MATLAB is a programming and numeric computing platform used to analyze data, develop algorithms, and create models. This article explains how to get an authorization token in MATLAB for Azure Data Explorer, and how to use the token to interact with your cluster.
Select the tab for the operating system used to run MATLAB.
Install JDK 8 and Maven.
Clone the MATLAB Interface for Azure Services repository:
git clone https://github.com/mathworks-ref-arch/matlab-azure-services.git
Build the MATLAB Azure SDK jar:
cd matlab-azure-services/Software/Java
mvn clean package
The build should produce the file: Software/MATLAB/lib/jar/azure-common-sdk-0.2.0.jar.
In the MATLAB studio, load the jar file and add it to the java static class path:
edit(fullfile(prefdir,'javaclasspath.txt'));
In java static class path file, add an entry corresponding to the jar file:
(full-path)/matlab-azure-services/Software/MATLAB/lib/jar/azure-common-sdk-0.2.0.jar
Restart the MATLAB IDE. Once reloaded, run the startup.m script available at matlab-azure-services\Software\MATLAB\startup.m. This ensures all the prerequisite functions are set up for access to Azure services. The output should look as follows:
>> startup
Adding Azure Paths
------------------
Adding: matlab-azure-services/Software/MATLAB/app
Adding: matlab-azure-services/Software/MATLAB/app/functions
Adding: matlab-azure-services/Software/MATLAB/app/system
Adding: matlab-azure-services/Software/MATLAB/lib
Adding: matlab-azure-services/Software/MATLAB/config
Skipping: matlab-azure-services/Software/Utilities
Checking the static Java classpath for: matlab-azure-services/Software/MATLAB/lib/jar/azure-common-sdk-0.2.0.jar
Found: azure-common-sdk-0.2.0.jar
With user authentication, the user is prompted to sign-in through a browser window. Upon successful sign-in, a user authorization token is granted. This section shows how to configure this interactive sign-in flow.
To perform user authentication:
Define the constants needed for the authorization. For more information about these values, see Authentication parameters.
% The Azure Data Explorer cluster URL
clusterUrl = 'https://<adx-cluster>.kusto.chinacloudapi.cn';
% The Azure AD tenant ID
tenantId = '';
% Send a request to https://<adx-cluster>.kusto.chinacloudapi.cn/v1/rest/auth/metadata
% The appId should be the value of KustoClientAppId
appId = '';
% The Azure AD scopes
scopesToUse = strcat(clusterUrl,'/.default ');
In MATLAB studio, load the extracted DLL files:
% Access the folder that contains the DLL files
dllFolder = fullfile("C:","Matlab","DLL");
% Load the referenced assemblies in the MATLAB session
matlabDllFiles = dir(fullfile(dllFolder,'*.dll'));
for k = 1:length(matlabDllFiles)
baseFileName = matlabDllFiles(k).name;
fullFileName = fullfile(dllFolder,baseFileName);
fprintf(1, 'Reading %s\n', fullFileName);
end
% Load the downloaded assembly in MATLAB
NET.addAssembly(fullFileName);
Use the PublicClientApplicationBuilder to prompt a user interactive sign-in:
% Create an PublicClientApplicationBuilder
app = Microsoft.Identity.Client.PublicClientApplicationBuilder.Create(appId)...
.WithAuthority(Microsoft.Identity.Client.AzureCloudInstance.AzurePublic,tenantId)...
.WithRedirectUri('http://localhost:8675')...
.Build();
% System.Net.ServicePointManager.SecurityProtocol = System.Net.SecurityProtocolType.Tls12;
NET.setStaticProperty ('System.Net.ServicePointManager.SecurityProtocol',System.Net.SecurityProtocolType.Tls12)
% Start with creating a list of scopes
scopes = NET.createGeneric('System.Collections.Generic.List',{'System.String'});
% Add the actual scopes
scopes.Add(scopesToUse);
fprintf(1, 'Using appScope %s\n', scopesToUse);
% Get the token from the service
% and show the interactive dialog in which the user can login
tokenAcquirer = app.AcquireTokenInteractive(scopes);
result = tokenAcquirer.ExecuteAsync;
% Extract the token and when it expires
% and retrieve the returned token
token = char(result.Result.AccessToken);
fprintf(2, 'User token aquired and will expire at %s & extended expires at %s', result.Result.ExpiresOn.LocalDateTime.ToString,result.Result.ExtendedExpiresOn.ToLocalTime.ToString);
Use the authorization token to query your cluster through the REST API:
options=weboptions('HeaderFields',{'RequestMethod','POST';'Accept' 'application/json';'Authorization' ['Bearer ', token]; 'Content-Type' 'application/json; charset=utf-8'; 'Connection' 'Keep-Alive'; 'x-ms-app' 'Matlab'; 'x-ms-client-request-id' 'Matlab-Query-Request'});
% The DB and KQL variables represent the database and query to execute
querydata = struct('db', "<DB>", 'csl', "<KQL>");
querryresults = webwrite("https://sdktestcluster.chinaeast2.dev.kusto.chinacloudapi.cn/v2/rest/query", querydata, options);
% Extract the results row
results=querryresults{3}.Rows
Create an Auth.json file in your working directory with the relevant credentials. The KustoClientAppId returned from https://<adx-cluster>.kusto.chinacloudapi.cn/v1/rest/auth/metadata needs to be replaced with the value of ClientId below:
{
"AuthMethod": "InteractiveBrowser",
"TenantId" : "<AAD Tenant / Authority ID for the login>",
"ClientId" : "<Client ID>",
"RedirectUrl": "http://localhost:8675"
}
Query your cluster as follows:
% References the credential file created in the previous step
credentials = configureCredentials('Auth.json');
% Point the scopes to the Azure Data Explorer cluster that we want to query
request = azure.core.credential.TokenRequestContext();
request.addScopes('<https://adx-cluster-changeme.kusto.chinacloudapi.cn/.default>');
token=credentials.getToken(request);
% Prepare to query the cluster
options=weboptions('HeaderFields',{'RequestMethod','POST';'Accept' 'application/json';'Authorization' ['Bearer ', token.getToken()]; 'Content-Type' 'application/json; charset=utf-8'; 'Connection' 'Keep-Alive'; 'x-ms-app' 'Matlab'; 'x-ms-client-request-id' 'Matlab-Query-Request'});
% The DB and KQL variables represent the database and query to execute
querydata = struct('db', "<DB>", 'csl', "<KQL>");
querryresults = webwrite('<https://adx-cluster-changeme.kusto.chinacloudapi.cn/v2/rest/query>', querydata, options);
% The results row can be extracted as follows
results=querryresults{3}.Rows
Microsoft Entra application authorization can be used for scenarios where interactive sign-in isn't desired and automated runs are necessary.
To perform application authentication:
Provision a Microsoft Entra application. For the Redirect URI, select Web and input http://localhost:8675 as the URI.
Define the constants needed for the authorization. For more information about these values, see Authentication parameters.
% The Azure Data Explorer cluster URL
clusterUrl = 'https://<adx-cluster>.kusto.chinacloudapi.cn';
% The Azure AD tenant ID
tenantId = '';
% The Azure AD application ID and key
appId = '';
appSecret = '';
In MATLAB studio, load the extracted DLL files:
% Access the folder that contains the DLL files
dllFolder = fullfile("C:","Matlab","DLL");
% Load the referenced assemblies in the MATLAB session
matlabDllFiles = dir(fullfile(dllFolder,'*.dll'));
for k = 1:length(matlabDllFiles)
baseFileName = matlabDllFiles(k).name;
fullFileName = fullfile(dllFolder,baseFileName);
fprintf(1, 'Reading %s\n', fullFileName);
end
% Load the downloaded assembly
NET.addAssembly(fullFileName);
Use the ConfidentialClientApplicationBuilder to perform a non-interactive automated sign-in with the Microsoft Entra application:
% Create an ConfidentialClientApplicationBuilder
app = Microsoft.Identity.Client.ConfidentialClientApplicationBuilder.Create(appId)...
.WithAuthority(Microsoft.Identity.Client.AzureCloudInstance.AzurePublic,tenantId)...
.WithRedirectUri('http://localhost:8675')...
.WithClientSecret(appSecret)...
.Build();
% System.Net.ServicePointManager.SecurityProtocol = System.Net.SecurityProtocolType.Tls12;
NET.setStaticProperty ('System.Net.ServicePointManager.SecurityProtocol',System.Net.SecurityProtocolType.Tls12)
% Start with creating a list of scopes
scopes = NET.createGeneric('System.Collections.Generic.List',{'System.String'});
% Add the actual scopes
scopes.Add(scopesToUse);
fprintf(1, 'Using appScope %s\n', scopesToUse);
% Get the token from the service and cache it until it expires
tokenAcquirer = app.AcquireTokenForClient(scopes);
result = tokenAcquirer.ExecuteAsync;
% Extract the token and when it expires
% retrieve the returned token
token = char(result.Result.AccessToken);
fprintf(2, 'User token aquired and will expire at %s & extended expires at %s', result.Result.ExpiresOn.LocalDateTime.ToString,result.Result.ExtendedExpiresOn.ToLocalTime.ToString);
Use the authorization token to query your cluster through the REST API:
options=weboptions('HeaderFields',{'RequestMethod','POST';'Accept' 'application/json';'Authorization' ['Bearer ', token]; 'Content-Type' 'application/json; charset=utf-8'; 'Connection' 'Keep-Alive'; 'x-ms-app' 'Matlab'; 'x-ms-client-request-id' 'Matlab-Query-Request'});
% The DB and KQL variables represent the database and query to execute
querydata = struct('db', "<DB>", 'csl', "<KQL>");
querryresults = webwrite("https://sdktestcluster.chinaeast2.dev.kusto.chinacloudapi.cn/v2/rest/query", querydata, options);
% Extract the results row
results=querryresults{3}.Rows
Create an Auth.json file in your working directory with the relevant credentials:
{
"AuthMethod": "ClientSecret",
"TenantId" : "<AAD Tenant / Authority ID for the login>",
"ClientId" : "<Client ID of the Azure AD application>",
"ClientSecret": "<Client Key of the Azure AD application>",
"RedirectUrl": "http://localhost:8675"
}
Query your cluster as follows:
% References the credential file created in the previous step
credentials = configureCredentials('Auth.json');
% Point the scopes to the Azure Data Explorer cluster that we want to query
request = azure.core.credential.TokenRequestContext();request.addScopes('<https://adx-cluster-changeme.kusto.chinacloudapi.cn/.default>');
token=credentials.getToken(request);
% Prepare to query the cluster
options=weboptions('HeaderFields',{'RequestMethod','POST';'Accept' 'application/json';'Authorization' ['Bearer ', token.getToken()]; 'Content-Type' 'application/json; charset=utf-8'; 'Connection' 'Keep-Alive'; 'x-ms-app' 'Matlab'; 'x-ms-client-request-id' 'Matlab-Query-Request'});
% The DB and KQL variables represent the database and query to execute
querydata = struct('db', "<DB>", 'csl', "<KQL>");
querryresults = webwrite('<https://adx-cluster-changeme.kusto.chinacloudapi.cn/v2/rest/query>', querydata, options);
% The results row can be extracted as follows
results=querryresults{3}.Rows