netgovern/ai repository overview

This container provides a RESTful API for interacting with the AI service.

To use the AI service you need to specify two parameters:

• MAX_HEAP: maximum size of the JVM heap space
• ml.properties: this properties file contains the API key and secrets for Microsoft Azure or Amazon AWS

Docker command to start the container: docker run -d -t -e MAX_HEAP=2g -v /path/to/ml.properties:/etc/config/ml.properties netgovern/ai:latest

By default the AI service will use the port 8089. If the port is not available you can use: docker run -d -t -p :8089 -e MAX_HEAP=2g -v /path/to/ml.properties:/home/netmail/ml.properties netgovern/ai:latest

⁠============================== Default template for the ml.properties file

server.properties=8089 #select either Sagemaker or Azure as cloud service cloud.service.class=com.netmail.ml.azure.AzureProvider #cloud.service.class=com.netmail.ml.sagemaker.SagemakerMlService

############################################

⁠configuration for Hadoop

############################################ #path to winutils.exe file on windows hadoop.winutils=/

############################################

⁠configuration for AWS

############################################ aws.key.id=<AWS_KEY_ID> aws.key.secret=<AWS_SECRET> aws.bucket=sagemaker-us-netmail aws.region=us-east-1 nbThreads=5

#spark configuration spark.master=local[4] #spark.master=spark://<SPARK_MASTER_IP>:7077

############################################

⁠configuration for MS Azure

############################################ azure.account.name=ngmlworkspacestorage #key for the storage account azure.storage.key=<AZURE_STORAGE_KEY> #key for machine learning models azure.ml.account.key=<AZURE_ML_KEY> azure.endpoint.suffix=core.windows.net azure.default.endpoints.protocol=https azure.workspace=<AZURE_WORKSPACE_ID> azure.subscriptionId=<AZURE_SUBSCRIPTION_ID>