On 'Kogito', we don't have a real WID parser. Historically, William wrote a toy implementation to demo resource content API and we never invested the time to properly write one.

Now this affects the execution of real v7 projects. Some blocker issues that a real parser would fix:

• If you put an empty wid or malformed wid, breaks BPMN editor
• if you put some types that are not supported, breaks BPMN editor (we have an example with 4 wids types that breaks BPMN for some reason)

The goal of this task is to create a proper wid parser implementation

Acceptance Criteria

1. WID parser should parse all supported parameters:
   • name
   • displayName
   • icon
   • parameters
   • results
   • category
   • documentation
2. WID parser shouldn’t fail on unknown/unsupported/engine only parameters see KOGITO-4372 for more details.
3. All supported by engine MVEL types should be converted to Java types during WID loading:
   • BooleanDataType -> Boolean
   • EnumDataType -> java.lang.Object
   • FloatDataType -> Float
   • IntegerDataType -> Integer
   • ListDataType -> java.util.List
   • ObjectDataType -> java.lang.Object
   • StringDataType -> String
   • UndefinedDataType -> java.lang.Object
4. Users' custom WID files can be placed under global directory as well as any package under src/main/resources directory (but same with bpmn file).
5. It should be possible to add an icon path inside of the global directory and set up an icon in base 64 format.