Using Anypoint Code Builder to Configure Einstein AI Connector 1.2

To use Anypoint Code Builder to configure a connector or module, create a basic integration project in Anypoint Code Builder, add connectors and other components to your Mule application to process your business logic, and configure the attributes.

Before You Begin

Before creating an integration project, you must:

Set up and access the Anypoint Code Builder web or desktop IDE
Have credentials to access the connector’s API
Ensure that the Visual Studio Code Extension Pack for Java is installed.

To use this connector, you must be familiar with:

The connector’s API
Anypoint Connectors
Mule runtime engine (Mule)

Create a New Integration Project

To create a new integration project:

In the activity bar of the IDE, click the (Anypoint Code Builder) icon.
From Quick Actions, click Develop an Integration:

Complete the following in the Develop an Integration form:

Attribute Value

Attribute	Value
Project Name	Unique name for your project. This name is used as the title and name of the integration project file. For example, if the project name is "Salesforce Integration," the project file name is `salesforce-integration`.
Project Location	Your home directory or another directory you create. Click Browse to select a different directory for the integration project. Don’t create the project within another project directory. See Adding Folders to Your Home Directory.

Project Name

Unique name for your project.

This name is used as the title and name of the integration project file. For example, if the project name is "Salesforce Integration," the project file name is salesforce-integration.

Project Location

Your home directory or another directory you create. Click Browse to select a different directory for the integration project.

Don’t create the project within another project directory.

See Adding Folders to Your Home Directory.

Select the type of project to create by selecting either the Empty Project or Template or Example card.
Select the Mule runtime and Java Version for your app.
Click Create Project.

Add Components to Your Project

Add components to your integration project to build a flow:

In the Explorer view, open the XML file for your project, such as my-project-name.xml.
Select Build a Flow from the start card to create an empty flow:
Change the name of the flow:
1. Click Flow.
2. Click the edit icon.
3. Enter the flow name, and click the checkmark.
4. In the canvas UI, click the (Add component) icon.
5. In the Add Component panel, search for and select your component from the results. The following example shows the Listener component from the HTTP search results:
  
  The configuration XML file now includes the XML for the HTTP Listener into the <flow/> element, for example:
  <flow name="getFlights" > <http:listener path="" config-ref="" doc:name="Listener" doc:id="ojzfry" /> </flow>

Configure the Trigger For the Flow

A trigger (source) starts the flow when specified conditions are met. If a connector has its own associated sources, you can use a connector-specific trigger to initiate the flow. Each connector trigger has a unique configuration.

You can configure the following sources for this connector:

HTTP > Listener

Initiates a flow each time it receives a request on the configured host and port
Scheduler

Initiates a flow when a time-based condition is met

For example, to configure an HTTP Listener trigger, follow these steps:

Click the Listener component on the canvas UI.
(Optional) Click the edit icon to change the name of the Listener component.
(Optional) If you created a configuration file for the connector, select it.
In the General tab, in the Path field, enter the relative path for the path set in the HTTP configuration element. See Configure Paths.

Select the Advanced tab to configure the following attributes:

Attribute Description Required

Attribute	Description	Required
Allowed methods	To allow all HTTP methods (default), do not define this attribute. To restrict certain methods, specify all allowed HTTP methods, for example, `GET, POST`, as a comma-separated list.	No
Response streaming mode	Whether to use streaming when the response is sent.	No
Output MIME type	MIME type of the payload that this component outputs.	No
Parameters	Appends a key-value pair for a reader property to the value of the Output MIME attribute. Multiple key-value pairs are allowed.	No
Output encoding	Encoding of the payload that this component outputs.	No
Streaming Strategy	Configures the way data streams are consumed in Mule apps.	No
Primary node only	Select this option to execute the Listener trigger on the primary node only when running in a cluster.	No
Redelivery policy	Configures the redelivery policy for executing requests that generate errors. You can add a redelivery policy to any source in a flow.	No
Reconnection strategy	When an operation in a Mule application fails to connect to an external server, the default behavior is for the operation to fail immediately and return a connectivity error. You can modify the default behavior by configuring the reconnection strategy for the operation.	No
Response	See HTTP Response Attributes.	No
Error Response	HTTP error response returns a status code for the error.	No
Body	Body of the error message.	No
Headers	Headers to include in the error message.	No
Status code	Status code of the error response, for example, `500`.	No
Reason phrase	Reason for the error, for example, `Gateway Timeout` or `Internal Server Error`.	No

Allowed methods

To allow all HTTP methods (default), do not define this attribute. To restrict certain methods, specify all allowed HTTP methods, for example, GET, POST, as a comma-separated list.

Response streaming mode

Whether to use streaming when the response is sent.

Output MIME type

MIME type of the payload that this component outputs.

Parameters

Appends a key-value pair for a reader property to the value of the Output MIME attribute. Multiple key-value pairs are allowed.

Output encoding

Encoding of the payload that this component outputs.

Streaming Strategy

Configures the way data streams are consumed in Mule apps.

Primary node only

Select this option to execute the Listener trigger on the primary node only when running in a cluster.

Redelivery policy

Configures the redelivery policy for executing requests that generate errors. You can add a redelivery policy to any source in a flow.

Reconnection strategy

When an operation in a Mule application fails to connect to an external server, the default behavior is for the operation to fail immediately and return a connectivity error. You can modify the default behavior by configuring the reconnection strategy for the operation.

Response

See HTTP Response Attributes.

Error Response

HTTP error response returns a status code for the error.

Body

Body of the error message.

Headers

Headers to include in the error message.

Status code

Status code of the error response, for example, 500.

Reason phrase

Reason for the error, for example, Gateway Timeout or Internal Server Error.

Add the Connector to Your Integration Project

In the Explorer view, open the configuration XML file for your project, such as my-project-name.xml.
Click the (Show Mule graphical mode) icon in the activity bar to open the canvas UI if it doesn’t open automatically.
Add the connector the same way you added other components from the canvas UI:
1. In the canvas UI, click the (Add component) icon.
2. In the Add Component panel, click Connectors.
Click the (Search in Exchange) toggle:
Enter einstein ai in the search box.
Select the connector to add it to your project.

Adding a connector to an integration project doesn’t make that connector available to other projects in your Anypoint Code Builder workspace.

Add a Global Configuration

In the configuration XML, place your cursor before the <flow/> element and type einstein-ai.

Ensure that the cursor is not inside the <flow/> element.

Select the ms-einstein-ai:config snippet:

The snippet adds the following code:

<ms-einstein-ai:config llmType="#[]" configType="#[]" filePath="#[]" modelName="#[]" name="name" />

Add the attributes to complete the configuration, for example:

<ms-einstein-ai:config name="Einstein_AI_Config1" doc:name="Einstein AI Config" doc:id="96f110e3-f17c-4405-8911-80345985da12" >
	<ms-einstein-ai:oauth-client-credentials-connection >
			<ms-einstein-ai:oauth-client-credentials clientId="myclientid" clientSecret="myclientsecret" />
	</ms-einstein-ai:oauth-client-credentials-connection>
</ms-einstein-ai:config>

This example shows a flow named "Define_Prompt" with two global elements defined (http:listener-config and ms-einstein-ai:config):

<http:listener-config name="HTTP_Listener_config" doc:name="HTTP Listener config" doc:id="0868b5a9-90e8-4470-a255-8b9e8e3938a7" >
	<http:listener-connection host="0.0.0.0" port="8081" />
</http:listener-config>
<ms-einstein-ai:config name="Einstein_AI_Config" doc:name="Einstein AI Config" doc:id="a68f8f2c-a560-4d8e-9ebe-98a7d8d716cd" >
	<ms-einstein-ai:oauth-client-credentials-connection >
			<ms-einstein-ai:oauth-client-credentials clientId="myclientid" clientSecret="myclientsecret" />
	</ms-einstein-ai:oauth-client-credentials-connection>
</ms-einstein-ai:config>
<flow name="Define_Prompt">

Example Customer Feedback Flow

This example flow is designed to handle customer feedback analysis and responses using the Einstein AI service.

In the first route of the scatter-gather component, the flow defines a prompt template for an agent to analyze customer feedback provided in a dataset. If the feedback is negative, the agent is instructed to open a service case and apologize to the customer.

In the second route of the scatter-gather component, the flow defines a chat answer prompt for interacting with the Einstein AI service.

A flow with a Tranform component, a scatter-gather component, and two routes

Configure Additional Connector Fields

After you configure a global element for the connector, configure the other required fields. The required fields vary depending on which connector actions you use.

For more information about configuring Einstein AI Connector actions, see Einstein AI Connector - Additional Configuration Information.