import {z} from 'zod' export type CreateApiV1ModelsModelIdDeploymentsDeploymentIdRetryOptions = {baseUrl: string, BearerAuth?: string | undefined, fetch?: typeof fetch | undefined}; export const deploymentStatusV1 = z.enum(["BUILDING", "DEPLOYING", "DEPLOY_FAILED", "LOADING_MODEL", "ACTIVE", "UNHEALTHY", "BUILD_FAILED", "BUILD_STOPPED", "DEACTIVATING", "INACTIVE", "FAILED", "UPDATING", "SCALED_TO_ZERO", "WAKING_UP"]); export const autoscalingSettingsV1 = z.object({min_replica: z.number().int(), max_replica: z.number().int(), autoscaling_window: z.number().int().nullable(), scale_down_delay: z.number().int().nullable(), concurrency_target: z.number().int(), target_utilization_percentage: z.number().int().nullable(), target_in_flight_tokens: z.number().int().nullable().optional(), max_scale_down_rate: z.number().int().nullable().optional()}); export const deploymentV1 = z.object({id: z.string(), created_at: z.string(), name: z.string(), model_id: z.string(), is_production: z.boolean(), is_development: z.boolean(), status: deploymentStatusV1, active_replica_count: z.number().int(), autoscaling_settings: autoscalingSettingsV1.nullable(), instance_type_name: z.string().nullable(), environment: z.string().nullable(), labels: z.record(z.string(), z.unknown()).nullable().optional()}); export const retryDeploymentResponseV1 = z.object({retried: z.boolean(), reason: z.string().nullable().optional(), deployment: deploymentV1}); export const validateResponse = async ( schema: { readonly '~standard': { readonly validate: (value: unknown) => ({ readonly value: Output; readonly issues?: undefined } | { readonly issues: ReadonlyArray<{ readonly message: string; readonly path?: ReadonlyArray | undefined }> }) | Promise<{ readonly value: Output; readonly issues?: undefined } | { readonly issues: ReadonlyArray<{ readonly message: string; readonly path?: ReadonlyArray | undefined }> }> } }, value: unknown ): Promise => { const result = await schema['~standard'].validate(value) if (result.issues !== undefined) { const detail = result.issues .map(({ message, path }) => path?.length ? `${path.map((segment) => String(typeof segment === 'object' ? segment.key : segment)).join('.')}: ${message}` : message ) .join('; ') throw new Error(`Response does not match the schema: ${detail}`, { cause: { issues: result.issues } }) } return result.value }; /** * Retries a failed deployment * * POST /v1/models/{model_id}/deployments/{deployment_id}/retry */ export const createApiV1ModelsModelIdDeploymentsDeploymentIdRetry = async (options: CreateApiV1ModelsModelIdDeploymentsDeploymentIdRetryOptions) => { const {baseUrl, fetch: fetchFn = fetch, BearerAuth} = options ?? {} const query = new URLSearchParams() const headers = new Headers() headers.set('accept', "application/json") if (BearerAuth !== undefined) headers.set('authorization', `Bearer ${BearerAuth}`) const url = `${baseUrl.replace(/\/+$/, '')}/v1/models/${encodeURIComponent(String(args.model_id))}/deployments/${encodeURIComponent(String(args.deployment_id))}/retry${query.size > 0 ? `?${query}` : ''}` const res = await fetchFn(url, {method: "POST", headers}) if (!res.ok) throw new Error(`Request failed with status ${res.status}`, {cause: {status: res.status, body: await res.text()}}) return validateResponse(retryDeploymentResponseV1, await res.json()) };