Feature Request: `embed_sampleset` or similar

It would be useful sometimes to be able to "embed" a sampleset. Something like
```python
import numpy as np

def embed_sampleset(sampleset, embedding, *, num_qubits=None, fill_value=2):

if num_qubits is None:
num_qubits = max(max(chain) for chain in embedding.values()) + 1

num_samples = len(sampleset)

samples = np.full((num_samples, num_qubits), fill_value)

for i, v in enumerate(sampleset.variables):
for c in embedding[v]:
samples[:, c] = sampleset.record.sample[:, i]

return samples

if __name__ == '__main__':
import dimod

from dwave.system import DWaveSampler, EmbeddingComposite

bqm = dimod.generators.gnm_random_bqm(100, 1000, "SPIN")

sampleset = EmbeddingComposite(DWaveSampler()).sample(bqm, return_embedding=True, num_reads=100, answer_mode="raw")

embedding = sampleset.info["embedding_context"]["embedding"]

unembedded_samples = embed_sampleset(sampleset, embedding)
```