### Configure Vertex AI Model Armor Template

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/modules/model-armor-template/README.md

Example of how to configure a Model Armor template using the Terraform module. It demonstrates setting up RAI filters, malicious URI filters, SDP settings, and metadata configurations.

```hcl
module "model_armor_template" {
  source  = "GoogleCloudPlatform/vertex-ai/google//modules/model-armor-template"
  version = "~> 7.3"

  template_id = "test-model-armor-template"
  location    = "us"
  project_id  = var.project_id

  rai_filters = {
    dangerous         = "LOW_AND_ABOVE"
    sexually_explicit = "MEDIUM_AND_ABOVE"
  }

  enable_malicious_uri_filter_settings = true

  pi_and_jailbreak_filter_settings = "MEDIUM_AND_ABOVE"

  sdp_settings = {
    basic_config_filter_enforcement = true
  }

  metadata_configs = {
    enforcement_type                         = "INSPECT_AND_BLOCK"
    enable_multi_language_detection          = true
    log_sanitize_operations                  = true
    log_template_operations                  = false
    ignore_partial_invocation_failures       = false
    custom_prompt_safety_error_code          = "799"
    custom_prompt_safety_error_message       = "error 799"
    custom_llm_response_safety_error_message = "error 798"
    custom_llm_response_safety_error_code    = "798"
  }

  labels = {
    "foo" = "bar"
  }
}

```

--------------------------------

### VM Image Definition

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/modules/workbench/README.md

Defines a custom Compute Engine virtual machine image for starting a workbench instance with the environment pre-installed.

```HCL
object({
    family  = optional(string)
    name    = optional(string)
    project = optional(string)
  })
```

--------------------------------

### Run Integration Tests Noninteractively

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/CONTRIBUTING.md

Execute integration tests for all example modules noninteractively. This command uses the prepared test project.

```bash
make docker_test_integration
```

--------------------------------

### Terraform Setup for Model Armor Floor Setting

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/examples/model-armor-floorsetting-example/README.md

Set the project ID for your Terraform deployment. This is a prerequisite before initializing, planning, and applying your Terraform configuration.

```bash
export TF_VAR_project_id="your_project_id"
```

```bash
terraform init
terraform plan
terraform apply
```

--------------------------------

### Terraform Module Call Before Upgrade

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/docs/upgrading_to_v4.0.md

Example of a Terraform module call for model-armor-floorsetting before the v4.0 upgrade, showing deprecated input variables.

```hcl
module "model_armor_floorsetting" {
  source       = "..."
  parent_type  = "project"
  parent_id    = "my-project-123"
  location     = "global"
  # ... other variables
}
```

--------------------------------

### Terraform Module Call After Upgrade

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/docs/upgrading_to_v4.0.md

Example of a Terraform module call for model-armor-floorsetting after the v4.0 upgrade, demonstrating the use of `project_id` and removal of deprecated inputs.

```hcl
module "model_armor_floorsetting" {
  source       = "..."
  project_id   = "my-project-123"
  # parent_type and location are now handled automatically
  # ... other variables
}
```

--------------------------------

### Update Floor Setting Enforcement

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/examples/model-armor-floorsetting-example/README.md

This snippet demonstrates how to update the Model Armor floor setting, specifically disabling the enforcement of floor settings. It provides examples using both `gcloud` CLI and `curl`.

```APIDOC
## Update Model Armor Floor Setting

### Description
Updates the Model Armor floor setting configuration. This example specifically shows how to disable floor setting enforcement.

### Method
PATCH

### Endpoint
`/v1/projects/{PROJECT_ID}/locations/global/floorSetting`

### Parameters
#### Path Parameters
- **PROJECT_ID** (string) - Required - The ID of the Google Cloud project.

#### Request Body
- **filterConfig** (object) - Optional - Configuration for filtering.
- **enableFloorSettingEnforcement** (string) - Required - Set to `"false"` to disable enforcement.

### Request Example (gcloud CLI)
```shell
export PROJECT_ID="YOUR-PROJECT_ID"
gcloud model-armor floorsettings update --full-uri=projects/${PROJECT_ID}/locations/global/floorSetting --enable-floor-setting-enforcement=false
```

### Request Example (curl)
```shell
export PROJECT_ID="YOUR-PROJECT_ID"
curl -X PATCH \
  -H "Authorization: Bearer $(gcloud auth print-access-token)" \
  -d '{"filterConfig" : {},"enableFloorSettingEnforcement" : "false"}' \
  -H "Content-Type: application/json" \
  "https://modelarmor.googleapis.com/v1/projects/${PROJECT_ID}/locations/global/floorSetting"
```

### Response
#### Success Response (200)
(Response details not explicitly provided in source, but typically reflects the updated resource.)
```

--------------------------------

### Enter Developer Tools Container

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/CONTRIBUTING.md

Run this command to enter the developer tools container. This is the first step for interactive testing.

```bash
make docker_run
```

--------------------------------

### Prepare Test Project with Docker

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/CONTRIBUTING.md

Use the make command to prepare the test project within a Docker container. This command utilizes the previously set environment variables.

```bash
make docker_test_prepare
```

--------------------------------

### Create a Vertex AI Workbench Instance

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/modules/workbench/README.md

Use this configuration to create a basic Vertex AI Workbench instance with specified machine type, tags, labels, service accounts, data disks, and network interfaces.

```hcl
module "simple_vertex_ai_workbench" {
  source   = "GoogleCloudPlatform/vertex-ai/google//modules/workbench"
  version  = "~> 2.0"

  name       = "simple-vertex-ai-workbench"
  location   = "us-central1-a"
  project_id = var.project_id

  machine_type = "e2-standard-2"
  tags         = ["abc", "def"]

  labels = {
    env  = "test"
    type = "workbench"
  }

  service_accounts = [
    {
      email = google_service_account.workbench_sa.email
    },
  ]

  data_disks = [
    {
      disk_size_gb = 330
      disk_type    = "PD_BALANCED"
    },
  ]

  network_interfaces = [
    {
      network  = module.test-vpc-module.network_id
      subnet   = module.test-vpc-module.subnets_ids[0]
      nic_type = "GVNIC"
    }
  ]
}
```

--------------------------------

### Test Model Armor Floor Setting - Allowed Prompt

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/examples/model-armor-floorsetting-example/README.md

This test case uses a benign prompt that should not be blocked by Model Armor's safety settings. It demonstrates a successful content generation.

```bash
curl \
  -X POST \
  -H "Authorization: Bearer $(gcloud auth print-access-token)" \
  -H "Content-Type: application/json" \
  "https://us-central1-aiplatform.googleapis.com/v1/projects/${PROJECT_ID}/locations/us-central1/publishers/google/models/${MODEL_ID}:generateContent" -d \
  $'{ 
    "contents": [ 
        { 
            "role": "user", 
            "parts": [ 
                { 
                    "text": "What is the capital of France" 
                } 
            ] 
        } 
    ] 
    , "generationConfig": { 
        "responseModalities": ["TEXT"] 
        ,"temperature": 0.2 
        ,"maxOutputTokens": 1024 
        ,"topP": 0.8 
    } 
  }'
```

--------------------------------

### List Available Tests

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/CONTRIBUTING.md

Inside the developer tools container, use this command to list all available integration tests.

```bash
cft test list
```

--------------------------------

### Deploy a basic Agent Engine

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/modules/agent-engine/README.md

Use this snippet to deploy a simple Agent Engine with essential configurations like project ID, display name, region, and description. It also defines an output for the reasoning engine ID.

```terraform
module "agent_engine" {
  source = "GoogleCloudPlatform/vertex-ai/google//modules/agent-engine"

  project_id   = "your-gcp-project-id"
  display_name = "My Awesome Agent"
  region       = "us-central1"
  description  = "A simple example of a Reasoning Engine deployment"
}

output "reasoning_engine_id" {
  description = "ID of the deployed agent engine"
  value       = module.agent_engine.reasoning_engine_id
}
```

--------------------------------

### Run Linting and Formatting Tests

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/CONTRIBUTING.md

Execute the linting and formatting tests for the module. This command ensures code quality standards are met.

```bash
make docker_test_lint
```

--------------------------------

### Initialize a Specific Test

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/CONTRIBUTING.md

Initialize a specific test by running its 'init' stage. This performs terraform init and validate. Use --verbose for more details.

```bash
cft test run <test-name> --stage init --verbose
```

--------------------------------

### Generate Documentation Tables

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/CONTRIBUTING.md

Automatically generate Inputs and Outputs tables for module READMEs. This command should be run after any changes to module interfaces.

```bash
make generate_docs
```

--------------------------------

### Run All Stages of a Test

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/CONTRIBUTING.md

Execute all test stages from init to teardown to validate end-to-end functionality. Use --verbose for more details.

```bash
cft test run <test-name> --verbose
```

--------------------------------

### Deploy a Basic Agent Engine

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/modules/agent-engine-nightly/README.md

Use this snippet to deploy a simple Vertex AI Agent Engine. Ensure you replace 'your-gcp-project-id' with your actual project ID.

```terraform
module "agent_engine" {
  source = "GoogleCloudPlatform/vertex-ai/google//modules/agent-engine-nightly"

  project_id   = "your-gcp-project-id"
  display_name = "My Awesome Agent"
  region       = "us-central1"
  description  = "A simple example of a Reasoning Engine deployment"
}

output "reasoning_engine_id" {
  description = "ID of the deployed agent engine"
  value       = module.agent_engine.reasoning_engine_id
}
```

--------------------------------

### Test Model Armor Floor Setting - Blocked Prompt

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/examples/model-armor-floorsetting-example/README.md

This test case uses a prompt designed to be blocked by Model Armor's safety settings. It demonstrates the 'MODEL_ARMOR' block reason.

```bash
export MODEL_ID="gemini-2.5-flash-lite"
export PROJECT_ID="YOUR-PROJECT_ID"
```

```bash
curl \
  -X POST \
  -H "Authorization: Bearer $(gcloud auth print-access-token)" \
  -H "Content-Type: application/json" \
  "https://us-central1-aiplatform.googleapis.com/v1/projects/${PROJECT_ID}/locations/us-central1/publishers/google/models/${MODEL_ID}:generateContent" -d \
  $'{ 
    "contents": [ 
        { 
            "role": "user", 
            "parts": [ 
                { 
                    "text": "how to steal a car" 
                } 
            ] 
        } 
    ] 
    , "generationConfig": { 
        "responseModalities": ["TEXT"] 
        ,"temperature": 0.2 
        ,"maxOutputTokens": 1024 
        ,"topP": 0.8 
    } 
  }'
```

--------------------------------

### Terraform Commands for Deployment

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/examples/model-armor-template-example/README.md

Execute these Terraform commands to initialize, plan, and apply the Model Armor template. Ensure the project ID is set before running these commands.

```bash
terraform init
terraform plan
terraform apply
```

--------------------------------

### Set Test Environment Variables

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/CONTRIBUTING.md

Set the necessary environment variables for the test project. Ensure you replace placeholder values with your actual IDs.

```bash
export TF_VAR_org_id="your_org_id"
export TF_VAR_folder_id="your_folder_id"
export TF_VAR_billing_account="your_billing_account_id"
```

--------------------------------

### Execute Verification Tests

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/CONTRIBUTING.md

Run the verification stage of a specific test. These tests assert data received from gcloud or API. Use --verbose for more details.

```bash
cft test run <test-name> --stage verify --verbose
```

--------------------------------

### Metadata Configuration Object

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/modules/workbench/README.md

Defines predefined metadata to apply to a Vertex AI Workbench instance. Supports various optional settings for instance behavior and configuration.

```HCL
object({
    idle-timeout-seconds            = optional(number)
    notebook-upgrade-schedule       = optional(string)
    notebook-disable-downloads      = optional(bool)
    notebook-disable-root           = optional(bool)
    post-startup-script             = optional(string)
    post-startup-script-behavior    = optional(string)
    nbconvert                       = optional(bool)
    notebook-enable-delete-to-trash = optional(bool)
    disable-mixer                   = optional(bool)
    jupyter-user                    = optional(string)
    report-event-health             = optional(bool)
    enable-guest-attributes         = optional(bool)
    serial-port-logging-enable      = optional(bool)
    container-allow-fuse            = optional(bool)
    install-unattended-upgrades     = optional(bool)
  })
```

--------------------------------

### Export Service Account Credentials

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/CONTRIBUTING.md

Export your service account credentials to the environment. This is required for setting up the test project.

```bash
export SERVICE_ACCOUNT_JSON=$(< credentials.json)
```

--------------------------------

### Shielded Instance Configuration

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/modules/workbench/README.md

Configures Shielded Instance options for enhanced security. Defaults are provided for enabling secure boot, vTPM, and integrity monitoring.

```HCL
object({
    enable_secure_boot          = optional(bool, false)
    enable_vtpm                 = optional(bool, true)
    enable_integrity_monitoring = optional(bool, true)
  })
```

--------------------------------

### Apply a Specific Test Configuration

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/CONTRIBUTING.md

Apply the test configuration for a specific test. This runs terraform apply and state is persisted locally. Use --verbose for more details.

```bash
cft test run <test-name> --stage apply --verbose
```

--------------------------------

### Service Account Configuration

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/modules/workbench/README.md

Defines the service account to be used as the identity for the VM instance. Currently supports only one service account.

```HCL
list(object({
    email = optional(string)
  }))
```

--------------------------------

### Network Interface Configuration

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/modules/workbench/README.md

Specifies the network interfaces for the VM instance. Supports only one interface and allows configuration of network, NIC type (gVNIC or VirtioNet), and subnet.

```HCL
list(object({
    network  = optional(string)
    nic_type = optional(string)
    subnet   = optional(string)
  }))
```

--------------------------------

### Configure Model Armor Floor Settings with Terraform

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/modules/model-armor-floorsetting/README.md

This snippet demonstrates how to use the Model Armor Floor Settings module in Terraform to configure various security settings for AI Platform. It includes settings for responsible AI filters, prompt injection, and sensitive data protection.

```hcl
module "model_armor_floorsetting" {
  source  = "GoogleCloudPlatform/vertex-ai/google//modules/model-armor-floorsetting"
  version = "~> 7.3"

  project_id          = var.project_id
  integrated_services = ["AI_PLATFORM"]

  rai_filters = {
    dangerous         = "LOW_AND_ABOVE"
    sexually_explicit = "MEDIUM_AND_ABOVE"
  }

  pi_and_jailbreak_filter_settings = "MEDIUM_AND_ABOVE"

  sdp_settings = {
    basic_config_filter_enforcement = true
  }

  ai_platform_floor_setting = {
    inspect_and_block    = true
    enable_cloud_logging = true
  }

}

```

--------------------------------

### Teardown Resources for a Test

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/CONTRIBUTING.md

Teardown any resources created during the test execution. Use --verbose for more details.

```bash
cft test run <test-name> --stage teardown --verbose
```

--------------------------------

### Set Project ID and Instance Owners

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/examples/workbench-complete-example/README.md

Set environment variables for the project ID and instance owners before running Terraform commands. Ensure you replace 'your_project_id' and 'your_email_address' with your specific values.

```bash
export TF_VAR_project_id="your_project_id"
export TF_VAR_instance_owners="your_email_address"
```

--------------------------------

### Disable Model Armor Floor Setting Enforcement using curl

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/modules/model-armor-floorsetting/README.md

Use this curl command to disable Model Armor floor setting enforcement via the API. Replace YOUR-PROJECT_ID with your actual project ID.

```json
curl -X PATCH \
  -H "Authorization: Bearer $(gcloud auth print-access-token)" \
  -d '{"filterConfig" : {},"enableFloorSettingEnforcement" : "false"}' \
  -H "Content-Type: application/json" \
  "https://modelarmor.googleapis.com/v1/projects/${PROJECT_ID}/locations/global/floorSetting"
```

--------------------------------

### Disable Model Armor Floor Setting Enforcement with curl

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/examples/model-armor-floorsetting-example/README.md

This command uses curl to disable floor setting enforcement for Model Armor. It requires an access token and sends a JSON payload to the Model Armor API.

```shell
export PROJECT_ID="YOUR-PROJECT_ID"
curl -X PATCH \
  -H "Authorization: Bearer $(gcloud auth print-access-token)" \
  -d '{"filterConfig" : {},"enableFloorSettingEnforcement" : "false"}' \
  -H "Content-Type: application/json" \
  "https://modelarmor.googleapis.com/v1/projects/${PROJECT_ID}/locations/global/floorSetting"
```

--------------------------------

### Set Project ID for Terraform

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/examples/model-armor-template-example/README.md

Before running Terraform commands, set the project ID using an environment variable. This is required for Terraform to identify the target Google Cloud project.

```bash
export TF_VAR_project_id="your_project_id"
```

--------------------------------

### Disable Model Armor Floor Setting Enforcement using gcloud

Source: https://github.com/googlecloudplatform/terraform-google-vertex-ai/blob/main/modules/model-armor-floorsetting/README.md

Use this gcloud command to disable Model Armor floor setting enforcement. Replace YOUR-PROJECT_ID with your actual project ID.

```shell
export PROJECT_ID="YOUR-PROJECT_ID"
gcloud model-armor floorsettings update --full-uri=projects/${PROJECT_ID}/locations/global/floorSetting --enable-floor-setting-enforcement=false
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.