError
if
source
is an empty list
Notes
The available builtin formats are:
Format
Read
Write
Autoidentify
csv
Yes
Yes
Yes
gwf
Yes
Yes
Yes
gwf.framecpp
Yes
Yes
No
gwf.framel
Yes
Yes
No
gwf.lalframe
Yes
Yes
No
hdf5
Yes
Yes
Yes
hdf5.gwosc
Yes
No
No
txt
Yes
Yes
Yes
wav
Yes
No
No

repeat(repeats, axis=
None
)¶ Repeat elements of an array.
Refer to
numpy.repeat
for full documentation.See also
numpy.repeat
equivalent function

resample(rate, window=
'hamming'
, ftype='fir'
, n=None
)[source]¶ Resample this Series to a new rate
 Parameters:¶
 rate
float
rate to which to resample this
Series
 window
str
,numpy.ndarray
, optional window function to apply to signal in the Fourier domain, see
scipy.signal.get_window()
for details on acceptable formats, only used forftype='fir'
or irregular downsampling ftype
str
, optional type of filter, either ‘fir’ or ‘iir’, defaults to ‘fir’
 n
int
, optional if
ftype='fir'
the number of taps in the filter, otherwise the order of the Chebyshev type I IIR filter
 rate
 Returns:¶
 Series
a new Series with the resampling applied, and the same metadata

reshape(shape, /, *, order=
'C'
, copy=None
)¶ Returns an array containing the same data with a new shape.
Refer to
numpy.reshape
for full documentation.See also
numpy.reshape
equivalent function
Notes
Unlike the free function
numpy.reshape
, this method onndarray
allows the elements of the shape parameter to be passed in as separate arguments. For example,a.reshape(10, 11)
is equivalent toa.reshape((10, 11))
.

resize(new_shape, refcheck=
True
)¶ Change shape and size of array inplace.
 Parameters:¶
 new_shapetuple of ints, or
n
ints Shape of resized array.
 refcheckbool, optional
If False, reference count will not be checked. Default is True.
 new_shapetuple of ints, or
 Returns:¶
 None
 Raises:¶
 ValueError
If
a
does not own its own data or references or views to it exist, and the data memory must be changed. PyPy only: will always raise if the data memory must be changed, since there is no reliable way to determine if references or views to it exist. SystemError
If the
order
keyword argument is specified. This behaviour is a bug in NumPy.
See also
resize
Return a new array with the specified shape.
Notes
This reallocates space for the data area if necessary.
Only contiguous arrays (data elements consecutive in memory) can be resized.
The purpose of the reference count check is to make sure you do not use this array as a buffer for another Python object and then reallocate the memory. However, reference counts can increase in other ways so if you are sure that you have not shared the memory for this array with another Python object, then you may safely set
refcheck
to False.Examples
Shrinking an array: array is flattened (in the order that the data are stored in memory), resized, and reshaped:
>>> import numpy as np
>>> a = np.array([[0, 1], [2, 3]], order='C') >>> a.resize((2, 1)) >>> a array([[0], [1]])
>>> a = np.array([[0, 1], [2, 3]], order='F') >>> a.resize((2, 1)) >>> a array([[0], [2]])
Enlarging an array: as above, but missing entries are filled with zeros:
>>> b = np.array([[0, 1], [2, 3]]) >>> b.resize(2, 3) # new_shape parameter doesn't have to be a tuple >>> b array([[0, 1, 2], [3, 0, 0]])
Referencing an array prevents resizing…
>>> c = a >>> a.resize((1, 1)) Traceback (most recent call last): ... ValueError: cannot resize an array that references or is referenced ...
Unless
refcheck
is False:>>> a.resize((1, 1), refcheck=False) >>> a array([[0]]) >>> c array([[0]])

rms(stride=
1
)[source]¶ Calculate the rootmeansquare value of this
TimeSeries
once per stride. Parameters:¶
 stride
float
stride (seconds) between RMS calculations
 stride
 Returns:¶
 rms
TimeSeries
a new
TimeSeries
containing the RMS value with dt=stride
 rms

round(decimals=
0
, out=None
)¶ Return
a
with each element rounded to the given number of decimals.Refer to
numpy.around
for full documentation.See also
numpy.around
equivalent function

searchsorted(v, side=
'left'
, sorter=None
)¶ Find indices where elements of v should be inserted in a to maintain order.
For full documentation, see
numpy.searchsorted
See also
numpy.searchsorted
equivalent function

setfield(val, dtype, offset=
0
)¶ Put a value into a specified place in a field defined by a datatype.
Place
val
intoa
’s field defined bydtype
and beginningoffset
bytes into the field. Parameters:¶
 valobject
Value to be placed in field.
 dtypedtype object
Datatype of the field in which to place
val
. offsetint, optional
The number of bytes into the field at which to place
val
.
 Returns:¶
 None
See also
Examples
>>> import numpy as np >>> x = np.eye(3) >>> x.getfield(np.float64) array([[1., 0., 0.], [0., 1., 0.], [0., 0., 1.]]) >>> x.setfield(3, np.int32) >>> x.getfield(np.int32) array([[3, 3, 3], [3, 3, 3], [3, 3, 3]], dtype=int32) >>> x array([[1.0e+000, 1.5e323, 1.5e323], [1.5e323, 1.0e+000, 1.5e323], [1.5e323, 1.5e323, 1.0e+000]]) >>> x.setfield(np.eye(3), np.int32) >>> x array([[1., 0., 0.], [0., 1., 0.], [0., 0., 1.]])

setflags(write=
None
, align=None
, uic=None
)¶ Set array flags WRITEABLE, ALIGNED, WRITEBACKIFCOPY, respectively.
These Booleanvalued flags affect how numpy interprets the memory area used by
a
(see Notes below). The ALIGNED flag can only be set to True if the data is actually aligned according to the type. The WRITEBACKIFCOPY flag can never be set to True. The flag WRITEABLE can only be set to True if the array owns its own memory, or the ultimate owner of the memory exposes a writeable buffer interface, or is a string. (The exception for string is made so that unpickling can be done without copying memory.) Parameters:¶
 writebool, optional
Describes whether or not
a
can be written to. alignbool, optional
Describes whether or not
a
is aligned properly for its type. uicbool, optional
Describes whether or not
a
is a copy of another “base” array.
Notes
Array flags provide information about how the memory area used for the array is to be interpreted. There are 7 Boolean flags in use, only three of which can be changed by the user: WRITEBACKIFCOPY, WRITEABLE, and ALIGNED.
WRITEABLE (W) the data area can be written to;
ALIGNED (A) the data and strides are aligned appropriately for the hardware (as determined by the compiler);
WRITEBACKIFCOPY (X) this array is a copy of some other array (referenced by .base). When the CAPI function PyArray_ResolveWritebackIfCopy is called, the base array will be updated with the contents of this array.
All flags can be accessed using the single (upper case) letter as well as the full name.
Examples
>>> import numpy as np >>> y = np.array([[3, 1, 7], ... [2, 0, 0], ... [8, 5, 9]]) >>> y array([[3, 1, 7], [2, 0, 0], [8, 5, 9]]) >>> y.flags C_CONTIGUOUS : True F_CONTIGUOUS : False OWNDATA : True WRITEABLE : True ALIGNED : True WRITEBACKIFCOPY : False >>> y.setflags(write=0, align=0) >>> y.flags C_CONTIGUOUS : True F_CONTIGUOUS : False OWNDATA : True WRITEABLE : False ALIGNED : False WRITEBACKIFCOPY : False >>> y.setflags(uic=1) Traceback (most recent call last): File "<stdin>", line 1, in <module> ValueError: cannot set WRITEBACKIFCOPY flag to True
 shift(delta)[source]¶
Shift this
Series
forward on the Xaxis bydelta
This modifies the series inplace.
 Parameters:¶
Examples
>>> from gwpy.types import Series >>> a = Series([1, 2, 3, 4, 5], x0=0, dx=1, xunit='m') >>> print(a.x0) 0.0 m >>> a.shift(5) >>> print(a.x0) 5.0 m >>> a.shift('1 km') 995.0 m

sort(axis=
1
, kind=None
, order=None
)¶ Sort an array inplace. Refer to
numpy.sort
for full documentation. Parameters:¶
 axisint, optional
Axis along which to sort. Default is 1, which means sort along the last axis.
 kind{‘quicksort’, ‘mergesort’, ‘heapsort’, ‘stable’}, optional
Sorting algorithm. The default is ‘quicksort’. Note that both ‘stable’ and ‘mergesort’ use timsort under the covers and, in general, the actual implementation will vary with datatype. The ‘mergesort’ option is retained for backwards compatibility.
Changed in version 1.15.0: The ‘stable’ option was added.
 orderstr or list of str, optional
When
a
is an array with fields defined, this argument specifies which fields to compare first, second, etc. A single field can be specified as a string, and not all fields need be specified, but unspecified fields will still be used, in the order in which they come up in the dtype, to break ties.
See also
numpy.sort
Return a sorted copy of an array.
numpy.argsort
Indirect sort.
numpy.lexsort
Indirect stable sort on multiple keys.
numpy.searchsorted
Find elements in sorted array.
numpy.partition
Partial sort.
Notes
See
numpy.sort
for notes on the different sorting algorithms.Examples
>>> import numpy as np >>> a = np.array([[1,4], [3,1]]) >>> a.sort(axis=1) >>> a array([[1, 4], [1, 3]]) >>> a.sort(axis=0) >>> a array([[1, 3], [1, 4]])
Use the
order
keyword to specify a field to use when sorting a structured array:>>> a = np.array([('a', 2), ('c', 1)], dtype=[('x', 'S1'), ('y', int)]) >>> a.sort(order='y') >>> a array([(b'c', 1), (b'a', 2)], dtype=[('x', 'S1'), ('y', '<i8')])

spectral_variance(stride, fftlength=
None
, overlap=None
, method='median'
, window='hann'
, nproc=1
, filter=None
, bins=None
, low=None
, high=None
, nbins=500
, log=False
, norm=False
, density=False
)[source]¶ Calculate the
SpectralVariance
of thisTimeSeries
. Parameters:¶
 stride
float
number of seconds in single PSD (column of spectrogram)
 fftlength
float
number of seconds in single FFT
 method
str
, optional FFTaveraging method (default:
'median'
), see Notes for more details overlap
float
, optional number of seconds of overlap between FFTs, defaults to the recommended overlap for the given window (if given), or 0
 window
str
,numpy.ndarray
, optional window function to apply to timeseries prior to FFT, see
scipy.signal.get_window()
for details on acceptable formats nproc
int
maximum number of independent frame reading processes, default is set to singleprocess file reading.
 bins
numpy.ndarray
, optional, defaultNone
array of histogram bin edges, including the rightmost edge
 low
float
, optional left edge of lowest amplitude bin, only read if
bins
is not given high
float
, optional right edge of highest amplitude bin, only read if
bins
is not given nbins
int
, optional number of bins to generate, only read if
bins
is not given log
bool
, optional calculate amplitude bins over a logarithmic scale, only read if
bins
is not given norm
bool
, optional normalise bin counts to a unit sum
 density
bool
, optional normalise bin counts to a unit integral
 stride
 Returns:¶
 specvar
SpectralVariance
2Darray of spectral frequencyamplitude counts
 specvar
See also
numpy.histogram
for details on specifying bins and weights
Notes
The accepted
method
arguments are:'bartlett'
: a mean average of nonoverlapping periodograms'median'
: a median average of overlapping periodograms'welch'
: a mean average of overlapping periodograms

spectrogram(stride, fftlength=
None
, overlap=None
, window='hann'
, method='median'
, nproc=1
, **kwargs)[source]¶ Calculate the average power spectrogram of this
TimeSeries
using the specified average spectrum method.Each timebin of the output
Spectrogram
is calculated by taking a chunk of theTimeSeries
in the segment[t  overlap/2., t + stride + overlap/2.)
and calculating thepsd()
of those data.As a result, each timebin is calculated using
stride + overlap
seconds of data. Parameters:¶
 stride
float
number of seconds in single PSD (column of spectrogram).
 fftlength
float
number of seconds in single FFT.
 overlap
float
, optional number of seconds of overlap between FFTs, defaults to the recommended overlap for the given window (if given), or 0
 window
str
,numpy.ndarray
, optional window function to apply to timeseries prior to FFT, see
scipy.signal.get_window()
for details on acceptable formats method
str
, optional FFTaveraging method (default:
'median'
), see Notes for more details nproc
int
number of CPUs to use in parallel processing of FFTs
 stride
 Returns:¶
 spectrogram
Spectrogram
timefrequency power spectrogram as generated from the input timeseries.
 spectrogram
Notes
The accepted
method
arguments are:'bartlett'
: a mean average of nonoverlapping periodograms'median'
: a median average of overlapping periodograms'welch'
: a mean average of overlapping periodograms

spectrogram2(fftlength, overlap=
None
, window='hann'
, **kwargs)[source]¶ Calculate the nonaveraged power
Spectrogram
of thisTimeSeries
 Parameters:¶
 fftlength
float
number of seconds in single FFT.
 overlap
float
, optional number of seconds of overlap between FFTs, defaults to the recommended overlap for the given window (if given), or 0
 window
str
,numpy.ndarray
, optional window function to apply to timeseries prior to FFT, see
scipy.signal.get_window()
for details on acceptable formats scaling[ ‘density’  ‘spectrum’ ], optional
selects between computing the power spectral density (‘density’) where the
Spectrogram
has units of V**2/Hz if the input is measured in V and computing the power spectrum (‘spectrum’) where theSpectrogram
has units of V**2 if the input is measured in V. Defaults to ‘density’. **kwargs
other parameters to be passed to
scipy.signal.periodogram
for each column of theSpectrogram
 fftlength
 Returns:¶
 spectrogram:
Spectrogram
a power
Spectrogram
with1/fftlength
frequency resolution and (fftlength  overlap) time resolution.
 spectrogram:
See also
scipy.signal.periodogram
for documentation on the Fourier methods used in this calculation
Notes
This method calculates overlapping periodograms for all possible chunks of data entirely containing within the span of the input
TimeSeries
, then normalises the power in overlapping chunks using a triangular window centred on that chunk which most overlaps the givenSpectrogram
time sample.

squeeze(axis=
None
)¶ Remove axes of length one from
a
.Refer to
numpy.squeeze
for full documentation.See also
numpy.squeeze
equivalent function

std(axis=
None
, dtype=None
, out=None
, ddof=0
, keepdims=False
, *, where=True
)¶ Returns the standard deviation of the array elements along given axis.
Refer to
numpy.std
for full documentation.See also
numpy.std
equivalent function

sum(axis=
None
, dtype=None
, out=None
, keepdims=False
, initial=0
, where=True
)¶ Return the sum of the array elements over the given axis.
Refer to
numpy.sum
for full documentation.See also
numpy.sum
equivalent function
 swapaxes(axis1, axis2)¶
Return a view of the array with
axis1
andaxis2
interchanged.Refer to
numpy.swapaxes
for full documentation.See also
numpy.swapaxes
equivalent function

take(indices, axis=
None
, out=None
, mode='raise'
)¶ Return an array formed from the elements of
a
at the given indices.Refer to
numpy.take
for full documentation.See also
numpy.take
equivalent function

taper(side=
'leftright'
, duration=None
, nsamples=None
)[source]¶ Taper the ends of this
TimeSeries
smoothly to zero. Parameters:¶
 side
str
, optional the side of the
TimeSeries
to taper, must be one of'left'
,'right'
, or'leftright'
 duration
float
, optional the duration of time to taper, will override
nsamples
if both are provided as arguments nsamples
int
, optional the number of samples to taper, will be overridden by
duration
if both are provided as arguments
 side
 Returns:¶
 out
TimeSeries
a copy of
self
tapered at one or both ends
 out
 Raises:¶
 ValueError
if
side
is not one of('left', 'right', 'leftright')
Notes
The
TimeSeries.taper()
automatically tapers from the second stationary point (local maximum or minimum) on the specified side of the input. However, the method will never taper more than half the full width of theTimeSeries
, and will fail if there are no stationary points.See
planck()
for the generic Planck taper window, and seescipy.signal.get_window()
for other common window formats.Examples
To see the effect of the Plancktaper window, we can taper a sinusoidal
TimeSeries
at both ends:>>> import numpy >>> from gwpy.timeseries import TimeSeries >>> t = numpy.linspace(0, 1, 2048) >>> series = TimeSeries(numpy.cos(10.5*numpy.pi*t), times=t) >>> tapered = series.taper()
We can plot it to see how the ends now vary smoothly from 0 to 1:
>>> from gwpy.plot import Plot >>> plot = Plot(series, tapered, separate=True, sharex=True) >>> plot.show()
(
png
)

to(unit, equivalencies=
[]
, copy=True
)¶ Return a new
Quantity
object with the specified unit. Parameters:¶
 unitunitlike
An object that represents the unit to convert to. Must be an
UnitBase
object or a string parseable by theunits
package. equivalencieslist of tuple
A list of equivalence pairs to try if the units are not directly convertible. See Equivalencies. If not provided or
[]
, class default equivalencies will be used (none forQuantity
, but may be set for subclasses) IfNone
, no equivalencies will be applied at all, not even any set globally or within a context. copybool, optional
If
True
(default), then the value is copied. Otherwise, a copy will only be made if necessary.
See also
to_value
get the numerical value in a given unit.
 to_device()¶
 to_lal()[source]¶
Convert this
TimeSeries
into a LAL TimeSeries.Note
This operation always copies data to new memory.

to_pycbc(copy=
True
)[source]¶ Convert this
TimeSeries
into a PyCBCTimeSeries
 Parameters:¶
 Returns:¶
 timeseries
TimeSeries
a PyCBC representation of this
TimeSeries
 timeseries

to_string(unit=
None
, precision=None
, format=None
, subfmt=None
)¶ Generate a string representation of the quantity and its unit.
The behavior of this function can be altered via the
numpy.set_printoptions
function and its various keywords. The exception to this is thethreshold
keyword, which is controlled via the[units.quantity]
configuration itemlatex_array_threshold
. This is treated separately because the numpy default of 1000 is too big for most browsers to handle. Parameters:¶
 unitunitlike, optional
Specifies the unit. If not provided, the unit used to initialize the quantity will be used.
 precisionnumber, optional
The level of decimal precision. If
None
, or not provided, it will be determined from NumPy print options. formatstr, optional
The format of the result. If not provided, an unadorned string is returned. Supported values are:
‘latex’: Return a LaTeXformatted string
‘latex_inline’: Return a LaTeXformatted string that uses negative exponents instead of fractions
 subfmtstr, optional
Subformat of the result. For the moment, only used for
format='latex'
andformat='latex_inline'
. Supported values are:‘inline’: Use
$ ... $
as delimiters.‘display’: Use
$\displaystyle ... $
as delimiters.
 Returns:¶
 str
A string with the contents of this Quantity

to_value(unit=
None
, equivalencies=[]
)¶ The numerical value, possibly in a different unit.
 Parameters:¶
 unitunitlike, optional
The unit in which the value should be given. If not given or
None
, use the current unit. equivalencieslist of tuple, optional
A list of equivalence pairs to try if the units are not directly convertible (see Equivalencies). If not provided or
[]
, class default equivalencies will be used (none forQuantity
, but may be set for subclasses). IfNone
, no equivalencies will be applied at all, not even any set globally or within a context.
 Returns:¶
 valuendarray or scalar
The value in the units specified. For arrays, this will be a view of the data if no unit conversion was necessary.
See also
to
Get a new instance in a different unit.