xuml-populate 0.5.0

Populates the SM Metamodel schema with a user model and outputs this as a TclRAL *.ral database

Blueprint MBSE Repository Populator

This package transforms human readable text files describing an Executable UML model of a system into a populated metamodel database. With your models loaded into this database it is now possible to produce a variety of useful artifacts to support model execution and verification, code generation and anything else that performs detailed analyis on the model components and their relationships.

The text files are expressed in an easy to read markdown format so you can browse through the classes, relationships, states and transitions, actions and all other model components.

Here we support the Shlaer-Mellor variant of Executable UML exclusively.

Command usage

% modeldb -s elevator-case-study

Assuming your system is named 'elevator-case-study' and is in the current working directory with the internal structure defined below, will output a file named mmdb_elevator-case-study.ral. This *.ral file is a text file serialization of a TclRAL database.

Your user model is loaded into the Shlaer-Mellor Metamodel and, if there are no errors, you know you have a Shlaer-Mellor Executable Model that doesn't break any of the rules defined by the metamodel.

You can load and view it using TclRAL or PyRAL or feed it to other downstream Blueprint tools such as the Model Executor which among its tasks will generate a user model database.

Input to the populator

Each system is defined in a single package broken down into standard hierarchy of folders like so:

system
    domain
        subsystem
            class-model
                classmodel.xcm
            types.yaml
            methods
                m1.mtd
                m2.mtd
                ...
            state-machines
                s1.xsm
                ...
            external
                EE
                    op1.op
                    ...
        subsystem2
        ...
    domain2
    ...

Here is a partial layout for The Elevator Case Study as an example:

elevator-case-study // system
    elevator-management // application domain
        elevator // All defined in one subsystem
            class-model
                elevator.xcm // the class model
            methods // methods for all classes in subsystem
                cabin // methods on 'cabin' class
                    ping.mtd // the ping method
                    ...
                ...
            state-machines // lifecycles and assigners for this subsystem
                cabin.xsm // lifecycles named by class, assigners by association
                transfer.xsm
                R53.xsm // assigner state machine on association R53
                ...
            external // external entities, each a proxy for some class
                CABIN // proxy for 'cabin' class
                    arrived-at-floor.op // two ee operations
                    goto-floor.op
        types.yaml // data types for all subsystems in domain
    transport // two more domains (not broken down yet)
    signal io

Each modeled domain has its own folder. Above we just see one for the Elevator Managment domain.

Each domain requires at least one subsystem folder. Here we see only one and that is the Elevator domain.

Within a subsystem folder there is a class-model subfolder with one class model expressed as an .xcm (executable class model) file.

The following folders are optional:

• external – external entities and their operations, one subfolder per external entity
• methods – class methods each in a folder matching the class name with each method in a separate .mtd file
• state-machines – each state machine, assigner or lifecycle, in its own .xsm (executable state machine) file

Also within a domain you have a types.yaml file which specifies each domain specific type (Pressure, Speed, etc) and selects a corresponding system (database) type. This is a stop gap measure as we have not yet provided a more robust typing domain, so, for now we settle with what our database has to offer (int, string, float, etc). Unltimately, though, a full featured typing facility will support a variety of types and operations on those types as well as a type definition system. Note that the typing facility can be, but need not necessarily be a modeled domain.