Trustwise SDK

https://trustwiseai.github.io/trustwise
Admin
Trustwise SDK is a Python tool for evaluating AI-generated content with metrics like faithfulness...
Tokens:156,037
Snippets:1,390
Trust Score:9.7
Update:5 months ago
Show doc for...
Tokens:
Raw
### Complete Migration Example: After TrustwiseClient

Source: https://trustwiseai.github.io/trustwise/_sources/sdk_migration.rst.txt

Shows the equivalent setup and usage with TrustwiseClient, including accessing other platform features.

```python
import os
from trustwise.sdk.config import TrustwiseConfig
from trustwise.sdk.platform import TrustwiseClient

config = TrustwiseConfig(api_key=os.environ["TW_API_KEY"])
client = TrustwiseClient(config)

result = client.metrics.faithfulness(
    query="What is the capital of France?",
    response="The capital of France is Paris.",
    context=[{"chunk_text": "Paris is the capital of France.", "chunk_id": "1"}],
)
print(result.score)

# Plus: access agents, policies, guardrails, evaluations, etc.
agents = client.agents.list(limit=5)
```

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

### Complete Migration Example: Before TrustwiseSDK

Source: https://trustwiseai.github.io/trustwise/_sources/sdk_migration.rst.txt

Illustrates the complete setup and usage of TrustwiseSDK for evaluating metric faithfulness.

```python
import os
from trustwise.sdk import TrustwiseSDK
from trustwise.sdk.config import TrustwiseConfig

config = TrustwiseConfig(api_key=os.environ["TW_API_KEY"])
sdk = TrustwiseSDK(config)

result = sdk.metrics.faithfulness.evaluate(
    query="What is the capital of France?",
    response="The capital of France is Paris.",
    context=[{"chunk_text": "Paris is the capital of France.", "chunk_id": "1"}],
)
print(result.score)
```

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

### Install Trustwise SDK using pip

Source: https://trustwiseai.github.io/trustwise/_sources/installation.rst.txt

Use this command to install the Trustwise SDK and its dependencies.

```bash
pip install trustwise
```

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

### TrustwiseClient Quickstart

Source: https://trustwiseai.github.io/trustwise/_sources/platform.rst.txt

Demonstrates basic usage of TrustwiseClient for listing agents and evaluating metrics.

```APIDOC
## Quickstart

.. code-block:: python

    import os
    from trustwise.sdk.config import TrustwiseConfig
    from trustwise.sdk.platform import TrustwiseClient

    config = TrustwiseConfig(api_key=os.environ["TW_API_KEY"])
    client = TrustwiseClient(config)

    # List agents within a project
    agents = client.agents.list("my-project-id", limit=5)

    # Evaluate faithfulness
    result = client.metrics.faithfulness(
        query="What is the capital of France?",
        response="The capital of France is Paris.",
        context=[{"chunk_text": "Paris is the capital.", "chunk_id": "1"}],
    )
```

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

### Quickstart: Initialize Client and List Agents

Source: https://trustwiseai.github.io/trustwise/platform.html

Initialize the TrustwiseClient with configuration and list agents within a project. Requires TW_API_KEY environment variable.

```python
import os
from trustwise.sdk.config import TrustwiseConfig
from trustwise.sdk.platform import TrustwiseClient

config = TrustwiseConfig(api_key=os.environ["TW_API_KEY"])
client = TrustwiseClient(config)

# List agents within a project
agents = client.agents.list("my-project-id", limit=5)

# Evaluate faithfulness
result = client.metrics.faithfulness(
    query="What is the capital of France?",
    response="The capital of France is Paris.",
    context=[{"chunk_text": "Paris is the capital.", "chunk_id": "1"}],
)

```

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

### Initialize Legacy TrustwiseSDK

Source: https://trustwiseai.github.io/trustwise/_sources/usage.rst.txt

The TrustwiseSDK is deprecated. Use TrustwiseClient instead. This example shows the legacy initialization process.

```python
import os
from trustwise.sdk import TrustwiseSDK
from trustwise.sdk.config import TrustwiseConfig

# Configure using environment variables
os.environ["TW_API_KEY"] = "your-api-key"
os.environ["TW_BASE_URL"] = "https://api.trustwise.ai"
config = TrustwiseConfig()

# Or configure directly
# config = TrustwiseConfig(
#     api_key="your-api-key",
#     base_url="https://api.trustwise.ai"
# )

# Initialize the SDK
trustwise = TrustwiseSDK(config)
```

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

### Initialize TrustwiseClient and List Agents

Source: https://trustwiseai.github.io/trustwise/_sources/platform.rst.txt

Initialize the TrustwiseConfig with an API key and create a TrustwiseClient instance. This example demonstrates listing agents within a specified project.

```python
import os
from trustwise.sdk.config import TrustwiseConfig
from trustwise.sdk.platform import TrustwiseClient

config = TrustwiseConfig(api_key=os.environ["TW_API_KEY"])
client = TrustwiseClient(config)

# List agents within a project
agents = client.agents.list("my-project-id", limit=5)
```

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

### Complete Migration Example: Before TrustwiseSDK

Source: https://trustwiseai.github.io/trustwise/sdk_migration.html

This code block demonstrates the usage of the deprecated TrustwiseSDK for evaluating faithfulness metrics.

```python
import os
from trustwise.sdk import TrustwiseSDK
from trustwise.sdk.config import TrustwiseConfig

config = TrustwiseConfig(api_key=os.environ["TW_API_KEY"])
 sdk = TrustwiseSDK(config)

result = sdk.metrics.faithfulness.evaluate(
    query="What is the capital of France?",
    response="The capital of France is Paris.",
    context=[{"chunk_text": "Paris is the capital of France.", "chunk_id": "1"}],
)
print(result.score)

```

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

### Policies

Source: https://trustwiseai.github.io/trustwise/_sources/platform.rst.txt

Evaluate policies in batch or get recommendations based on tags.

```APIDOC
## Evaluate Policies in Batch

### Description
Evaluates an input against multiple policies simultaneously.

### Method
`client.policies.evaluate_batch

### Parameters
- **policy_ids** (list[str]) - Required - A list of policy IDs to evaluate against.
- **input** (dict) - Required - The input data for evaluation, typically containing a 'text' field.

### Request Example
```python
results = client.policies.evaluate_batch(
    policy_ids=["policy-1", "policy-2"],
    input={"text": "test input"},
)
```

## Get Policy Recommendations

### Description
Retrieves policy recommendations based on specified tags.

### Method
`client.policies.recommend

### Parameters
- **tags** (list[str]) - Required - A list of tags to filter recommendations.

### Request Example
```python
recs = client.policies.recommend(tags=["safety", "pii"])
```
```

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

### List Supported Auth Schemas

Source: https://trustwiseai.github.io/trustwise/_sources/platform.rst.txt

Get a list of supported authentication schemas for runtime providers. This helps in configuring secure agent deployments.

```python
# List supported auth schemas
schemas = client.runtime_providers.auth_schemas()
```

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

### Manage Policies: List, Evaluate, and Recommend

Source: https://trustwiseai.github.io/trustwise/platform.html

List policies, evaluate content against single or multiple policies, and get policy recommendations based on tags. Requires `policy_id` for evaluation.

```python
# List policies
policies = client.policies.list()

# Evaluate content against a policy
result = client.policies.evaluate(policy_id, input="test input")

# Evaluate against multiple policies in one call
results = client.policies.evaluate_batch(
    policy_ids=["policy-1", "policy-2"],
    input={"text": "test input"},
)

# Get policy recommendations based on tags
recs = client.policies.recommend(tags=["safety", "pii"])

```

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

### Get Available API Versions

Source: https://trustwiseai.github.io/trustwise/versioning.html

Retrieve a list of all available API versions supported by the SDK. Ensure TrustwiseSDK and TrustwiseConfig are imported and initialized.

```python
from trustwise.sdk import TrustwiseSDK
from trustwise.sdk.config import TrustwiseConfig

config = TrustwiseConfig(api_key="your-api-key")
trustwise = TrustwiseSDK(config)

# Get available versions
versions = trustwise.get_versions()

```

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

### Accessing Agents with TrustwiseClient

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/platform/client.html

Demonstrates how to interact with the agents sub-client to retrieve and list agents within a specific project. The `get` method fetches a single agent, while `list` retrieves multiple agents, supporting search and limiting results. By default, methods return the inner data model.

```python
# Agents are now project-scoped
agent = client.agents.get("project-id", "agent-id")
print(agent.name)

agents = client.agents.list("project-id", search="my-bot", limit=5)
for a in agents:
    print(a.name)
```

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

### Get Available API Versions

Source: https://trustwiseai.github.io/trustwise/_sources/versioning.rst.txt

Retrieve a list of all available API versions supported by the Trustwise SDK. Ensure TrustwiseSDK and TrustwiseConfig are imported and the SDK is initialized with your API key.

```python
from trustwise.sdk import TrustwiseSDK
from trustwise.sdk.config import TrustwiseConfig

config = TrustwiseConfig(api_key="your-api-key")
trustwise = TrustwiseSDK(config)

# Get available versions
versions = trustwise.get_versions()
```

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

### Deploy Agent to AI Gateway and Get Status History

Source: https://trustwiseai.github.io/trustwise/_sources/platform.rst.txt

Deploy an agent to the AI Gateway with specified models and retrieve its status history. Both actions are project-scoped.

```python
# Deploy to AI Gateway
deploy = client.agents.deploy_gateway(project_id, agent["id"], models=["gpt-4.1-nano"])

# Status history
history = client.agents.status_history(project_id, agent["id"])
```

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

### Carbon Footprint Evaluation Example Response

Source: https://trustwiseai.github.io/trustwise/v4_metrics.html

An example of the JSON response structure detailing the carbon footprint, including total and component breakdowns.

```json
{
    "carbon": {
        "value": 0.0011949989480068127,
        "unit": "kg_co2e"
    },
    "components": [
        {
            "component": "operational_gpu",
            "carbon": {
                "value": 0.0,
                "unit": "kg_co2e"
            }
        },
        {
            "component": "operational_cpu",
            "carbon": {
                "value": 0.00021294669026962343,
                "unit": "kg_co2e"
            }
        },
        {
            "component": "embodied_cpu",
            "carbon": {
                "value": 0.0009820522577371892,
                "unit": "kg_co2e"
            }
        }
    ]
}
```

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

### PlatformHTTP GET Method

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/platform/_http.html

Performs an HTTP GET request to the specified path with optional query parameters. It utilizes the internal _request method for execution.

```python
def get(self, path: str, **params: Any) -> dict[str, Any]:
    return self._request("GET", path, params=params or None)
```

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

### TrustwiseClient Initialization and Usage

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/platform/client.html

Demonstrates how to initialize the TrustwiseClient with configuration and access various platform resources like agents, gateway, guardrails, and more. It also shows how to retrieve agents by project and agent ID, list agents with search parameters, and access raw API responses.

```APIDOC
## TrustwiseClient

### Description
User-friendly client for the Trustwise Harmony Platform API. Provides access to all platform resources through sub-clients: ``agents``, ``gateway``, ``guardrails``, ``policies``, ``evaluations``, ``evaluators``, ``organizations``, ``projects``, ``risk_engine``, ``generation``, ``metrics``, ``events``, ``subscriptions``, and ``runtime_providers``.

### Example

```python
from trustwise.sdk.config import TrustwiseConfig
from trustwise.sdk.platform import TrustwiseClient

config = TrustwiseConfig(api_key="tw-...")
client = TrustwiseClient(config)

# Agents are now project-scoped
agent = client.agents.get("project-id", "agent-id")
print(agent.name)

agents = client.agents.list("project-id", search="my-bot", limit=5)
for a in agents:
    print(a.name)

# Raw — returns the full API envelope
resp = client.agents.get("project-id", "agent-id", raw=True)
print(resp["success"], resp["data"]["name"])
```

### Parameters

- **config** (TrustwiseConfig) - Trustwise configuration instance.
- **timeout** (int, optional) - HTTP request timeout in seconds. Defaults to 30.
```

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

### Get Agent Status History

Source: https://trustwiseai.github.io/trustwise/api.html

Retrieve the status history for an agent. Supports filtering and pagination through `kwargs`. Set `raw=True` to get the full API envelope.

```python
status_history = client.agents.status_history(project_id="project-id", agent_id="agent-id")
```

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

### Initialize TrustwiseSDK and Evaluate Faithfulness (Sync)

Source: https://trustwiseai.github.io/trustwise/_sources/v4_metrics.rst.txt

Demonstrates initializing the TrustwiseSDK with configuration and evaluating the faithfulness metric using synchronous calls. Ensure you have your API key and necessary types imported.

```python
from trustwise.sdk import TrustwiseSDK
from trustwise.sdk.config import TrustwiseConfig
from trustwise.sdk.metrics.v4.types import ContextChunk

config = TrustwiseConfig(api_key="your-api-key")
trustwise = TrustwiseSDK(config)

# v4 context format
context = [
    ContextChunk(chunk_text="Paris is the capital of France.", chunk_id="doc:idx:1")
]

# v4 metric calls (now default)
result = trustwise.metrics.faithfulness.evaluate(
    query="What is the capital of France?",
    response="The capital of France is Paris.",
    context=context
)
```

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

### get() method

Source: https://trustwiseai.github.io/trustwise/genindex.html

The generic get() method is available across multiple resource classes for retrieving specific items. It is used in AgentResource, EvaluationResource, EventResource, GuardrailResource, OrganizationResource, PolicyResource, ProjectResource, RuntimeProviderResource, and SubscriptionResource.

```APIDOC
## GET

### Description
Retrieves a specific resource.

### Method
GET

### Endpoint
Not specified, depends on the resource.

### Parameters
None explicitly defined in the source.

### Request Example
None provided.

### Response
None explicitly defined.
```

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

### Initialize Trustwise SDK Async

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/async_sdk.html

Initialize the Trustwise SDK asynchronously with configuration. This sets up the HTTP client and metrics access.

```python
from trustwise.sdk.async_sdk import TrustwiseSDKAsync
from trustwise.config import TrustwiseConfig

config = TrustwiseConfig(api_key="YOUR_API_KEY")
sdk = TrustwiseSDKAsync(config)

```

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

### Get Organization Logo

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/platform/client.html

Retrieves the logo for a specific organization.

```APIDOC
## Get Organization Logo

### Description
Get the logo for an organization.

### Method
GET

### Endpoint
/v1/organizations/{org_id}/logo

### Parameters
#### Path Parameters
- **org_id** (str) - Required - Unique organization identifier.
#### Query Parameters
- **raw** (bool) - Optional - Return the full API envelope.

### Response
#### Success Response (200)
- **data** (LogoResponse) - The logo details if raw=False.
- **envelope** (dict) - The full API envelope if raw=True.
```

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

### Get Guardrail

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/platform/client.html

Retrieves a single guardrail by its unique identifier.

```APIDOC
## Get Guardrail

### Description
Get a single guardrail by ID.

### Method
GET

### Endpoint
/v1/guardrails/{guardrail_id}

### Parameters
#### Path Parameters
- **guardrail_id** (str) - Required - Unique guardrail identifier.
#### Query Parameters
- **raw** (bool) - Optional - Return the full API envelope.

### Response
#### Success Response (200)
- **guardrail** (dict) - The guardrail object.
- **raw** (dict) - Full API envelope if raw=True.
```

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

### TrustwiseClient Initialization and Basic Usage

Source: https://trustwiseai.github.io/trustwise/platform.html

Initialize the TrustwiseClient with configuration and demonstrate listing agents and evaluating faithfulness.

```APIDOC
## Initialization and Basic Usage

### Description
Initialize the `TrustwiseClient` and demonstrate common operations like listing agents and evaluating metrics.

### Code Example
```python
import os
from trustwise.sdk.config import TrustwiseConfig
from trustwise.sdk.platform import TrustwiseClient

config = TrustwiseConfig(api_key=os.environ["TW_API_KEY"])
client = TrustwiseClient(config)

# List agents within a project
agents = client.agents.list("my-project-id", limit=5)

# Evaluate faithfulness
result = client.metrics.faithfulness(
    query="What is the capital of France?",
    response="The capital of France is Paris.",
    context=[{"chunk_text": "Paris is the capital.", "chunk_id": "1"}],
)
```
```

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

### Initialize and Use TrustwiseClient

Source: https://trustwiseai.github.io/trustwise/api.html

Demonstrates how to initialize the TrustwiseClient with configuration and access agents within a project. Use this client for all new development to interact with the Trustwise Platform API.

```python
from trustwise.sdk.config import TrustwiseConfig
from trustwise.sdk.platform import TrustwiseClient

config = TrustwiseConfig(api_key="tw-...")
client = TrustwiseClient(config)

# Agents are now project-scoped
agent = client.agents.get("project-id", "agent-id")
print(agent.name)

agents = client.agents.list("project-id", search="my-bot", limit=5)
for a in agents:
    print(a.name)

```

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

### Get Provider Schemas

Source: https://trustwiseai.github.io/trustwise/api.html

Fetches the configuration schemas for all guardrail providers.

```APIDOC
## provider_schemas

### Description
Get the configuration schemas for all guardrail providers.

### Return Type
`dict`[`str`, `Any`]

### Returns
Configuration schemas for all guardrail providers.
```

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

### Get Policy by ID

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/platform/client.html

Retrieves a single policy by its unique identifier.

```APIDOC
## Get Policy by ID

### Description
Get a single policy by ID.

### Method
GET

### Endpoint
/v1/policies/{policy_id}

### Parameters
#### Path Parameters
- **policy_id** (str) - Required - Unique policy identifier.
#### Query Parameters
- **raw** (bool) - Optional - Return the full API envelope.
```

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

### Initialize TrustwiseClient

Source: https://trustwiseai.github.io/trustwise/_sources/usage.rst.txt

Use TrustwiseClient for new development to access all platform APIs. Configure using environment variables or directly with API key and base URL.

```python
import os
from trustwise.sdk.config import TrustwiseConfig
from trustwise.sdk.platform import TrustwiseClient

# Configure using environment variables
config = TrustwiseConfig(api_key=os.environ["TW_API_KEY"])

# Or configure directly
# config = TrustwiseConfig(
#     api_key="your-api-key",
#     base_url="https://api.trustwise.ai"
# )

client = TrustwiseClient(config)

# Evaluate a metric
result = client.metrics.faithfulness(
    query="What is the capital of France?",
    response="The capital of France is Paris.",
    context=[{"chunk_text": "Paris is the capital of France.", "chunk_id": "doc:idx:1"}],
)

# Manage agents
agents = client.agents.list(limit=5)

# Manage guardrails
guardrails = client.guardrails.list()
```

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

### Get Guardrail Provider Schemas

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/platform/client.html

Retrieves the configuration schemas for all guardrail providers.

```APIDOC
## Get Guardrail Provider Schemas

### Description
Get the configuration schemas for all guardrail providers.

### Method
GET

### Endpoint
/v1/guardrail-provider-schemas

### Response
#### Success Response (200)
- **schemas** (dict) - A dictionary containing configuration schemas for each provider.
```

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

### Get Guardrail Evaluation

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/platform/client.html

Retrieves a specific guardrail evaluation result by its IDs.

```APIDOC
## Get Guardrail Evaluation

### Description
Get a specific guardrail evaluation result.

### Method
GET

### Endpoint
/v1/guardrails/{guardrail_id}/evaluations/{evaluation_id}

### Parameters
#### Path Parameters
- **guardrail_id** (str) - Required - Unique guardrail identifier.
- **evaluation_id** (str) - Required - Unique evaluation identifier.
#### Query Parameters
- **raw** (bool) - Optional - Return the full API envelope.

### Response
#### Success Response (200)
- **result** (GuardrailEvaluationResult) - The guardrail evaluation result.
- **raw** (dict) - Full API envelope if raw=True.
```

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

### Run First Evaluation and List Agents with Trustwise SDK

Source: https://trustwiseai.github.io/trustwise/_sources/index.rst.txt

Set up the Trustwise client with your API key and run your first evaluation for faithfulness. This snippet also shows how to list agents. Ensure the TW_API_KEY environment variable is set.

```python
import os
from trustwise.sdk.config import TrustwiseConfig
from trustwise.sdk.platform import TrustwiseClient

config = TrustwiseConfig(api_key=os.environ["TW_API_KEY"])
client = TrustwiseClient(config)

# Evaluate faithfulness
result = client.metrics.faithfulness(
    query="What is the capital of France?",
    response="The capital of France is Paris.",
    context=[{"chunk_text": "Paris is the capital of France.", "chunk_id": "doc:idx:1"}],
)
print(result.score)

# List agents
agents = client.agents.list(limit=5)
```

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

### Initialize TrustwiseConfig with Environment Variables

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/config.html

Instantiate TrustwiseConfig without arguments to automatically load API key and base URL from environment variables TW_API_KEY and TW_BASE_URL. Defaults to 'https://api.trustwise.ai' if TW_BASE_URL is not set.

```python
>>> from trustwise.sdk.config import TrustwiseConfig
>>> # Using environment variables
>>> config = TrustwiseConfig()
```

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

### Get Agent

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/platform/client.html

Retrieves a single agent by its unique identifier within a project.

```APIDOC
## get

### Description
Get a single agent by ID.

### Method
GET

### Endpoint
/projects/{project_id}/agents/{agent_id}

### Parameters
#### Path Parameters
- **project_id** (str) - Required - Project identifier.
- **agent_id** (str) - Required - Unique agent identifier.

#### Query Parameters
- **raw** (bool) - Optional - Return the full API envelope.

### Response
#### Success Response (200)
- **agent** (dict[str, Any]) - Dictionary representing the agent.
- **envelope** (dict[str, Any]) - The raw API envelope if *raw* is True.
```

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

### Get Classification Results

Source: https://trustwiseai.github.io/trustwise/api.html

Retrieves the classification results for a specific evaluation within a project.

```APIDOC
## classification

### Description
Get classification results for an evaluation.

### Parameters
* **project_id** (str) - Project identifier.
* **evaluation_id** (str) - Unique evaluation identifier.

### Return Type
`dict`[`str`, `Any`]
```

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

### Initialize TrustwiseSDK

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/sdk.html

Initializes the Trustwise SDK with a configuration object. This class is deprecated; use TrustwiseClient instead.

```python
import warnings

from trustwise.sdk.client import TrustwiseHTTPClient
from trustwise.sdk.config import TrustwiseConfig
from trustwise.sdk.metrics import Metrics




[docs]
class TrustwiseSDK:
    """
    Main SDK class for Trustwise API access.

    Provides access to all metrics through version-specific paths.

    .. deprecated::
        Use :class:`trustwise.sdk.platform.client.TrustwiseClient` instead.
        ``TrustwiseClient`` is a superset of this SDK and covers all
        metrics, guardrails, policies, and platform APIs.
    """




[docs]
    def __init__(self, config: TrustwiseConfig) -> None:
        """
        Initialize the Trustwise SDK with path-based versioning support.

        Args:
            config: Trustwise configuration instance.
        """
        warnings.warn(
            "TrustwiseSDK is deprecated and will be removed in a future version. "
            "Please migrate to TrustwiseClient for continued support and enhanced features. "
            "See the migration guide for more details: "
            "https://trustwiseai.github.io/trustwise/sdk_migration.html",
            FutureWarning,
            stacklevel=2,
        )
        self.client = TrustwiseHTTPClient(config)
        self.metrics = Metrics(self.client)

```

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

### Initialize TrustwiseSDKAsync and Evaluate Faithfulness (Async)

Source: https://trustwiseai.github.io/trustwise/_sources/v4_metrics.rst.txt

Shows how to set up the asynchronous TrustwiseSDK and evaluate the faithfulness metric using await. This is suitable for non-blocking operations in async applications.

```python
import asyncio
from trustwise.sdk import TrustwiseSDKAsync
from trustwise.sdk.config import TrustwiseConfig
from trustwise.sdk.metrics.v4.types import ContextChunk

async def main():
    config = TrustwiseConfig(api_key="your-api-key")
    trustwise = TrustwiseSDKAsync(config)
    
    context = [
        ContextChunk(chunk_text="Paris is the capital of France.", chunk_id="doc:idx:1")
    ]
    
    result = await trustwise.metrics.faithfulness.evaluate(
        query="What is the capital of France?",
        response="The capital of France is Paris.",
        context=context
    )

asyncio.run(main())
```

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

### Get Agent Status History

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/platform/client.html

Retrieves the historical status records for a specific agent.

```APIDOC
## GET /projects/{project_id}/agents/{agent_id}/status

### Description
Get the status history for an agent.

### Method
GET

### Endpoint
/projects/{project_id}/agents/{agent_id}/status

### Parameters
#### Path Parameters
- **project_id** (str) - Required - Project identifier.
- **agent_id** (str) - Required - Unique agent identifier.

#### Query Parameters
- **limit** (int) - Optional - Maximum number of records to return.
- **offset** (int) - Optional - Number of records to skip.

### Response
#### Success Response (200)
- **AgentStatusHistory** (object) - Parsed agent status history.

#### Response Example
```json
{
  "history": [
    {
      "timestamp": "2023-10-27T10:00:00Z",
      "status": "online",
      "reason": ""
    }
  ],
  "total": 1
}
```
```

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

### Use Server-Side Guardrails with TrustwiseClient

Source: https://trustwiseai.github.io/trustwise/_sources/sdk_migration.rst.txt

Demonstrates how to create and evaluate against platform-managed guardrails using TrustwiseClient, replacing the old local guardrail system.

```python
# Create a guardrail on the platform (once)
guardrail = client.guardrails.create(
    name="My Guardrail",
    evaluators=[
        {"evaluator": "faithfulness", "threshold": 0.8},
        {"evaluator": "toxicity", "threshold": 0.3},
    ],
)

# Evaluate against it
result = client.guardrails.evaluate(
    guardrail["id"], input="user message", output="assistant reply"
)
```

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

### Risk Engine

Source: https://trustwiseai.github.io/trustwise/_sources/platform.rst.txt

Interact with the risk engine to get questionnaire scaffolds and classify risk.

```APIDOC
## Get Risk Engine Scaffold

### Description
Retrieves the questionnaire scaffold for a specified risk engine version.

### Method
`client.risk_engine.scaffold

### Parameters
- **version** (str) - Required - The version of the risk engine (e.g., "arc-v3").

### Request Example
```python
scaffold = client.risk_engine.scaffold("arc-v3")
```

## Classify Risk

### Description
Classifies risk based on provided answers for a given risk engine version.

### Method
`client.risk_engine.classify

### Parameters
- **version** (str) - Required - The version of the risk engine (e.g., "arc-v3").
- **answers** (dict) - Required - A dictionary of answers to the questionnaire.

### Request Example
```python
result = client.risk_engine.classify("arc-v3", answers={...})
```
```

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

### Get Specific Event

Source: https://trustwiseai.github.io/trustwise/_sources/platform.rst.txt

Retrieve details for a single event using its unique ID.

```python
# Get a specific event
event = client.events.get("event-id")
```

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

### PlatformHTTP Initialization

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/platform/_http.html

Initializes the PlatformHTTP client with configuration and sets up the requests session with necessary headers for authentication and content negotiation.

```python
def __init__(self, config: TrustwiseConfig, *, timeout: int = 30) -> None:
    self._base = config.base_url.rstrip("/")
    self._timeout = timeout
    self._session = requests.Session()
    self._session.headers.update({
        "Authorization": f"Bearer {config.api_key}",
        "Content-Type": "application/json",
        "Accept": "application/json",
    })
    logger.debug("Initialized PlatformHTTP with base URL: %s", self._base)
```

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

### TrustwiseClient Initialization and Usage

Source: https://trustwiseai.github.io/trustwise/api.html

The TrustwiseClient is the recommended client for new development, offering a unified interface to platform resources. It provides access to agents, gateway, guardrails, policies, evaluations, and more through sub-clients. You can initialize the client with a configuration object and use its sub-clients to interact with specific resources.

```APIDOC
## TrustwiseClient

### Description
User-friendly client for the Trustwise Harmony Platform API. Provides access to all platform resources through sub-clients: `agents`, `gateway`, `guardrails`, `policies`, `evaluations`, `evaluators`, `organizations`, `projects`, `risk_engine`, `generation`, `metrics`, `events`, `subscriptions`, and `runtime_providers`.

### Initialization
```python
from trustwise.sdk.config import TrustwiseConfig
from trustwise.sdk.platform import TrustwiseClient

config = TrustwiseConfig(api_key="tw-...")
client = TrustwiseClient(config)
```

### Usage Examples

#### Agents
```python
# Agents are now project-scoped
agent = client.agents.get("project-id", "agent-id")
print(agent.name)

agents = client.agents.list("project-id", search="my-bot", limit=5)
for a in agents:
    print(a.name)
```

### Parameters

- **config** (_TrustwiseConfig_): Configuration object for the client.
- **\*\*kwargs**: Additional keyword arguments.
- **timeout** (_int_): Request timeout in seconds. Defaults to 30.
```

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

### trustwise.sdk.platform.client.SubscriptionResource.get

Source: https://trustwiseai.github.io/trustwise/api.html

Get a single event subscription by its ID. Accepts 'subscription_id' and optional 'raw' boolean.

```APIDOC
## get

### Description
Get a single event subscription.

### Parameters
* **subscription_id** (`str`) – Unique subscription identifier.
* **raw** (`bool`) – Return the full API envelope.
```

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

### trustwise.sdk.platform.client.EventResource.get

Source: https://trustwiseai.github.io/trustwise/api.html

Get a single event notification by its ID. Accepts 'event_id' and optional 'raw' boolean.

```APIDOC
## get

### Description
Get a single event notification.

### Parameters
* **event_id** (`str`) – Unique event identifier.
* **raw** (`bool`) – Return the full API envelope.
```

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

### Initialize TrustwiseClient

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/platform/client.html

Instantiate the TrustwiseClient with a TrustwiseConfig object. The client provides access to various platform resources via sub-clients. You can optionally set an HTTP request timeout.

```python
from trustwise.sdk.config import TrustwiseConfig
from trustwise.sdk.platform import TrustwiseClient

config = TrustwiseConfig(api_key="tw-...")
client = TrustwiseClient(config)
```

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

### Get Organization by ID

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/platform/client.html

Retrieves a single organization's details using its unique identifier.

```APIDOC
## Get Organization by ID

### Description
Get a single organization by ID.

### Method
GET

### Endpoint
/v1/organizations/{org_id}

### Parameters
#### Path Parameters
- **org_id** (str) - Required - Unique organization identifier.
#### Query Parameters
- **raw** (bool) - Optional - Return the full API envelope.

### Response
#### Success Response (200)
- **data** (OrganizationResponse) - The organization object if raw=False.
- **envelope** (dict) - The full API envelope if raw=True.
```

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

### Initialize TrustwiseConfig with Direct Values

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/config.html

Instantiate TrustwiseConfig by providing the API key and base URL directly as arguments. This overrides any environment variable settings.

```python
>>> # Using direct initialization
>>> config = TrustwiseConfig(api_key="your-api-key", base_url="https://api.trustwise.ai")
```

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

### Get Current Default Version

Source: https://trustwiseai.github.io/trustwise/versioning.html

Accesses the current default API version being used by the SDK.

```APIDOC
## Get Current Default Version

### Description
Accesses the current default API version being used by the SDK for metrics.

### Method
`trustwise.metrics.version`

### Parameters
None

### Response
- `version` (string): The current default version string.
```

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

### TrustwiseSDK Initialization

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/sdk.html

Initializes the Trustwise SDK with a configuration object. This class is deprecated and users should migrate to TrustwiseClient.

```APIDOC
## __init__ TrustwiseSDK

### Description
Initialize the Trustwise SDK with path-based versioning support.

### Args
- **config** (TrustwiseConfig) - Required - Trustwise configuration instance.

### Deprecation Warning
This class is deprecated. Use :class:`trustwise.sdk.platform.client.TrustwiseClient` instead.
```

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

### Replace Client Instantiation: TrustwiseSDK to TrustwiseClient

Source: https://trustwiseai.github.io/trustwise/_sources/sdk_migration.rst.txt

Update how the SDK or client is instantiated, changing from TrustwiseSDK to TrustwiseClient.

```diff
- sdk = TrustwiseSDK(config)
+ client = TrustwiseClient(config)
```

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

### Get Field Description

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/types.html

Retrieves the description for a given field from its Pydantic field information, if available.

```python
def _get_field_description(self, field_name: str, field_info: object | None = None) -> str:
    """Get the description for a field from its field info."""
    if field_info and hasattr(field_info, "description") and field_info.description:
```

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

### Get Available API Versions

Source: https://trustwiseai.github.io/trustwise/versioning.html

Retrieves a list of all available API versions supported by the Trustwise SDK.

```APIDOC
## Get Available API Versions

### Description
Retrieves a list of all available API versions supported by the Trustwise SDK.

### Method
`trustwise.get_versions()`

### Parameters
None

### Response
- `versions` (list): A list of available API version strings.
```

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

### List All Runtime Providers

Source: https://trustwiseai.github.io/trustwise/platform.html

Retrieve a list of all available runtime providers, which define connector types for agent deployment. This is useful for understanding deployment options.

```python
providers = client.runtime_providers.list()
```

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

### Create a Webhook Subscription

Source: https://trustwiseai.github.io/trustwise/platform.html

Set up a new subscription to forward event notifications to a specified webhook URL. Filter events by type and severity.

```python
sub = client.subscriptions.create(
    name="My Webhook",
    event_type_filter=["evaluation.completed", "evaluation.failed"],
    severity_filter=["critical", "warning"],
    handler_type="webhook",
    handler_config={"url": "https://my-service.example.com/events"},
)
```

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

### Get SDK Version Async

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/async_sdk.html

Retrieve the current version of the SDK, which is set to 'v4' for the default metrics.

```python
@property
def version(self) -> str:
    return "v4"
```

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

### TrustwiseSDKAsync Initialization

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/async_sdk.html

Initializes the Trustwise asynchronous SDK client and exposes metrics directly.

```APIDOC
## TrustwiseSDKAsync

### Description
Async SDK entrypoint for Trustwise. Use this class to access async metrics.

### Initialization
```python
from trustwise.sdk import TrustwiseSDKAsync, TrustwiseConfig

config = TrustwiseConfig(...)
async_sdk = TrustwiseSDKAsync(config)

# Access metrics directly
answer_relevancy = async_sdk.answer_relevancy
carbon = async_sdk.carbon
# ... and so on for other metrics

# Access metrics through the metrics object
answer_relevancy_v4 = async_sdk.metrics.v4.answer_relevancy
```

### Attributes
- **client**: TrustwiseAsyncHTTPClient instance.
- **metrics**: MetricsAsync instance for accessing various metrics.
```

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

### Evaluate Faithfulness Metric

Source: https://trustwiseai.github.io/trustwise/_sources/usage.rst.txt

Evaluate the faithfulness metric by providing a query, response, and context. This is an example of a synchronous evaluation.

```python
# Evaluate faithfulness
trustwise.metrics.faithfulness.evaluate(
    query="What is the capital of France?",
    response="The capital of France is Paris.",
    context=[
        {"chunk_text": "Paris is the capital of France.", "chunk_id": "doc:idx:1"}
    ]
)
```

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

### Configuration

Source: https://trustwiseai.github.io/trustwise/_sources/api.rst.txt

Documentation for the configuration module of the Trustwise SDK.

```APIDOC
## Configuration

.. automodule:: trustwise.sdk.config
   :members:
   :undoc-members:
   :show-inheritance:
```

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

### TrustwiseSDKAsync Initialization

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/async_sdk.html

Initializes the asynchronous Trustwise SDK client.

```APIDOC
## TrustwiseSDKAsync Initialization

Initializes the asynchronous client for the Trustwise SDK.

### `__init__()`

Constructor for the `TrustwiseSDKAsync` class.
```

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

### Get Single Evaluation

Source: https://trustwiseai.github.io/trustwise/api.html

Retrieves a single evaluation by its unique identifier within a project. Supports raw envelope returns.

```APIDOC
## get

### Description
Get a single evaluation by ID.

### Parameters
* **project_id** (str) - Project identifier.
* **evaluation_id** (str) - Unique evaluation identifier.
* **raw** (bool) - Return the full API envelope.

### Return Type
`dict`[`str`, `Any`]
```

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

### V4 Metrics Evaluation

Source: https://trustwiseai.github.io/trustwise/usage.html

Demonstrates how to evaluate various V4 metrics using the SDK.

```APIDOC
## V4 Metrics Evaluation

### Description
This section covers the evaluation of various metrics available in V4 of the Trustwise SDK.

### Available Metrics and Methods
- **Adherence**: `metrics.adherence.evaluate()`
- **Answer Relevancy**: `metrics.answer_relevancy.evaluate()`
- **Clarity**: `metrics.clarity.evaluate()`
- **Completion**: `metrics.completion.evaluate()`
- **Context Relevancy**: `metrics.context_relevancy.evaluate()`
- **Faithfulness**: `metrics.faithfulness.evaluate()`
- **Formality**: `metrics.formality.evaluate()`
- **Helpfulness**: `metrics.helpfulness.evaluate()`
- **PII Detection**: `metrics.pii.evaluate()`
- **Prompt Manipulation**: `metrics.prompt_manipulation.evaluate()`
- **Refusal**: `metrics.refusal.evaluate()`
- **Sensitivity**: `metrics.sensitivity.evaluate()`
- **Stability**: `metrics.stability.evaluate()`
- **Simplicity**: `metrics.simplicity.evaluate()`
- **Tone**: `metrics.tone.evaluate()`
- **Toxicity**: `metrics.toxicity.evaluate()`
- **Carbon**: `metrics.carbon.evaluate()`

### Usage Example (Conceptual)
```python
# Assuming 'client' is an initialized TrustwiseClient instance

# Example for Faithfulness metric
faithfulness_result = client.metrics.faithfulness.evaluate(prompt="...", response="...", context="...")
print(faithfulness_result)

# Example for Clarity metric
clarity_result = client.metrics.clarity.evaluate(response="...")
print(clarity_result)
```

### Parameters
(Specific parameters for each metric's `evaluate` method are not detailed in the source but typically include prompt, response, context, etc., depending on the metric.)

### Response
(The structure of the response for each metric evaluation is not detailed in the source.)
```

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

### Risk Engine - Scaffold and Classification

Source: https://trustwiseai.github.io/trustwise/platform.html

Interact with the risk engine to get questionnaire scaffolds and classify risk based on answers.

```APIDOC
## Risk Engine

### Description
Utilize the risk engine to obtain questionnaire scaffolds for risk assessment and to classify risk based on provided answers.

### Methods

#### Get Questionnaire Scaffold
```python
scaffold = client.risk_engine.scaffold("arc-v3")
```

#### Classify Risk
```python
result = client.risk_engine.classify("arc-v3", answers={...})
```
```

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

### Policies - Management and Evaluation

Source: https://trustwiseai.github.io/trustwise/platform.html

Manage policies, including listing, evaluating content against single or multiple policies, and getting recommendations.

```APIDOC
## Policies

### Description
Manage policies, including listing, evaluating content against policies, and obtaining policy recommendations based on tags.

### Methods

#### List Policies
```python
policies = client.policies.list()
```

#### Evaluate Content Against Policy
```python
result = client.policies.evaluate(policy_id, input="test input")
```

#### Evaluate Content Against Multiple Policies
```python
results = client.policies.evaluate_batch(
    policy_ids=["policy-1", "policy-2"],
    input={"text": "test input"},
)
```

#### Get Policy Recommendations
```python
recs = client.policies.recommend(tags=["safety", "pii"])
```
```

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

### Synchronous V4 Metrics Evaluation

Source: https://trustwiseai.github.io/trustwise/v4_metrics.html

Demonstrates how to use the Trustwise SDK for synchronous evaluation of v4 metrics, including setting up the configuration and defining context chunks.

```python
from trustwise.sdk import TrustwiseSDK
from trustwise.sdk.config import TrustwiseConfig
from trustwise.sdk.metrics.v4.types import ContextChunk

config = TrustwiseConfig(api_key="your-api-key")
trustwise = TrustwiseSDK(config)

# v4 context format
context = [
    ContextChunk(chunk_text="Paris is the capital of France.", chunk_id="doc:idx:1")
]

# v4 metric calls (now default)
result = trustwise.metrics.faithfulness.evaluate(
    query="What is the capital of France?",
    response="The capital of France is Paris.",
    context=context
)

```

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

### Get Policy Recommendations

Source: https://trustwiseai.github.io/trustwise/api.html

Provides policy recommendations based on tags describing an agent or use-case. Additional context can be provided in the body.

```APIDOC
## policies.recommend

### Description
Get policy recommendations for a given agent or use-case.

### Parameters
- **tags** (list[`str`]) - Required - Tags describing the agent or use-case.
- **raw** (bool) - Optional - Return the full API envelope.
- **body** (Any) - Optional - Extra context for recommendation, such as `agent_id`, `description`, etc.

### Return Type
`list`[`PolicyRecommendation`]

### Returns
List of `PolicyRecommendation`, or the raw envelope.
```

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

### Get a Single Agent

Source: https://trustwiseai.github.io/trustwise/api.html

Retrieve a specific agent by its ID within a project. Set `raw=True` to obtain the full API envelope.

```python
agent = client.agents.get(project_id="project-id", agent_id="agent-id")
```

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

### TrustwiseClient Initialization

Source: https://trustwiseai.github.io/trustwise/changelog.html

Initializes the Trustwise client for interacting with the platform API.

```APIDOC
## TrustwiseClient

### Description

The main client class for interacting with the Trustwise Platform API. It provides access to various resources.

### Method

`TrustwiseClient.__init__(self, api_key: str, api_url: str = "https://api.trustwise.io/v1")`

### Parameters

#### Path Parameters

None

#### Query Parameters

None

#### Request Body

None

### Request Example

```python
from trustwise import TrustwiseClient

client = TrustwiseClient(api_key="YOUR_API_KEY")
```

### Response

None (initialization)

#### Success Response (200)

None

#### Response Example

None
```

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

### TrustwiseConfig Initialization

Source: https://trustwiseai.github.io/trustwise/api.html

Initialize the Trustwise API configuration. Supports environment variables or direct parameter passing for API key and base URL.

```APIDOC
## TrustwiseConfig

### Description
Configuration for Trustwise API. This class handles configuration for the Trustwise API, including authentication and endpoint URLs. It supports both direct initialization and environment variables.

### Parameters

#### Path Parameters
None

#### Query Parameters
None

#### Request Body
None

### Method
__init__

### Endpoint
N/A (Class constructor)

### Parameters
- **api_key** (str, Optional) - API key for authentication. If not provided, will be read from TW_API_KEY environment variable. Required for API access.
- **base_url** (str, Optional) - Base URL for API endpoints. If not provided, will be read from TW_BASE_URL environment variable or default to https://api.trustwise.ai.

### Request Example
```python
from trustwise.sdk.config import TrustwiseConfig

# Using environment variables
config = TrustwiseConfig()

# Using direct initialization
config = TrustwiseConfig(api_key="your-api-key", base_url="https://api.trustwise.ai")
```

### Response
#### Success Response
None (Constructor does not return a value, but initializes the object)

#### Response Example
None
```

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

### Get a Specific Event

Source: https://trustwiseai.github.io/trustwise/platform.html

Fetch details for a single event using its unique identifier. This is useful for inspecting individual event occurrences.

```python
event = client.events.get("event-id")
```

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

### Manage Evaluations: List and Get Audit Logs

Source: https://trustwiseai.github.io/trustwise/platform.html

List evaluations within a project and retrieve their audit logs. Evaluations are project-scoped.

```python
# List evaluations in a project
evals = client.evaluations.list("my-project-id", limit=10)

# Get audit logs
logs = client.evaluations.audit_logs("my-project-id")

```

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

### TrustwiseConfig Initialization

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/config.html

Initialize the TrustwiseConfig class, which handles API authentication and endpoint URLs. It supports reading credentials from environment variables or direct input.

```APIDOC
## TrustwiseConfig

### Description
Configuration for Trustwise API. This class handles configuration for the Trustwise API, including authentication and endpoint URLs. It supports both direct initialization and environment variables.

### Parameters
- **api_key** (str, Optional): API key for authentication. If not provided, will be read from TW_API_KEY environment variable. Required for API access.
- **base_url** (str, Optional): Base URL for API endpoints. If not provided, will be read from TW_BASE_URL environment variable or default to https://api.trustwise.ai.

### Environment Variables
- TW_API_KEY: API key for authentication
- TW_BASE_URL: Base URL for API endpoints (defaults to https://api.trustwise.ai)

### Example
```python
>>> from trustwise.sdk.config import TrustwiseConfig
>>> # Using environment variables
>>> config = TrustwiseConfig()
>>> # Using direct initialization
>>> config = TrustwiseConfig(api_key="your-api-key", base_url="https://api.trustwise.ai")
```

### Raises
- ValueError: If API key is missing or invalid, or if base URL is invalid
```

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

### Get Project Audit Logs

Source: https://trustwiseai.github.io/trustwise/_sources/platform.rst.txt

Fetch audit logs for a given project. This is essential for tracking changes and activities within the project.

```python
# Get audit logs
logs = client.evaluations.audit_logs("my-project-id")
```

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

### List Runtime Providers

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/platform/client.html

Retrieves a list of all available runtime providers. Use the `raw` parameter to get the full API response envelope.

```python
client.runtime_providers.list()
```

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

### TrustwiseConfig Initialization and Attributes

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/config.html

Details on initializing the TrustwiseConfig class and its core attributes like api_key and base_url.

```APIDOC
## TrustwiseConfig

### Description
Configuration class for the Trustwise SDK.

### Attributes
- **api_key** (str): The API key for authentication.
- **base_url** (str): The base URL for the Trustwise API.

### Methods
- **__init__(api_key: str, base_url: str = None)**: Initializes the configuration with an API key and an optional base URL.
- **_validate_url()**: Internal method to validate the base URL.
- **get_performance_url()**: Returns the URL for performance-related endpoints.
- **get_metrics_url()**: Returns the URL for metrics-related endpoints.
```

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

### similar_prompt

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/platform/client.html

Generate prompts similar to a seed prompt. Accepts optional `raw` flag and seed prompt along with generation parameters.

```APIDOC
## similar_prompt

Generate prompts similar to a seed prompt.

### Method

POST

### Endpoint

/v1/generation/similar_prompt

### Parameters

#### Query Parameters

- **raw** (bool) - Optional - Return the full API envelope.

#### Request Body

- **seed_prompt** (str) - Required - The prompt to generate similar prompts from.
- **...** (Any) - Other generation parameters.
```

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

### Get Available API Versions

Source: https://trustwiseai.github.io/trustwise/_modules/trustwise/sdk/sdk.html

Retrieves the available API versions for metrics. The returned dictionary maps 'metrics' to a list of its supported versions.

```python
def get_versions(self) -> dict[str, list[str]]:
    """
    Get the available API versions for the metrics.

    Returns:
        Dictionary mapping 'metrics' to its available versions.
        Example: {"metrics": ["v3"]}
    """
    return {"metrics": ["v3", "v4"]}

```
Tokens:
Raw