Batch size in WorkerSettings

We define `batch_size` in the `WorkerSettings`, which is initialized using the `sample_batch_size` argument from neps.run(). However, at runtime, although multiple trials are sampled according to batch_size, only the first one is actually used.

Start by tracing how neps.run() passes sample_batch_size into WorkerSettings.batch_size, then follow the runtime path that samples and selects trials. Verify where the sampled batch is reduced to its first trial. Done means all trials sampled according to batch_size are used during runtime, with behavior checked against the existing execution flow.

Written by the indexing model from the issue text.