Function Calling using Google Vertex AI

Back Next

Test Function Calling using Google Vertex AI

You can test the Function Calling using Google Vertex AI recipe with the function declaration value that is set as default in the Prepare Default Function Declaration
step of the Google Vertex Function Calling
process.

The default value applies to and can execute the Synchronize ServiceNow Incidents with Jira Issues recipe and the Update Amazon Bedrock Knowledge Base recipe.

For example, in the Function Calling using Google Vertex AI
guide, you can enter one of the following prompts for the correct function call for these recipes:

Please make an ingestion job for KB ON1FS86KFX with DS REQJVGUX8S and send the result to example@gmail.com

Synchronize my incidents in the Software category for November 19, 2024, project id 1001, Jira issue type id 1001, you also need to synchronize the Assignee and send the result to example@gmail.com

The function name must start with a letter or an underscore, and can contain alphanumeric characters, underscores (_), periods (.), and hyphens (-). For more information, see Google Vertex AI documentation.

The structure of the Function_Call_Answer
is important for further execution of the request to the Application Integration process. To set your Function_Declaration
in the guide, you must specify it in the JSON format as shown in the following example:

{
    "tools": [{
            "functionDeclarations": [{
                    "name": "Update_Amazon_Bedrock_Knowledge_Base_executor",
                    "description": "This function processes a user-provided prompt to identify and extract key information based on predefined parameters. For instance, it can detect specific identifiers such as Knowledge_Base_ID or Email_Address by scanning the text for patterns or matches associated with these keywords. The function returns the extracted data for further processing or validation, facilitating automated workflows where structured information is derived from unstructured input.",
                    "parameters": {
                        "type": "object",
                        "properties": {
                            "Process_URL": {
                                "type": "string",
                                "description": "Fixed URL to execution process",
                              "enum": ["<Service_URL_of_the_Application_Integration_Process>"]
                            },
                            "Payload": {
                                "type": "array",
                                "description": "Structured payload for processes",
                                "items": {
                                    "description": "Fields in payload for processes",
                                    "type": "object",
                                    "properties": {
                                        "Bucket_Name": {
                                            "type": "string",
                                            "description": "The bucket name in AWS S3 is a unique identifier for an S3 bucket."
                                        },
                                        "Knowledge_Base_ID": {
                                            "type": "string",
                                            "description": "The unique identifier of the knowledge base. Sometimes, the knowledge base is abbreviated as KB."
                                        },
                                        "Email_Address": {
                                            "type": "string",
                                            "description": "Email address where the process result will be sent."
                                        },
                                        "Data_Source_ID": {
                                            "type": "string",
                                            "description": "The unique identifier of the data source. Sometimes, the data source is abbreviated as DS."
                                        }
                                    },
                                    "required": ["Knowledge_Base_ID", "Email_Address"]
                                }
                            }
                        },
                        "required": ["Process_URL", "Payload"]
                    }
                }, {
                    "name": "Synchronize_ServiceNow_Incidents_with_Jira_Issues_executor",
                    "description": "This function processes a user-provided prompt to identify and extract key information based on predefined parameters. For instance, it can detect specific identifiers such as Jira_Issue_Type_ID or Jira_Project_ID by scanning the text for patterns or matches associated with these keywords. The function returns the extracted data for further processing or validation, facilitating automated workflows where structured information is derived from unstructured input.",
                    "parameters": {
                        "type": "object",
                        "properties": {
                            "Process_URL": {
                                "type": "string",
                              "description": "Fixed URL to execution process",
                              "enum": ["<Service_URL_of_the_Application_Integration_Process>"]
                            },
                            "Payload": {
                                "type": "array",
                                "description": "Structured payload for processes",
                                "items": {
                                    "description": "Fields in payload for processes",
                                    "type": "object",
                                    "properties": {
                                        "Date_For_Search": {
                                            "type": "string",
                                            "description": "Search date in the YYYY-MM-DD format."
                                        },
                                        "Filter": {
                                            "type": "string",
                                            "description": "Name of the category to be filtered"
                                        },
                                        "Email": {
                                            "type": "string",
                                            "description": "Email address where the process result will be sent."
                                        },
                                        "Jira_Issue_Type_ID": {
                                            "type": "string",
                                            "description": "The unique identifier of Jira issue type"
                                        },
                                        "Jira_Project_ID": {
                                            "type": "string",
                                            "description": "The unique identifier of Jira project"
                                        },
                                        "Sync_Assignee": {
                                            "type": "boolean",
                                            "description": "Boolean variable whether to synchronize the Assignee"
                                        }
                                    },
                                    "required": ["Email", "Jira_Issue_Type_ID", "Jira_Project_ID"]
                                }
                            }
                        },
                        "required": ["Process_URL", "Payload"]
                    }
                }

            ]

        }

    ]
}

For testing, you can construct another structure of a Function Declaration as required.

Using the Function Calling using Google Vertex AI recipe

Download Guide

Watch

Comments

Cloud Application Integration Homepage