What Can You Do with the Model Manager?
When working with a simulation model for an extended period of time, you will inevitably have a need to keep a backup of old versions. You may, for example, want to recover an older version in case your current modeling work goes astray, or perhaps you want to use an older version as a template for a completely new model. Your solution to this may vary from something as straightforward as saving files with different filenames on your local hard drive, saving to a file-based version control system, or by uploading files to a product lifecycle management (PLM) system provided by your organization. If you are working on models in a collaborative setting, you are also used to sharing files with your coworkers, for example via email, by placing them on a shared file system, or by allowing your coworkers to download them from a centralized version control system or PLM system.
As the amount of simulation models and data grows, you and your coworkers might find yourselves spending a large part of your time managing these models and data. This may involve working with multiple tools and software — keeping you away from your main modeling and simulation work. Some of the challenges and concerns you might face are:
•
Efficient storage of models. Store only relevant data for the purpose of archiving and future retrieval, all while keeping disk space usage manageable.
•
Automatic extraction of model metadata. Extract metadata when saving simulation models so that you and others may later find them, without requiring manual data entry for keywords and other search terms.
•
Reuse of models. Use previously saved simulation models and data as building blocks when creating new models, perhaps generating an extensive library of reusable parts.
•
Track and compare changes to models over time. Get an automatic audit trail for your simulation models to compare and restore older versions, or to reproduce modeling steps.
•
Manage relationships between models and data files. Find out which simulation models use a particular data file as input or what simulation model generated a particular data file as output.
•
Access control. Control who can find, open, and save simulation models and data.
The Model Manager comes with a set of tools for addressing these points — all while staying within the COMSOL Desktop modeling environment. From the COMSOL Desktop, you can create a new database on your own computer to keep track of your private models and data files, or you and your colleagues may share such models and files by uploading them to a server database accessed via a Model Manager server.
A Model Manager database is tailor-made for the storage needs of a model built in COMSOL Multiphysics®. The Model Manager makes sure to never store duplicates of simulation data when saving multiple versions of the same model. You can also avoid storing built, computed, and plotted data that may instead be reproduced from the model when needed.
The powerful Model Manager search syntax enables you to search deep into models based on their properties, features, settings, and other metadata. You can, for example, perform search queries answering:
•
Which models use a Time Dependent study step?
•
Which models have a Length parameter between 5 cm and 15 cm?
•
Which models were last saved by me?
The Model Manager also comes with standard version control tools such as viewing the version history of models and data files, automatically detecting version conflicts when saving, and comparing versions with each other. You can, for example, open an older model version to create a completely new model with its own split-off version history, or see all the changes made to a model from one version to the next.
You can do exploratory work on an existing model in a database by creating a draft of the model. The draft is version controlled in its own right, enabling you to experiment with various simulation ideas without polluting the version history of the original model. Once you have finished your draft work, you can choose to either keep it as a new version of the original model or discard it.
More advanced version control tools such as branching, merging, and reverting are also available. Branching enables you, for example, to work on an entire collection of models and data files in isolation, while at the same time postponing the decision whether or not your changes are worth preserving. Reverting enables you, for example, to restore models and data files that you have previously deleted, perhaps by accident.
The Model Manager also comes with an application programming interface (API) for use with the Java® programming language. The Model Manager API enables you to easily perform repetitive tasks by writing and running a few lines of code from the COMSOL Desktop. The API also enables you to perform more advanced database operations that would otherwise be difficult to do using built-in Model Manager tools.
When using a server database accessed via a Model Manager server, you can control who can access models and data files by setting permissions. You can, for example, set which users are permitted to open or save a particular model, or set which users are permitted to search and browse a collection of models. You can also use the web-based asset management system included with a Model Manager server installation to link your simulation models and results to various documents, presentations, project notes, slides, and other supplementary files and metadata — all while keeping everything in the same database storing your models. Using the asset management system is also a simple way of collaborating on simulation projects with people in your organization who may not have access to the COMSOL Multiphysics software. Simulation engineers can, for example, share output files from simulation runs by exporting them to the Model Manager server, while other engineers can upload new versions of data files via the Model Manager server web interface — versions which are then immediately available to the simulation engineer from the COMSOL Desktop.