This PR does two things

• It removes the check for subgroup inclusion when getting a signature from bytes, since the signature were already verified.
• It aggregates attestations directly when they are a single bit attestation, there is no need to run the max coverage algo

These numbers were run on a NUC i5-8259U

On 293 aggregations, the average on mainnet were

• at 7 seconds: 1905 ms. (standard deviation 310ms or 17%)
• at 9 seconds: 142 ms. (standard deviation 119ms)
• at 11 seconds: 17ms. (standard deviation 16ms)

variation happened up to 3500ms on the first batch regularly (hence the high standard deviation).

At 9 seconds it got regularly > 1000 ms and both second and third batches had standard deviation of 100%.

mean(field-1)	median(field-1)	perc:90(field-1)	perc:95(field-1)	perc:99(field-1)
1896.361774744	1892	2087.2	2292.8	3320

With the PR applied we obtain

• 7 seconds: 789ms (std deviation 206ms)
• 9 seconds: 86ms (std deviation 74 ms)
• 11 seconds: 9ms (std deviation 7ms)

Variation happened up to 2256ms in a couple of places.

mean(field-1)	median(field-1)	perc:90(field-1)	perc:95(field-1)	perc:99(field-1)
794.21164021164	765	854	1055.5	2130.58