def compress(self, nsamples=None):
"""Reduce the number of samples by discarding low-weights.
Parameters
----------
neff: int, optional
effective number of samples after compression. If not supplied,
then reduce to the channel capacity (theoretical optimum
compression). If <=0, then compress so that all weights are unity.
"""
i = compress_weights(self.weights, self._rand, nsamples)
return self.repeat(i)
def compress(self, nsamples=None):
"""Reduce the number of samples by discarding low-weights.
Parameters
----------
neff: int, optional
effective number of samples after compression. If not supplied,
then reduce to the channel capacity (theoretical optimum
compression). If <=0, then compress so that all weights are unity.
"""
i = compress_weights(self.weights, self._rand, nsamples)
data = np.repeat(self.values, i, axis=0)
index = self.index.repeat(i)
df = DataFrame(data=data, index=index, columns=self.columns)
df.index = df.index.get_level_values('#')
return df
def compress(self, nsamples=None):
"""Reduce the number of samples by discarding low-weights.
Parameters
----------
neff: int, optional
effective number of samples after compression. If not supplied,
then reduce to the channel capacity (theoretical optimum
compression). If <=0, then compress so that all weights are unity.
"""
i = compress_weights(self.weight, self._rand, nsamples)
data = numpy.repeat(self.values, i, axis=0)
index = numpy.repeat(self.index.values, i)
df = pandas.DataFrame(data=data, index=index, columns=self.columns)
if 'weight' in self:
return df.drop(columns='weight')
else:
return df