Fix NestedProcessingTransformation

[r0ot]

When using a custom pipeline written in Yaml like so

pipeline_classes = [ecs_windows(), ecs_zeek_corelight()]
pipeline_names = [i.name for i in pipeline_classes if i.name]
ProcessingPipelineResolver.from_pipeline_list(pipeline_classes).resolve(pipeline_names + ['custom_pipeline.yml'])

using NestedProcessingTransformation like in custom_pipeline.yml

name: Custom
priority: 100
transformations:
  - type: nest
    items:
      - type: field_name_mapping
        rule_conditions:
          - type: logsource
            service: signinlogs
        mapping:
            ResultType: azure.signinlogs.result_type

yields the following error

File .../python3.12/site-packages/sigma/processing/transformations.py:1021, in NestedProcessingTransformation.__post_init__(self)
-> 1021 self._nested_pipeline = ProcessingPipeline(items=self.items)

File .../python3.12/site-packages/sigma/processing/pipeline.py:433, in ProcessingPipeline.__post_init__(self)
    432 if not all((isinstance(item, ProcessingItem) for item in self.items)):
--> 433     raise TypeError(
    434         "Each item in a processing pipeline must be a ProcessingItem - don't use processing classes directly!"
    435     )
    436 if not all(
    437     (isinstance(item, QueryPostprocessingItem) for item in self.postprocessing_items)
    438 ):

TypeError: Each item in a processing pipeline must be a ProcessingItem - don't use processing classes directly!

So I propose the following fix. Somehow it appears NestedProcessingTransformation isn't getting instantiated with its from_dict method which would've assured each of its items were of type ProcessingItem. Remedying this would be another fix for this issue. There's a chance the way I'm using my custom yaml pipeline is incorrect (ProcessingPipelineResolver.from_pipeline_list(...).resolve(...)) in which case please correct me