onetimesetup.md

Create a Watson ML Service Instance

IBM Watson Machine Learning (WML) Service enables you to create, train, and deploy self-learning models using an automated, collaborative workflow.

Below pm-20 is the name of the Watson ML Service. You create a new instance and give it a name, e.g., myuseridMLinstance1. Below, replace <CLI_WMLi> by your chosen name.

$ bx service create pm-20 standard <CLI_WMLi>

Note: Please ignore warnings regarding payment.

Create an Access key for accessing your Watson ML service instance

Below, replace <key_CLI_WMLi> by a key name of your choice.

$ bx service key-create <CLI_WMLi> <key_CLI_WMLi>

Retrieve credentials:

$ instance_id=`bx service key-show <CLI_WMLi> <key_CLI_WMLi> | grep "instance_id"| awk -F": " '{print $2}'| cut -d'"' -f2`
$ username=`bx service key-show <CLI_WMLi> <key_CLI_WMLi> | grep "username"| awk -F": " '{print $2}'| cut -d'"' -f2`
$ password=`bx service key-show <CLI_WMLi> <key_CLI_WMLi> | grep "password"| awk -F": " '{print $2}'| cut -d'"' -f2`

$ echo ""; echo "ML Instance Credentials:"; echo "instance_id: $instance_id"; echo "username: $username "; echo "password: $password"; echo ""

Set up Environment Variables:

$ export ML_INSTANCE=$instance_id
$ export ML_USERNAME=$username
$ export ML_PASSWORD=$password

Create a bucket in the Cloud Object Storage (COS) to store data

IBM Cloud Object Storage is an unstructured data storage service designed for durability, resiliency and security. A bucket is a huge "folder" in the COS. You use the bucket to put and get any file or folder (e.g., your datasets).

Create a cloud storage instance:

Lets create a personal cloud storage instance to hold your bucket(s) and name the instance <my_COS_instance>. The service-instance-create command below creates the COS instance, and the service-instance command retrieves its attributes.

$ bx resource service-instance-create <my_COS_instance> cloud-object-storage standard 
$ bx resource service-instance <my_COS_instance>

Get security credentials:

Now create and get the credentials to access my_COS_instance. Give a name to your credentials (replace <my_COS_key> below).

Create key, store it and print it:

$ bx resource service-key-create <my_COS_key> Writer --instance-name <my_COS_instance> --parameters '{"HMAC":true}' > /dev/null 2>&1
$ access_key_id=`bx resource service-key <my_COS_key> | grep "access_key_id"| cut -d\:  -f2`
$ secret_access_key=`bx resource service-key <my_COS_key> | grep "secret_access_key"| cut -d\:  -f2`
$ echo ""; echo "COS Credentials:"; echo "access_key_id: $access_key_id"; echo "secret_access_key: - $secret_access_key"; echo ""

Save your keys to shell variables. (write down the keys as you'll need them again later to access your resources.)

export MY_BUCKET_KEY=$access_key_id
export MY_BUCKET_SECRET_KEY=$secret_access_key

Create and configure your aws profile.

Use the aws tool to add access_key_id and secret_access_key to a aws-profile, and give a name to your aws-profile (Replace <my_aws_profile> below).

$ aws configure --profile <my_aws_profile>

The above command will ask for your access_key_id and secret_access_key. Press enter for all other fields requested. [none]

Create an alias to simplify the invocation of the aws command:

First, lets create an alias for repeating parts of the command (to avoid typing too much).

Mac OS users:

$ alias bxaws='aws --profile <my_aws_profile> --endpoint-url=http://s3-api.us-geo.objectstorage.softlayer.net'

Windows OS users:

doskey bxaws=aws --profile <my_aws_profile> --endpoint-url=http://s3-api.us-geo.objectstorage.softlayer.net $*

Create a bucket:

Now, lets make a bucket and name it something unique! Buckets are named globally, which means that only one IBM Cloud account can have a bucket with a particular name. **NB: the bucket names may not contain upper-case, underscores, dashes, periods, etc. Just use simple text, and add your userid as part of the bucket name.

$ bxaws s3api create-bucket --bucket <your-bucket-name>

Congratulations you are done with the setup!