Decouple TritonInferenceStage from pipeline mode

[dagardner-nv]

Description

• Add needs_logits & inout_mapping a constructor arguments to TritonInferenceStage.
• Consolidate triton inference worker impls into a single TritonInferenceWorker class
• For compatibility the needs_logits & inout_mapping arguments default to default values inferred by the pipeline mode.
• Refactor custom triton inference stage from examples/log_parsing example as a subclass of TritonInferenceStage removing several lines of redundant code.
• examples/log_parsing example now works in C++ mode
• Add the ability to subclass a stage registered with the CLI, and register the sublcass with a new CLI name.
• Add the ability to parse dictionary types with click.
• Remove inf-triton from the AE mode in the CLI
• Misc linting/formatting pylint suggestions

Closes #1378

By Submitting this PR I confirm:

• I am familiar with the Contributing Guidelines.
• When the PR is ready for review, new or existing tests cover these changes.
• When the PR is ready for review, the documentation is up to date with these changes.

[mdemoret-nv]

Make the following changes:

• Remove AE support from the triton stage all together
• Remove all implementations of the worker class
• Make all options of the worker class instance values (need_logits, and inout_mapping)
• Set defaults for inout_mapping similar to default for need_logits (i.e. by mode if not supplied)

[mdemoret-nv]

Lets get these changes merged but make an issue before you do to try and remove the inference worker class entirely.

[dagardner-nv]

/merge