Merge pull request #49009 from penggu/patch-1

Update linux-dsvm-walkthrough.md

Author: ktoliver

articles/machine-learning/data-science-virtual-machine/linux-dsvm-walkthrough.md

@@ -184,6 +184,8 @@ To deploy the decision tree code from the preceding section, sign in to Azure Ma
    ![The Azure Machine Learning Studio (classic) primary authorization token](./media/linux-dsvm-walkthrough/workspace-token.png)
 1. Load the **AzureML** package, and then set values of the variables with your token and workspace ID in your R session on the DSVM:
 
+        if(!require("devtools")) install.packages("devtools")
+        devtools::install_github("RevolutionAnalytics/AzureML")
         if(!require("AzureML")) install.packages("AzureML")
         require(AzureML)
         wsAuth = "<authorization-token>"
@@ -203,9 +205,23 @@ To deploy the decision tree code from the preceding section, sign in to Azure Ma
         return(colnames(predictDF)[apply(predictDF, 1, which.max)])
         }
 
+1. Create a settings.json file for this workspace:
+
+        vim ~/.azureml/settings.json
+
+1. Make sure the following contents are put inside settings.json:
+
+         {"workspace":{
+           "id": "<workspace-id>",
+           "authorization_token": "<authorization-token>",
+           "api_endpoint": "https://studioapi.azureml.net",
+           "management_endpoint": "https://management.azureml.net"
+         }
+
 
 1. Publish the **predictSpam** function to AzureML by using the **publishWebService** function:
 
+        ws <- workspace()
         spamWebService <- publishWebService(ws, fun = predictSpam, name="spamWebService", inputSchema = smallTrainSet, data.frame=TRUE)
 
 1. This function takes the **predictSpam** function, creates a web service named **spamWebService** that has defined inputs and outputs, and then returns information about the new endpoint.