@dhbw-cloud/llm-client 0.1.0-test.7 → 0.1.1

package/dist/index.d.ts

@@ -1,2 +1,2 @@
-export { blockMachine, createAccessRule, createKey, deleteAccessRule, deleteFleetPackage, deleteKey, forgetMachine, getFleet, getFleetInventoryCsv, getFleetPackage, getFleetProfile, getFleetReadme, getMe, getUsage, listAccessRules, listKeys, listTiers, type Options, searchGroups, unblockMachine, updateAccessRule } from './sdk.gen';
-export type { AccessRole, AccessRule, BlockMachineData, BlockMachineError, BlockMachineErrors, BlockMachineResponse, BlockMachineResponses, ClientOptions, CreateAccessRuleData, CreateAccessRuleError, CreateAccessRuleErrors, CreateAccessRuleResponse, CreateAccessRuleResponses, CreateKeyData, CreateKeyError, CreateKeyErrors, CreateKeyResponse, CreateKeyResponses, DeleteAccessRuleData, DeleteAccessRuleError, DeleteAccessRuleErrors, DeleteAccessRuleResponse, DeleteAccessRuleResponses, DeleteFleetPackageData, DeleteFleetPackageResponse, DeleteFleetPackageResponses, DeleteKeyData, DeleteKeyError, DeleteKeyErrors, DeleteKeyResponse, DeleteKeyResponses, FleetFleetView, FleetPackageInfo, FleetPeerView, FleetProfile, ForgetMachineData, ForgetMachineError, ForgetMachineErrors, ForgetMachineResponse, ForgetMachineResponses, GetFleetData, GetFleetError, GetFleetErrors, GetFleetInventoryCsvData, GetFleetInventoryCsvResponse, GetFleetInventoryCsvResponses, GetFleetPackageData, GetFleetPackageError, GetFleetPackageErrors, GetFleetPackageResponse, GetFleetPackageResponses, GetFleetProfileData, GetFleetProfileResponse, GetFleetProfileResponses, GetFleetReadmeData, GetFleetReadmeResponse, GetFleetReadmeResponses, GetFleetResponse, GetFleetResponses, GetMeData, GetMeError, GetMeErrors, GetMeResponse, GetMeResponses, GetUsageData, GetUsageError, GetUsageErrors, GetUsageResponse, GetUsageResponses, KeysKeyView, KeysUsage, ListAccessRulesData, ListAccessRulesError, ListAccessRulesErrors, ListAccessRulesResponse, ListAccessRulesResponses, ListKeysData, ListKeysError, ListKeysErrors, ListKeysResponse, ListKeysResponses, ListTiersData, ListTiersResponse, ListTiersResponses, RoleproviderGroup, SearchGroupsData, SearchGroupsError, SearchGroupsErrors, SearchGroupsResponse, SearchGroupsResponses, UnblockMachineData, UnblockMachineError, UnblockMachineErrors, UnblockMachineResponse, UnblockMachineResponses, UpdateAccessRuleData, UpdateAccessRuleError, UpdateAccessRuleErrors, UpdateAccessRuleResponse, UpdateAccessRuleResponses, WebserverCreateKeyRequest, WebserverCreateKeyResponse, WebserverErrorResponse, WebserverMeResponse, WebserverRuleRequest, WebserverRuleRequest2 } from './types.gen';
+export { blockMachine, createAccessRule, createGpuEnvironment, createKey, deleteAccessRule, deleteFleetPackage, deleteGpuEnvironment, deleteKey, forgetMachine, getFleet, getFleetInference, getFleetInventoryCsv, getFleetPackage, getFleetProfile, getFleetReadme, getGpuEnvironment, getGpuEnvironmentLog, getMe, getUsage, listAccessRules, listGpuEnvironments, listGpuServers, listGpuTiers, listKeys, listTiers, type Options, rebuildGpuEnvironment, searchPrincipals, startGpuEnvironment, stopGpuServer, unblockMachine, updateAccessRule } from './sdk.gen';
+export type { AccessRole, AccessRule, BlockMachineData, BlockMachineError, BlockMachineErrors, BlockMachineResponse, BlockMachineResponses, ClientOptions, CreateAccessRuleData, CreateAccessRuleError, CreateAccessRuleErrors, CreateAccessRuleResponse, CreateAccessRuleResponses, CreateGpuEnvironmentData, CreateGpuEnvironmentError, CreateGpuEnvironmentErrors, CreateGpuEnvironmentResponse, CreateGpuEnvironmentResponses, CreateKeyData, CreateKeyError, CreateKeyErrors, CreateKeyResponse, CreateKeyResponses, DeleteAccessRuleData, DeleteAccessRuleError, DeleteAccessRuleErrors, DeleteAccessRuleResponse, DeleteAccessRuleResponses, DeleteFleetPackageData, DeleteFleetPackageResponse, DeleteFleetPackageResponses, DeleteGpuEnvironmentData, DeleteGpuEnvironmentError, DeleteGpuEnvironmentErrors, DeleteGpuEnvironmentResponse, DeleteGpuEnvironmentResponses, DeleteKeyData, DeleteKeyError, DeleteKeyErrors, DeleteKeyResponse, DeleteKeyResponses, FleetFleetView, FleetPackageInfo, FleetPeerView, FleetProfile, ForgetMachineData, ForgetMachineError, ForgetMachineErrors, ForgetMachineResponse, ForgetMachineResponses, GetFleetData, GetFleetError, GetFleetErrors, GetFleetInferenceData, GetFleetInferenceError, GetFleetInferenceErrors, GetFleetInferenceResponse, GetFleetInferenceResponses, GetFleetInventoryCsvData, GetFleetInventoryCsvResponse, GetFleetInventoryCsvResponses, GetFleetPackageData, GetFleetPackageError, GetFleetPackageErrors, GetFleetPackageResponse, GetFleetPackageResponses, GetFleetProfileData, GetFleetProfileResponse, GetFleetProfileResponses, GetFleetReadmeData, GetFleetReadmeResponse, GetFleetReadmeResponses, GetFleetResponse, GetFleetResponses, GetGpuEnvironmentData, GetGpuEnvironmentError, GetGpuEnvironmentErrors, GetGpuEnvironmentLogData, GetGpuEnvironmentLogError, GetGpuEnvironmentLogErrors, GetGpuEnvironmentLogResponse, GetGpuEnvironmentLogResponses, GetGpuEnvironmentResponse, GetGpuEnvironmentResponses, GetMeData, GetMeError, GetMeErrors, GetMeResponse, GetMeResponses, GetUsageData, GetUsageError, GetUsageErrors, GetUsageResponse, GetUsageResponses, GpuEnvironment, GpuInferencePod, GpuInferencePool, GpuServer, GpuStatus, KeysKeyView, KeysUsage, ListAccessRulesData, ListAccessRulesError, ListAccessRulesErrors, ListAccessRulesResponse, ListAccessRulesResponses, ListGpuEnvironmentsData, ListGpuEnvironmentsError, ListGpuEnvironmentsErrors, ListGpuEnvironmentsResponse, ListGpuEnvironmentsResponses, ListGpuServersData, ListGpuServersError, ListGpuServersErrors, ListGpuServersResponse, ListGpuServersResponses, ListGpuTiersData, ListGpuTiersResponse, ListGpuTiersResponses, ListKeysData, ListKeysError, ListKeysErrors, ListKeysResponse, ListKeysResponses, ListTiersData, ListTiersResponse, ListTiersResponses, RebuildGpuEnvironmentData, RebuildGpuEnvironmentError, RebuildGpuEnvironmentErrors, RebuildGpuEnvironmentResponse, RebuildGpuEnvironmentResponses, RoleproviderGroup, SearchPrincipalsData, SearchPrincipalsError, SearchPrincipalsErrors, SearchPrincipalsResponse, SearchPrincipalsResponses, StartGpuEnvironmentData, StartGpuEnvironmentError, StartGpuEnvironmentErrors, StartGpuEnvironmentResponse, StartGpuEnvironmentResponses, StopGpuServerData, StopGpuServerError, StopGpuServerErrors, StopGpuServerResponse, StopGpuServerResponses, UnblockMachineData, UnblockMachineError, UnblockMachineErrors, UnblockMachineResponse, UnblockMachineResponses, UpdateAccessRuleData, UpdateAccessRuleError, UpdateAccessRuleErrors, UpdateAccessRuleResponse, UpdateAccessRuleResponses, WebserverBuildLogResponse, WebserverCreateKeyRequest, WebserverCreateKeyResponse, WebserverEnvironmentRequest, WebserverErrorResponse, WebserverMeResponse, WebserverPrincipalSearchResponse, WebserverRuleRequest, WebserverRuleRequest2, WebserverStartRequest, WebserverStartResponse } from './types.gen';

package/dist/index.mjs

@@ -820,6 +820,11 @@ var getFleet = (options) => (options?.client ?? client).get({
   url: "/v1/fleet",
   ...options
 });
+var getFleetInference = (options) => (options?.client ?? client).get({
+  security: [{ name: "Authorization", type: "apiKey" }],
+  url: "/v1/fleet/inference",
+  ...options
+});
 var getFleetInventoryCsv = (options) => (options?.client ?? client).get({
   security: [{ name: "Authorization", type: "apiKey" }],
   url: "/v1/fleet/inventory.csv",
@@ -860,9 +865,62 @@ var unblockMachine = (options) => (options.client ?? client).post({
   url: "/v1/fleet/{serial}/unblock",
   ...options
 });
-var searchGroups = (options) => (options.client ?? client).get({
+var listGpuTiers = (options) => (options?.client ?? client).get({
+  security: [{ name: "Authorization", type: "apiKey" }],
+  url: "/v1/gpu-tiers",
+  ...options
+});
+var listGpuEnvironments = (options) => (options?.client ?? client).get({
   security: [{ name: "Authorization", type: "apiKey" }],
-  url: "/v1/groups",
+  url: "/v1/gpu/environments",
+  ...options
+});
+var createGpuEnvironment = (options) => (options.client ?? client).post({
+  security: [{ name: "Authorization", type: "apiKey" }],
+  url: "/v1/gpu/environments",
+  ...options,
+  headers: {
+    "Content-Type": "application/json",
+    ...options.headers
+  }
+});
+var deleteGpuEnvironment = (options) => (options.client ?? client).delete({
+  security: [{ name: "Authorization", type: "apiKey" }],
+  url: "/v1/gpu/environments/{id}",
+  ...options
+});
+var getGpuEnvironment = (options) => (options.client ?? client).get({
+  security: [{ name: "Authorization", type: "apiKey" }],
+  url: "/v1/gpu/environments/{id}",
+  ...options
+});
+var getGpuEnvironmentLog = (options) => (options.client ?? client).get({
+  security: [{ name: "Authorization", type: "apiKey" }],
+  url: "/v1/gpu/environments/{id}/log",
+  ...options
+});
+var rebuildGpuEnvironment = (options) => (options.client ?? client).post({
+  security: [{ name: "Authorization", type: "apiKey" }],
+  url: "/v1/gpu/environments/{id}/rebuild",
+  ...options
+});
+var startGpuEnvironment = (options) => (options.client ?? client).post({
+  security: [{ name: "Authorization", type: "apiKey" }],
+  url: "/v1/gpu/environments/{id}/start",
+  ...options,
+  headers: {
+    "Content-Type": "application/json",
+    ...options.headers
+  }
+});
+var listGpuServers = (options) => (options?.client ?? client).get({
+  security: [{ name: "Authorization", type: "apiKey" }],
+  url: "/v1/gpu/servers",
+  ...options
+});
+var stopGpuServer = (options) => (options.client ?? client).delete({
+  security: [{ name: "Authorization", type: "apiKey" }],
+  url: "/v1/gpu/servers/{name}",
   ...options
 });
 var listKeys = (options) => (options?.client ?? client).get({
@@ -889,6 +947,11 @@ var getMe = (options) => (options?.client ?? client).get({
   url: "/v1/me",
   ...options
 });
+var searchPrincipals = (options) => (options?.client ?? client).get({
+  security: [{ name: "Authorization", type: "apiKey" }],
+  url: "/v1/principals/search",
+  ...options
+});
 var listTiers = (options) => (options?.client ?? client).get({
   security: [{ name: "Authorization", type: "apiKey" }],
   url: "/v1/tiers",
@@ -903,22 +966,33 @@ export {
   blockMachine,
   client,
   createAccessRule,
+  createGpuEnvironment,
   createKey,
   deleteAccessRule,
   deleteFleetPackage,
+  deleteGpuEnvironment,
   deleteKey,
   forgetMachine,
   getFleet,
+  getFleetInference,
   getFleetInventoryCsv,
   getFleetPackage,
   getFleetProfile,
   getFleetReadme,
+  getGpuEnvironment,
+  getGpuEnvironmentLog,
   getMe,
   getUsage,
   listAccessRules,
+  listGpuEnvironments,
+  listGpuServers,
+  listGpuTiers,
   listKeys,
   listTiers,
-  searchGroups,
+  rebuildGpuEnvironment,
+  searchPrincipals,
+  startGpuEnvironment,
+  stopGpuServer,
   unblockMachine,
   updateAccessRule
 };

package/dist/sdk.gen.d.ts

@@ -1,5 +1,5 @@
 import type { Client, ClientMeta, Options as Options2, RequestResult, TDataShape } from './client';
-import type { BlockMachineData, BlockMachineErrors, BlockMachineResponses, CreateAccessRuleData, CreateAccessRuleErrors, CreateAccessRuleResponses, CreateKeyData, CreateKeyErrors, CreateKeyResponses, DeleteAccessRuleData, DeleteAccessRuleErrors, DeleteAccessRuleResponses, DeleteFleetPackageData, DeleteFleetPackageResponses, DeleteKeyData, DeleteKeyErrors, DeleteKeyResponses, ForgetMachineData, ForgetMachineErrors, ForgetMachineResponses, GetFleetData, GetFleetErrors, GetFleetInventoryCsvData, GetFleetInventoryCsvResponses, GetFleetPackageData, GetFleetPackageErrors, GetFleetPackageResponses, GetFleetProfileData, GetFleetProfileResponses, GetFleetReadmeData, GetFleetReadmeResponses, GetFleetResponses, GetMeData, GetMeErrors, GetMeResponses, GetUsageData, GetUsageErrors, GetUsageResponses, ListAccessRulesData, ListAccessRulesErrors, ListAccessRulesResponses, ListKeysData, ListKeysErrors, ListKeysResponses, ListTiersData, ListTiersResponses, SearchGroupsData, SearchGroupsErrors, SearchGroupsResponses, UnblockMachineData, UnblockMachineErrors, UnblockMachineResponses, UpdateAccessRuleData, UpdateAccessRuleErrors, UpdateAccessRuleResponses } from './types.gen';
+import type { BlockMachineData, BlockMachineErrors, BlockMachineResponses, CreateAccessRuleData, CreateAccessRuleErrors, CreateAccessRuleResponses, CreateGpuEnvironmentData, CreateGpuEnvironmentErrors, CreateGpuEnvironmentResponses, CreateKeyData, CreateKeyErrors, CreateKeyResponses, DeleteAccessRuleData, DeleteAccessRuleErrors, DeleteAccessRuleResponses, DeleteFleetPackageData, DeleteFleetPackageResponses, DeleteGpuEnvironmentData, DeleteGpuEnvironmentErrors, DeleteGpuEnvironmentResponses, DeleteKeyData, DeleteKeyErrors, DeleteKeyResponses, ForgetMachineData, ForgetMachineErrors, ForgetMachineResponses, GetFleetData, GetFleetErrors, GetFleetInferenceData, GetFleetInferenceErrors, GetFleetInferenceResponses, GetFleetInventoryCsvData, GetFleetInventoryCsvResponses, GetFleetPackageData, GetFleetPackageErrors, GetFleetPackageResponses, GetFleetProfileData, GetFleetProfileResponses, GetFleetReadmeData, GetFleetReadmeResponses, GetFleetResponses, GetGpuEnvironmentData, GetGpuEnvironmentErrors, GetGpuEnvironmentLogData, GetGpuEnvironmentLogErrors, GetGpuEnvironmentLogResponses, GetGpuEnvironmentResponses, GetMeData, GetMeErrors, GetMeResponses, GetUsageData, GetUsageErrors, GetUsageResponses, ListAccessRulesData, ListAccessRulesErrors, ListAccessRulesResponses, ListGpuEnvironmentsData, ListGpuEnvironmentsErrors, ListGpuEnvironmentsResponses, ListGpuServersData, ListGpuServersErrors, ListGpuServersResponses, ListGpuTiersData, ListGpuTiersResponses, ListKeysData, ListKeysErrors, ListKeysResponses, ListTiersData, ListTiersResponses, RebuildGpuEnvironmentData, RebuildGpuEnvironmentErrors, RebuildGpuEnvironmentResponses, SearchPrincipalsData, SearchPrincipalsErrors, SearchPrincipalsResponses, StartGpuEnvironmentData, StartGpuEnvironmentErrors, StartGpuEnvironmentResponses, StopGpuServerData, StopGpuServerErrors, StopGpuServerResponses, UnblockMachineData, UnblockMachineErrors, UnblockMachineResponses, UpdateAccessRuleData, UpdateAccessRuleErrors, UpdateAccessRuleResponses } from './types.gen';
 export type Options<TData extends TDataShape = TDataShape, ThrowOnError extends boolean = boolean, TResponse = unknown> = Options2<TData, ThrowOnError, TResponse> & {
     /**
      * You can provide a client instance returned by `createClient()` instead of
@@ -37,6 +37,12 @@ export declare const updateAccessRule: <ThrowOnError extends boolean = false>(op
  * Machines with enrolment state, LiteLLM health, script state and the JAMF onboarding data.
  */
 export declare const getFleet: <ThrowOnError extends boolean = false>(options?: Options<GetFleetData, ThrowOnError>) => RequestResult<GetFleetResponses, GetFleetErrors, ThrowOnError>;
+/**
+ * The GPU cluster's inference pool
+ *
+ * vLLM replicas (one gap filler per GPU node, the base load) with node, model and state. enabled=false without the GPU part or without an inference pool.
+ */
+export declare const getFleetInference: <ThrowOnError extends boolean = false>(options?: Options<GetFleetInferenceData, ThrowOnError>) => RequestResult<GetFleetInferenceResponses, GetFleetInferenceErrors, ThrowOnError>;
 /**
  * Fleet inventory as CSV
  */
@@ -76,11 +82,59 @@ export declare const blockMachine: <ThrowOnError extends boolean = false>(option
  */
 export declare const unblockMachine: <ThrowOnError extends boolean = false>(options: Options<UnblockMachineData, ThrowOnError>) => RequestResult<UnblockMachineResponses, UnblockMachineErrors, ThrowOnError>;
 /**
- * Search role-provider groups
+ * GPU tiers a rule may use
+ *
+ * Empty when the GPU part is not enabled.
+ */
+export declare const listGpuTiers: <ThrowOnError extends boolean = false>(options?: Options<ListGpuTiersData, ThrowOnError>) => RequestResult<ListGpuTiersResponses, unknown, ThrowOnError>;
+/**
+ * My environments
+ *
+ * Environments built from Git repositories, newest first, with the current build status.
+ */
+export declare const listGpuEnvironments: <ThrowOnError extends boolean = false>(options?: Options<ListGpuEnvironmentsData, ThrowOnError>) => RequestResult<ListGpuEnvironmentsResponses, ListGpuEnvironmentsErrors, ThrowOnError>;
+/**
+ * Add an environment from a Git repository
+ *
+ * Resolves the ref to a commit. If Harbor already has the image for that commit, the environment is ready at once; otherwise a build starts (status building). Public repositories on the allowed hosts only.
+ */
+export declare const createGpuEnvironment: <ThrowOnError extends boolean = false>(options: Options<CreateGpuEnvironmentData, ThrowOnError>) => RequestResult<CreateGpuEnvironmentResponses, CreateGpuEnvironmentErrors, ThrowOnError>;
+/**
+ * Forget an environment
+ *
+ * The image stays in Harbor until its retention rules remove it.
+ */
+export declare const deleteGpuEnvironment: <ThrowOnError extends boolean = false>(options: Options<DeleteGpuEnvironmentData, ThrowOnError>) => RequestResult<DeleteGpuEnvironmentResponses, DeleteGpuEnvironmentErrors, ThrowOnError>;
+/**
+ * One environment with its build status
+ */
+export declare const getGpuEnvironment: <ThrowOnError extends boolean = false>(options: Options<GetGpuEnvironmentData, ThrowOnError>) => RequestResult<GetGpuEnvironmentResponses, GetGpuEnvironmentErrors, ThrowOnError>;
+/**
+ * Tail of the build log
+ *
+ * Available while the build job exists (one day after it finished).
+ */
+export declare const getGpuEnvironmentLog: <ThrowOnError extends boolean = false>(options: Options<GetGpuEnvironmentLogData, ThrowOnError>) => RequestResult<GetGpuEnvironmentLogResponses, GetGpuEnvironmentLogErrors, ThrowOnError>;
+/**
+ * Build the environment again
+ */
+export declare const rebuildGpuEnvironment: <ThrowOnError extends boolean = false>(options: Options<RebuildGpuEnvironmentData, ThrowOnError>) => RequestResult<RebuildGpuEnvironmentResponses, RebuildGpuEnvironmentErrors, ThrowOnError>;
+/**
+ * Start a JupyterHub server for the environment
  *
- * For the access-rule editor: find group tokens by id or name.
+ * A named server per environment; the repository is checked out into ~/projekte/<repo>. With gpu=true it waits for a free GPU and counts against the GPU quota; JupyterHub's refusal comes back as 409 with its message.
  */
-export declare const searchGroups: <ThrowOnError extends boolean = false>(options: Options<SearchGroupsData, ThrowOnError>) => RequestResult<SearchGroupsResponses, SearchGroupsErrors, ThrowOnError>;
+export declare const startGpuEnvironment: <ThrowOnError extends boolean = false>(options: Options<StartGpuEnvironmentData, ThrowOnError>) => RequestResult<StartGpuEnvironmentResponses, StartGpuEnvironmentErrors, ThrowOnError>;
+/**
+ * My JupyterHub servers
+ */
+export declare const listGpuServers: <ThrowOnError extends boolean = false>(options?: Options<ListGpuServersData, ThrowOnError>) => RequestResult<ListGpuServersResponses, ListGpuServersErrors, ThrowOnError>;
+/**
+ * Stop one of my servers
+ *
+ * Named servers are removed after stopping; the home stays.
+ */
+export declare const stopGpuServer: <ThrowOnError extends boolean = false>(options: Options<StopGpuServerData, ThrowOnError>) => RequestResult<StopGpuServerResponses, StopGpuServerErrors, ThrowOnError>;
 /**
  * The caller's API keys
  */
@@ -101,6 +155,12 @@ export declare const deleteKey: <ThrowOnError extends boolean = false>(options:
  * Always 200 for an authenticated caller. Without access, role is empty.
  */
 export declare const getMe: <ThrowOnError extends boolean = false>(options?: Options<GetMeData, ThrowOnError>) => RequestResult<GetMeResponses, GetMeErrors, ThrowOnError>;
+/**
+ * Search groups and users
+ *
+ * For the access-rule editor. Groups match on ID, display name or description; users on their EMAIL ADDRESS ONLY and only once q is non-empty — the directory is not browsable by person or name.
+ */
+export declare const searchPrincipals: <ThrowOnError extends boolean = false>(options?: Options<SearchPrincipalsData, ThrowOnError>) => RequestResult<SearchPrincipalsResponses, SearchPrincipalsErrors, ThrowOnError>;
 /**
  * Quota tiers a rule may use
  */

package/dist/types.gen.d.ts

@@ -10,6 +10,7 @@ export type AccessRule = {
      */
     bootstrap?: boolean;
     comment?: string;
+    gpu_tier?: string;
     id?: number;
     role?: AccessRole;
     tier?: string;
@@ -69,6 +70,10 @@ export type FleetPeerView = {
     operator?: string;
     os?: string;
     pkg_version?: string;
+    /**
+     * the machine's own address in its local network (default route), as it reports it at enrolment; empty until a script that sends it has enrolled
+     */
+    primary_ip?: string;
     profile?: string;
     quiet_seconds?: number;
     ram_gb?: number;
@@ -86,6 +91,51 @@ export type FleetProfile = {
     weight?: number;
     wiredLimitMb?: number;
 };
+export type GpuEnvironment = {
+    branch?: string;
+    commit?: string;
+    created_at?: string;
+    git_url?: string;
+    id?: number;
+    image?: string;
+    message?: string;
+    name?: string;
+    owner?: string;
+    ref?: string;
+    status?: GpuStatus;
+    updated_at?: string;
+};
+export type GpuInferencePod = {
+    gpu_class?: string;
+    model?: string;
+    node?: string;
+    pod?: string;
+    replica?: string;
+    role?: 'gap-filler' | 'base-load';
+    /**
+     * start on the node, or creation while it still waits
+     */
+    since?: string;
+    /**
+     * name in the discovery job and in LiteLLM
+     */
+    site?: string;
+    state?: 'bereit' | 'startet' | 'wartet auf GPU' | 'wird beendet';
+};
+export type GpuInferencePool = {
+    enabled?: boolean;
+    pods?: Array<GpuInferencePod>;
+    url?: string;
+};
+export type GpuServer = {
+    image?: string;
+    name?: string;
+    pending?: string;
+    profile?: string;
+    ready?: boolean;
+    url?: string;
+};
+export type GpuStatus = 'building' | 'ready' | 'failed';
 export type KeysKeyView = {
     budget?: number;
     budget_duration?: string;
@@ -108,9 +158,16 @@ export type KeysUsage = {
     tpm?: number;
 };
 export type RoleproviderGroup = {
-    display_name?: string;
+    description?: string;
+    /**
+     * Display name, left out when it only repeats the group ID (imported groups).
+     */
+    label?: string;
     token?: string;
 };
+export type WebserverBuildLogResponse = {
+    log?: string;
+};
 export type WebserverCreateKeyRequest = {
     name: string;
 };
@@ -118,6 +175,14 @@ export type WebserverCreateKeyResponse = {
     key?: string;
     name?: string;
 };
+export type WebserverEnvironmentRequest = {
+    git_url: string;
+    name?: string;
+    /**
+     * Branch, tag or commit; empty = default branch.
+     */
+    ref?: string;
+};
 export type WebserverErrorResponse = {
     /**
      * Stable machine-readable reason.
@@ -136,16 +201,39 @@ export type WebserverMeResponse = {
     api_url?: string;
     chat_url?: string;
     email?: string;
+    /**
+     * GPU part: tier and JupyterHub, only with GPU access (and the GPU part enabled).
+     */
+    gpu_tier?: string;
+    jupyter_url?: string;
     matched_token?: string;
     role?: 'user' | 'fleet-admin' | 'admin';
     tier?: string;
 };
+export type WebserverPrincipalSearchResponse = {
+    groups?: Array<RoleproviderGroup>;
+    users?: Array<string>;
+};
 export type WebserverRuleRequest = {
     comment?: string;
+    /**
+     * GPU tier (GPU notebooks and environments); empty = no GPU access.
+     */
+    gpu_tier?: string;
     role: 'user' | 'fleet-admin' | 'admin';
     tier: string;
     token: string;
 };
+export type WebserverStartRequest = {
+    /**
+     * With a GPU (waits in the queue if all GPUs are busy; counts against the GPU quota).
+     */
+    gpu?: boolean;
+};
+export type WebserverStartResponse = {
+    server?: string;
+    url?: string;
+};
 export type WebserverRuleRequest2 = WebserverRuleRequest;
 export type ListAccessRulesData = {
     body?: never;
@@ -271,6 +359,26 @@ export type GetFleetResponses = {
     200: FleetFleetView;
 };
 export type GetFleetResponse = GetFleetResponses[keyof GetFleetResponses];
+export type GetFleetInferenceData = {
+    body?: never;
+    path?: never;
+    query?: never;
+    url: '/v1/fleet/inference';
+};
+export type GetFleetInferenceErrors = {
+    /**
+     * Forbidden
+     */
+    403: WebserverErrorResponse;
+};
+export type GetFleetInferenceError = GetFleetInferenceErrors[keyof GetFleetInferenceErrors];
+export type GetFleetInferenceResponses = {
+    /**
+     * OK
+     */
+    200: GpuInferencePool;
+};
+export type GetFleetInferenceResponse = GetFleetInferenceResponses[keyof GetFleetInferenceResponses];
 export type GetFleetInventoryCsvData = {
     body?: never;
     path?: never;
@@ -431,31 +539,239 @@ export type UnblockMachineResponses = {
     204: void;
 };
 export type UnblockMachineResponse = UnblockMachineResponses[keyof UnblockMachineResponses];
-export type SearchGroupsData = {
+export type ListGpuTiersData = {
     body?: never;
     path?: never;
-    query: {
+    query?: never;
+    url: '/v1/gpu-tiers';
+};
+export type ListGpuTiersResponses = {
+    /**
+     * OK
+     */
+    200: Array<string>;
+};
+export type ListGpuTiersResponse = ListGpuTiersResponses[keyof ListGpuTiersResponses];
+export type ListGpuEnvironmentsData = {
+    body?: never;
+    path?: never;
+    query?: never;
+    url: '/v1/gpu/environments';
+};
+export type ListGpuEnvironmentsErrors = {
+    /**
+     * Forbidden
+     */
+    403: WebserverErrorResponse;
+};
+export type ListGpuEnvironmentsError = ListGpuEnvironmentsErrors[keyof ListGpuEnvironmentsErrors];
+export type ListGpuEnvironmentsResponses = {
+    /**
+     * OK
+     */
+    200: Array<GpuEnvironment>;
+};
+export type ListGpuEnvironmentsResponse = ListGpuEnvironmentsResponses[keyof ListGpuEnvironmentsResponses];
+export type CreateGpuEnvironmentData = {
+    /**
+     * Repository
+     */
+    body: WebserverEnvironmentRequest;
+    path?: never;
+    query?: never;
+    url: '/v1/gpu/environments';
+};
+export type CreateGpuEnvironmentErrors = {
+    /**
+     * Bad Request
+     */
+    400: WebserverErrorResponse;
+    /**
+     * Bad Gateway
+     */
+    502: WebserverErrorResponse;
+};
+export type CreateGpuEnvironmentError = CreateGpuEnvironmentErrors[keyof CreateGpuEnvironmentErrors];
+export type CreateGpuEnvironmentResponses = {
+    /**
+     * Created
+     */
+    201: GpuEnvironment;
+};
+export type CreateGpuEnvironmentResponse = CreateGpuEnvironmentResponses[keyof CreateGpuEnvironmentResponses];
+export type DeleteGpuEnvironmentData = {
+    body?: never;
+    path: {
         /**
-         * Search text (min. 2 characters)
+         * Environment id
          */
-        q: string;
+        id: number;
     };
-    url: '/v1/groups';
+    query?: never;
+    url: '/v1/gpu/environments/{id}';
 };
-export type SearchGroupsErrors = {
+export type DeleteGpuEnvironmentErrors = {
+    /**
+     * Not Found
+     */
+    404: WebserverErrorResponse;
+};
+export type DeleteGpuEnvironmentError = DeleteGpuEnvironmentErrors[keyof DeleteGpuEnvironmentErrors];
+export type DeleteGpuEnvironmentResponses = {
+    /**
+     * No Content
+     */
+    204: void;
+};
+export type DeleteGpuEnvironmentResponse = DeleteGpuEnvironmentResponses[keyof DeleteGpuEnvironmentResponses];
+export type GetGpuEnvironmentData = {
+    body?: never;
+    path: {
+        /**
+         * Environment id
+         */
+        id: number;
+    };
+    query?: never;
+    url: '/v1/gpu/environments/{id}';
+};
+export type GetGpuEnvironmentErrors = {
+    /**
+     * Not Found
+     */
+    404: WebserverErrorResponse;
+};
+export type GetGpuEnvironmentError = GetGpuEnvironmentErrors[keyof GetGpuEnvironmentErrors];
+export type GetGpuEnvironmentResponses = {
+    /**
+     * OK
+     */
+    200: GpuEnvironment;
+};
+export type GetGpuEnvironmentResponse = GetGpuEnvironmentResponses[keyof GetGpuEnvironmentResponses];
+export type GetGpuEnvironmentLogData = {
+    body?: never;
+    path: {
+        /**
+         * Environment id
+         */
+        id: number;
+    };
+    query?: never;
+    url: '/v1/gpu/environments/{id}/log';
+};
+export type GetGpuEnvironmentLogErrors = {
+    /**
+     * Not Found
+     */
+    404: WebserverErrorResponse;
+};
+export type GetGpuEnvironmentLogError = GetGpuEnvironmentLogErrors[keyof GetGpuEnvironmentLogErrors];
+export type GetGpuEnvironmentLogResponses = {
+    /**
+     * OK
+     */
+    200: WebserverBuildLogResponse;
+};
+export type GetGpuEnvironmentLogResponse = GetGpuEnvironmentLogResponses[keyof GetGpuEnvironmentLogResponses];
+export type RebuildGpuEnvironmentData = {
+    body?: never;
+    path: {
+        /**
+         * Environment id
+         */
+        id: number;
+    };
+    query?: never;
+    url: '/v1/gpu/environments/{id}/rebuild';
+};
+export type RebuildGpuEnvironmentErrors = {
+    /**
+     * Not Found
+     */
+    404: WebserverErrorResponse;
+};
+export type RebuildGpuEnvironmentError = RebuildGpuEnvironmentErrors[keyof RebuildGpuEnvironmentErrors];
+export type RebuildGpuEnvironmentResponses = {
+    /**
+     * OK
+     */
+    200: GpuEnvironment;
+};
+export type RebuildGpuEnvironmentResponse = RebuildGpuEnvironmentResponses[keyof RebuildGpuEnvironmentResponses];
+export type StartGpuEnvironmentData = {
+    /**
+     * Options
+     */
+    body?: WebserverStartRequest;
+    path: {
+        /**
+         * Environment id
+         */
+        id: number;
+    };
+    query?: never;
+    url: '/v1/gpu/environments/{id}/start';
+};
+export type StartGpuEnvironmentErrors = {
+    /**
+     * Conflict
+     */
+    409: WebserverErrorResponse;
+};
+export type StartGpuEnvironmentError = StartGpuEnvironmentErrors[keyof StartGpuEnvironmentErrors];
+export type StartGpuEnvironmentResponses = {
+    /**
+     * Accepted
+     */
+    202: WebserverStartResponse;
+};
+export type StartGpuEnvironmentResponse = StartGpuEnvironmentResponses[keyof StartGpuEnvironmentResponses];
+export type ListGpuServersData = {
+    body?: never;
+    path?: never;
+    query?: never;
+    url: '/v1/gpu/servers';
+};
+export type ListGpuServersErrors = {
     /**
      * Bad Gateway
      */
     502: WebserverErrorResponse;
 };
-export type SearchGroupsError = SearchGroupsErrors[keyof SearchGroupsErrors];
-export type SearchGroupsResponses = {
+export type ListGpuServersError = ListGpuServersErrors[keyof ListGpuServersErrors];
+export type ListGpuServersResponses = {
     /**
      * OK
      */
-    200: Array<RoleproviderGroup>;
+    200: Array<GpuServer>;
 };
-export type SearchGroupsResponse = SearchGroupsResponses[keyof SearchGroupsResponses];
+export type ListGpuServersResponse = ListGpuServersResponses[keyof ListGpuServersResponses];
+export type StopGpuServerData = {
+    body?: never;
+    path: {
+        /**
+         * Server name
+         */
+        name: string;
+    };
+    query?: never;
+    url: '/v1/gpu/servers/{name}';
+};
+export type StopGpuServerErrors = {
+    /**
+     * Not Found
+     */
+    404: WebserverErrorResponse;
+};
+export type StopGpuServerError = StopGpuServerErrors[keyof StopGpuServerErrors];
+export type StopGpuServerResponses = {
+    /**
+     * No Content
+     */
+    204: void;
+};
+export type StopGpuServerResponse = StopGpuServerResponses[keyof StopGpuServerResponses];
 export type ListKeysData = {
     body?: never;
     path?: never;
@@ -548,6 +864,35 @@ export type GetMeResponses = {
     200: WebserverMeResponse;
 };
 export type GetMeResponse = GetMeResponses[keyof GetMeResponses];
+export type SearchPrincipalsData = {
+    body?: never;
+    path?: never;
+    query?: {
+        /**
+         * Search text
+         */
+        q?: string;
+        /**
+         * Maximum entries per kind (default 10, at most 50)
+         */
+        limit?: number;
+    };
+    url: '/v1/principals/search';
+};
+export type SearchPrincipalsErrors = {
+    /**
+     * Bad Gateway
+     */
+    502: WebserverErrorResponse;
+};
+export type SearchPrincipalsError = SearchPrincipalsErrors[keyof SearchPrincipalsErrors];
+export type SearchPrincipalsResponses = {
+    /**
+     * OK
+     */
+    200: WebserverPrincipalSearchResponse;
+};
+export type SearchPrincipalsResponse = SearchPrincipalsResponses[keyof SearchPrincipalsResponses];
 export type ListTiersData = {
     body?: never;
     path?: never;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@dhbw-cloud/llm-client",
-  "version": "0.1.0-test.7",
+  "version": "0.1.1",
   "description": "Generated TypeScript client for the DHBW Cloud llm-management-api.",
   "license": "Apache-2.0",
   "type": "module",