The NeuroML 2 toolbox for MATLAB works by facilitating access to the Java NeuroML 2 API functionality directly within Matlab.
Install jNeuroML following the instructions at https://github.com/NeuroML/jNeuroML. The quickest route is to go to https://github.com/NeuroML/jNeuroML/releases, download the latest jNeuroML.zip file, and unzip somewhere suitable on your machine.
Run Matlab and run the prefdir command to find the location of your preferences folder.
Create a file javaclasspath.txt within that folder containing, on a single line,
the full path to the jNeuroML-<version>-jar-with-dependencies.jar file you unpacked above.
For example, on my Mac the preferences folder is
/Users/name/Library/Application Support/MathWorks/MATLAB/R2016b
and so I created /Users/name/Library/Application Support/MathWorks/MATLAB/R2016b/javaclasspath.txt containing
/Users/name/projects/jNeuroMLJar/jNeuroML-0.7.7-jar-with-dependencies.jar
The specific paths will vary on your machine.
Now restart Matlab and it will know how to find the jNeuroML classes.
You can test that it is working by trying to validate a NeuroML file:
import org.neuroml.model.util.NeuroML2Validator
file = java.io.File('/full/path/to/model.nml');
validator = NeuroML2Validator();
validator.validateWithTests(file);
disp(validator.getValidity())There are further examples of usage in examples/run_examples.m.