[pkg/ottl] adapt mapGetter to handle nested map items within slices

[bacherfl]

Description

This PR fixes the limitations described in #37405. The recently introduced ValueExpression had to be adapted, and will now only return raw types, instead of their pcommon.Map/pcommon.Slice eqivalents, as suggested in #37280 (comment)

Link to tracking issue

Fixes #37405

Testing

Adapted and extended the unit and e2e tests

[bacherfl]

Converting back to draft for now, due to #37280 (comment)

[edmocosta]

If we end up stick to this approach, considering the mapGetter is now returning pcommon.Map values, I think we could keep it consistent and remove these changes, so all maps would still be parsed aspcommon.Map instead of raw values. I guess it would be a breaking change, though. WDYT?

[bacherfl]

I'd be fine with that, however there are also arguments for returning the raw types here (#37280 (comment)). I think I would also prefer to consistently return pcommon.Map values though. @evan-bradley @TylerHelmuth WDYT?

[evan-bradley]

Judging by #37280 (comment), we probably need to survey this a bit more to make sure we're either consistent everywhere, or can translate at any given point in the chain. I'm not sure I have a strong opinion here, but it sounds like at a minimum we could potentially use more tests.

Do the same issues apply to slices with []any vs. pcommon.Slice?

[bacherfl]

for slices, this issue doesn't apply, as the listGetter returns []any as opposed to pcommon.Slice:

opentelemetry-collector-contrib/pkg/ottl/expression.go

Lines 164 to 176 in ee2d165

	func (l *listGetter[K]) Get(ctx context.Context, tCtx K) (any, error) {
	evaluated := make([]any, len(l.slice))
	for i, v := range l.slice {
	val, err := v.Get(ctx, tCtx)
	if err != nil {
	return nil, err
	}
	evaluated[i] = val
	}
	return evaluated, nil
	}

[TylerHelmuth]

I agree with @edmocosta that we should continue returning pcommon.Map

[bacherfl]

Alright, I now changed it again to return pcommon.Map

[atoulme]

This looks about ready to merge. I think it'd be great to make sure @evan-bradley can review one more time.

[edmocosta]

Before merging it, I think we still need to address this #37408 (comment), if @evan-bradley agrees, we would need to undo those changes to keep returning pcommon.Map.

To to handle #37408 (comment)

[atoulme]

@evan-bradley please review and merge if ok?