This is a major release that has reworked the structure of both the attack graph and instance model which means that the older attack graphs and models may not be compatible with the current version.
- Instance Model and Attack Graph formats updated
- Use ids as primary identifiers for both assets and attack steps instead of names or qualified names
- Rename 'metaconcept' to 'type'
- Remove 'firstSteps' attack step node from the attack graph
- Split attack graph into attack steps and attackers sections
- Add `extras` field to assets, associations, and AttackGraphNodes
- Remove AttackGraphNodes `observations` and `reward` fields these should now just be part of `extras`
- Add mal-toolbox version to the model metadata to determine whether an older format is provided
- Add old format translator that can load older instance model formats (Note: Still requires an update for the latest changes and will not currently work)
- Assets and attack steps are generally serialized as dictionaries using the id as the key but they also contain the name in the value for increased legibility
- Save/load to/from file API more consistent between Model and AttackGraph
- Support both yaml for saving/loading Model/AttackGraph (Note: The default is now yaml, but you can change this in the `default.conf` file)
- Type hinting
- Make type hinting uniform and have it cover most of the code (the compiler is not currently covered)
- Added mypy checks to pipeline
- Optimizations
- Look up asset names and ids in sets instead of lists
- Find objects by name/id using dictionaries instead of lists
- Only run expensive expressions in debug statements if run log level is set to DEBUG
- Raise exceptions when duplicate associations are added to instance model
- Improve logging and error handling
- Wider use of exceptions instead of logging and returning None
- Make logging more uniform to always list the asset/attack step/attacker name in double quotes and the id in brackets(e.g. `"Test App"(1)`)
- Use logger formatter instead of f-string when logging
- More attack graph and instance model tests.
- Various bugfixes