A framework for automating the Manifold Boundary Approximation Method. This includes reparameterizing the model using templates stored in MongoDB as well as several algorithms for evaluating variable limits (singular limits).
This project contains an Engine, which attempts to approximate a model automatically, as well as a User-Interface, which allows the user more freedom when approximating her models.
import mbam
Load json model
model = mbam.engine(model_path, data_path)
model.run() # runs the geodesic
limits = [{"p1": "inf"}]
model.apply_limits(limits)The user interface is used to give more control over the approximation step. It is run with a command in the terminal, followed by opening up the corresponding HTML page.
To run locally, run the command in the temrinal
python main.py
If successful, the terminal should show
Ready for Connection
Once the terminal is ready for a connection, open ./static/mbamui.html with your favorite browser.
Any errors or problems will be reported to the terminal.
If you want to run the UI from a server, run the main.py script on the server. On your local machine run the following command.
ssh -N -f -L localhost:9000:localhost:9000 user@server_address
Note: port 9000 is hard coded into the main.py as well as in the mbamui.html. Feel free to adjust the port as necessary.
When finished, run this command on your local machine to close the port.
sudo kill `sudo lsof -t -i:9000`
Link to Thesis (Not online yet)
This implementation uses MongoDB. The installation guide can be found here!
On windows, I suggest installing the beta version of the Linux bash shell on Windows. Here's a great guide I have used to install it.
This version was created and tested on Python 3.6. The dependencies include:
pip install -r requirements.txt
Install Julia v1.0
Pkg.add("ZMQ")
Pkg.add("Parameters")
Pkg.add("HDF5")
Pkg.add("JSON")
Pkg.add("https://pulsar.byu.edu/Modeling/Geometry.jl.git")
Pkg.add("https://pulsar.byu.edu/Modeling/Models.jl.git")
Pkg.add("https://pulsar.byu.edu/Modeling/ParametricModels.jl.git")In addition, some recommended packages that may be useful for constructing models
Pkg.add("https://pulsar.byu.edu/Modeling/SmoothApproximations.jl.git")
Pkg.add("https://pulsar.byu.edu/Modeling/ModularLM.jl.git")
Pkg.add("https://pulsar.byu.edu/Modeling/ExampleModels.jl.git")The docs can be found here.
The current tests can be run by using the Engine and the engine and the user interface to approximate the models found under 'examples'.
Feel free to fork the repo and make any changes for you modeling pleasure!
MIT License