[Feature, Refactor] Iterable replay buffers

[vmoens]

Description

Implements an __iter__ method for replay buffer.
The goal is to be able to go through a replay buffer with a for loop.
The sampler without repetition should lead to a finite loop, and the drop-last argument should match the behaviour of dataloaders.

TODO:

• Implement tests of the feature
  • combinations of replay buffers with various samplers
  • check that the behaviour of the sampler without repetition matches what's expected
• Document the feature