Optimizations
Reference for client.optimizations in the Future AGI Python and TypeScript SDKs: optimizations.create, optimizations.get, optimizations.list.
client.optimizations on fi.Client (Python) and client.optimizations on Client from @future-agi/sdk (TypeScript). The models these methods return are on Datasets & evals response models; how arguments, pagination, errors and strict mode work is in the overview.
Methods
| Method | Description |
|---|---|
optimizations.create | Create a dataset optimization run. |
optimizations.get | Return a dataset optimization run by its identifier. |
optimizations.list | List dataset optimization runs. |
client.optimizations
optimizations.create
Create a dataset optimization run.
/model-hub/dataset-optimization/ MCP tool: create_optimization_run
result = client.optimizations.create(
name="your-name",
column_id="your-column-id",
optimizer_algorithm="random_search",
)const result = await client.optimizations.create({
name: "your-name",
column_id: "your-column-id",
optimizer_algorithm: "random_search",
}); Signature
client.optimizations.create(
*,
name: str,
column_id: str,
optimizer_algorithm: Literal['random_search', 'bayesian', 'metaprompt', 'protegi', 'promptwizard', 'gepa'],
optimizer_model_id: str | None | NotGiven = NOT_GIVEN,
optimizer_config: dict[str, Any] | None | NotGiven = NOT_GIVEN,
user_eval_template_ids: list[str] | None | NotGiven = NOT_GIVEN,
extra_query: dict[str, Any] | None = None,
extra_body: dict[str, Any] | None = None,
options: RequestOptions | None = None,
) -> models.OptimizationsCreateResultclient.optimizations.create(params: OptimizationsCreateParams, options?: RequestOptions): Promise<models.OptimizationsCreateResult>
export interface OptimizationsCreateParams {
name: string;
column_id: string;
optimizer_algorithm: "random_search" | "bayesian" | "metaprompt" | "protegi" | "promptwizard" | "gepa";
optimizer_model_id?: string | null;
optimizer_config?: Record<string, unknown> | null;
user_eval_template_ids?: Array<string> | null;
} Parameters
random_searchbayesianmetapromptprotegipromptwizardgepa Returns
optimizations.get
Return a dataset optimization run by its identifier.
/model-hub/dataset-optimization/{id}/ MCP tool: get_optimization_run
result = client.optimizations.get("your-id")const result = await client.optimizations.get("your-id"); Signature
client.optimizations.get(
id: str,
*,
extra_query: dict[str, Any] | None = None,
options: RequestOptions | None = None,
) -> models.DatasetOptimizationDetailclient.optimizations.get(id: string, options?: RequestOptions): Promise<models.DatasetOptimizationDetail> Parameters
A UUID string identifying this optimize dataset.
Returns
optimizations.list
List dataset optimization runs.
/model-hub/dataset-optimization/ Paginated: returns a pager (how pagination works) · MCP tool: list_optimization_runs
for item in client.optimizations.list():
print(item)for await (const item of client.optimizations.list()) {
console.log(item);
} Signature
client.optimizations.list(
*,
dataset_id: str | None | NotGiven = NOT_GIVEN,
column_id: str | None | NotGiven = NOT_GIVEN,
develop_id: str | None | NotGiven = NOT_GIVEN,
limit: int | None | NotGiven = NOT_GIVEN,
extra_query: dict[str, Any] | None = None,
options: RequestOptions | None = None,
) -> SyncPager[Any]client.optimizations.list(params: OptimizationsListParams = {}, options?: RequestOptions): Pager<unknown>
export interface OptimizationsListParams {
dataset_id?: string | null;
column_id?: string | null;
develop_id?: string | null;
limit?: number | null;
} Parameters
The pager sends page itself; pass limit to change the page size.
Returns
The decoded response. The contract does not describe its shape, so it is untyped.
Questions & Discussion