Misc DFP Documentation & fixes

[dagardner-nv]

• Fills in examples/digital_fingerprinting/production/README.md with a bit more detail
• Fix handling of command line enum values
• Add date window filtering to DFPFileBatcherStage, since this stage is already performing date calculations, and occurs prior to the downloading of any remote data. Previously this was a part of the S3SourceStage.
• Add a --start_time flag to dfp_azure_pipeline.py and dfp_duo_pipeline.py allowing an explicit time window.
• morpheus_training service in docker compose renamed to morpheus_pipeline to reflect that it is used for both training and inference
• Set nvidia runtime in docker compose yaml for users who don't have it as their default runtime.
• MLflow version restricted to <1.29 to avoid a bug (per @pdmack )
• Fix casing of class names in headings for Sphinx builds
• Spelling fixes in docs & help strings
• Fix launch script for morpheus_pipeline
• from-azure and from-duo added to CLI
• Remove usage of only_new_batches argument which no longer exists
• CI work-around for https://gitlab.com/karolinepauls/pytest-kafka/-/issues/10
• Update diagrams to reflect recent code changes
• Support _ as a time separator in iso_date_regex
• Adds the following to docs/source/developer_guide/guides/5_digital_fingerprinting.md:
  • Section on how to define a new data source schema
  • Explanation of Starter & Production examples
  • docker-compose commands for building & starting services in the production example
  • Helm Chart info for production example
  • Fix headings for stage classes

Fixes #345

[dagardner-nv]

rerun tests

[dagardner-nv]

/rerun tests

[dagardner-nv]

\rerun tests

[dagardner-nv]

/rerun tests

[drobison00]

@gpucibot merge