Skip to main content

MATLAB SDK

Windows Installation​

info

The OpenDataDSL MATLAB SDK requires a java SDK version of at least 11 - here are instructions from MathWorks on how to upgrade your version of java.

Follow these instructions to install the matlab client and configure it in MATLAB

  1. Download the matlab-client zip file
  2. Unzip the installation to a directory on your local drive
  3. Edit or create the file %USER_HOME%\AppData\Roaming\MathWorks\MATLAB\R2023b\javaclasspath.txt
  4. Add an entry in there that points to the odsl-sdk.jar file in the lib directory of the unzipped matlab-client

Example ( if you extracted the zip to c:\ ):

c:\matlab-client\lib\odsl-sdk.jar

  1. Start or restart MATLAB

Getting started​

Connecting​

To connect to OpenDataDSL, you need to create an instance of the sdk.ODSL class and then call login() as follows:

Interactive login​

% Connect
ODSL=sdk.ODSL();
ODSL.login();

Using a secret​

% Connect
ODSL=sdk.ODSL();
ODSL.loginWithSecret("tenant id", "client id", "secret");

Getting data​

To read from any of the ODSL services, you need to use the get method on the ODSL sdk with the following syntax:

var = odsl.get(service, source, id);

e.g. to the timeseries for the EUR/GBP FX rate from ABN Amro

fx = odsl.get('data', 'public', '#ABN_FX.EURGBP:SPOT');

Updating data​

To write to any of the ODSL services, you need to use the update method on the ODSL sdk with the following syntax:

odsl.update(service, source, Var, [options]);

e.g. to update some Master Data using the object service:

AAA = VarDynamicObject('AAA');
odsl.update('object', 'private', AAA);

Building curves​

We have added support for building curves externally using any language. Curves can be built either as:

  • Event triggered using Smart Curves
  • Manual or time triggered using Standard Curves

Event triggered curves​

Event triggered curves are Smart Curves that send a message to the external system to build the curve.

info

When you create a SMART curve that will be built externally, you need to set the caching type as External

The message is recommended to be a Batch variable with the following elements:

  • The Smart Curve configuration
  • The ondate as a String
  • Each curve input

An example ODSL curve function​

function triggerMatlab(BASE)
    inputs = Batch()
    inputs.add(#CURVE)
    inputs.add(#ONDATE)
    inputs.add(BASE)
    send inputs to "matlab"
    triggerMatlab = Message(406, "Triggered MATLAB curve build")    
end

Example ODSL function with multiple inputs​

function triggerMatlab(BASE, OTHER)
    inputs = Batch()
    inputs.add(#CURVE)
    inputs.add(#ONDATE)
    inputs.add(BASE)
    inputs.add(OTHER)
    send inputs to "matlab"
    triggerMatlab = Message(406, "Triggered MATLAB curve build")    
end

Example MATLAB script​

The following example waits for a new message, builds a curve and sends it to ODSL.

Connect to ODSL​

import sdk.*;
import com.opendatadsl.odsl.var.curve.*;

% Connect
odsl = ODSL();
odsl.setStage("dev");
odsl.login();

Wait for a message and create a curve shell​

% Wait for a message
messaging = odsl.messaging();
build = messaging.waitForCurveBuildMessage('matlab');

% Start Building
curve = build.createCurve();
logger = build.getLogger();
logger.info("Starting Curve Build");

Get input data​

% Get our input BASE curve
base = build.get("BASE");

Perform curve building logic​

% Curve Building Logic
% Use the input data and add contracts to the curve using
% curve.add(tenor, value);
% e.g. curve.add("M01", 25.1);

% Example Curve Building Logic - Add 10% to all the values in the BASE curve
for i=0:base.getContracts().size()-1
    vc = base.getContracts().get(i);
    newval = vc.getDoubleValue() * 1.1;
    vc.setDoubleValue(newval);
    curve.add(vc);
end

Save the curve and log information to ODSL​

% save the curve
logger.info("Saving Curve");
disp(['Saving Curve: ', char(build.getFullCurveId())]);
build.saveCurveAndComplete(curve, 'Built in MATLAB');