@zuplo/runtime 7.9.17 → 7.9.18

package/out/esm/mocks/index.js

@@ -22,5 +22,5 @@
  *  DEALINGS IN THE SOFTWARE.
  *--------------------------------------------------------------------------------------------*/
     
-import{E as p,J as c,M as u}from"../chunk-UJ56MYSD.js";import"../chunk-VTO3F6DK.js";import{a as d}from"../chunk-QGAJTUPE.js";import"../chunk-A3QGJZO7.js";import"../chunk-YJ2BHVUH.js";import{aa as i}from"../chunk-PBDT4FDO.js";import{a as o}from"../chunk-TYV53J3B.js";function w(l={request:new Request("https://api.example.com")}){let e=[];function t(s){e.push(Promise.resolve(s))}return o(t,"waitUntil"),{context:new a({event:{waitUntil:t},route:l.route,originalRequest:l.request}),invokeResponse:o(async()=>{await Promise.all(e)},"invokeResponse")}}o(w,"createMockContext");var g={path:"/",methods:["GET"],handler:{module:{},export:"default"},raw:o(()=>({}),"raw")},a=class extends EventTarget{static{o(this,"MockZuploContext")}#e;#t;contextId;requestId;log;route;custom;incomingRequestProperties;parentContext;analyticsContext;constructor({event:e,route:t=g,parentContext:n,originalRequest:r}){super(),this.requestId=crypto.randomUUID(),this.contextId=crypto.randomUUID(),this.log={info:i.console.info,log:i.console.log,debug:i.console.debug,warn:i.console.warn,error:i.console.error,setLogProperties:o(()=>{},"setLogProperties")},this.custom={},this.route=t,this.incomingRequestProperties={ip:"203.0.113.7",asn:1234,asOrganization:"ORGANIZATION",city:"Seattle",region:"Washington",regionCode:"WA",colo:"SEA",continent:"NA",country:"US",postalCode:"98004",metroCode:"SEA",latitude:void 0,longitude:void 0,timezone:void 0,httpProtocol:void 0,clientCert:void 0,clientMtlsVerificationStatus:void 0,clientMtlsVerificationReason:void 0,clientCertFingerprintSha256:void 0,clientCertNotBefore:void 0,clientCertNotAfter:void 0,clientCertIssuerDn:void 0,clientCertSubjectDn:void 0},this.parentContext=n,this.#e=e,this.#t=r,this.analyticsContext=new c(this.requestId)}waitUntil(e){this.#e.waitUntil(e)}invokeInboundPolicy(e,t){throw new Error("Not implemented")}invokeOutboundPolicy(e,t,n){throw new Error("Not implemented")}invokeRoute(e,t){throw new Error("Not implemented")}addResponseSendingHook(e){throw new Error("Not implemented")}addResponseSendingFinalHook(e){throw new Error("Not implemented")}addPreRequestHandler(e,t){let n;try{n=u.getContextExtensions(this)}catch{n=u.initialize(this,new d(this.#t??new Request("https://api.example.com")))}n.preRequestPhase??="active",p(this,n,e,t)}get preRequestHandlers(){let e;try{e=u.getContextExtensions(this)}catch{return[]}return(e.preRequestState?.entries??[]).map(t=>({policyName:t.policyName,name:t.name}))}addEventListener(e,t,n){let r=o(s=>{try{typeof t=="function"?t(s):t.handleEvent(s)}catch(m){throw this.log.error(`Error invoking event ${e}. See following logs for details.`),m}},"wrapped");super.addEventListener(e,r,n)}};export{a as MockZuploContext,w as createMockContext};
+import{E as p,J as c,M as u}from"../chunk-GYEX3OZM.js";import"../chunk-VTO3F6DK.js";import{a as d}from"../chunk-QGAJTUPE.js";import"../chunk-A3QGJZO7.js";import"../chunk-YJ2BHVUH.js";import{aa as i}from"../chunk-PBDT4FDO.js";import{a as o}from"../chunk-TYV53J3B.js";function w(l={request:new Request("https://api.example.com")}){let e=[];function t(s){e.push(Promise.resolve(s))}return o(t,"waitUntil"),{context:new a({event:{waitUntil:t},route:l.route,originalRequest:l.request}),invokeResponse:o(async()=>{await Promise.all(e)},"invokeResponse")}}o(w,"createMockContext");var g={path:"/",methods:["GET"],handler:{module:{},export:"default"},raw:o(()=>({}),"raw")},a=class extends EventTarget{static{o(this,"MockZuploContext")}#e;#t;contextId;requestId;log;route;custom;incomingRequestProperties;parentContext;analyticsContext;constructor({event:e,route:t=g,parentContext:n,originalRequest:r}){super(),this.requestId=crypto.randomUUID(),this.contextId=crypto.randomUUID(),this.log={info:i.console.info,log:i.console.log,debug:i.console.debug,warn:i.console.warn,error:i.console.error,setLogProperties:o(()=>{},"setLogProperties")},this.custom={},this.route=t,this.incomingRequestProperties={ip:"203.0.113.7",asn:1234,asOrganization:"ORGANIZATION",city:"Seattle",region:"Washington",regionCode:"WA",colo:"SEA",continent:"NA",country:"US",postalCode:"98004",metroCode:"SEA",latitude:void 0,longitude:void 0,timezone:void 0,httpProtocol:void 0,clientCert:void 0,clientMtlsVerificationStatus:void 0,clientMtlsVerificationReason:void 0,clientCertFingerprintSha256:void 0,clientCertNotBefore:void 0,clientCertNotAfter:void 0,clientCertIssuerDn:void 0,clientCertSubjectDn:void 0},this.parentContext=n,this.#e=e,this.#t=r,this.analyticsContext=new c(this.requestId)}waitUntil(e){this.#e.waitUntil(e)}invokeInboundPolicy(e,t){throw new Error("Not implemented")}invokeOutboundPolicy(e,t,n){throw new Error("Not implemented")}invokeRoute(e,t){throw new Error("Not implemented")}addResponseSendingHook(e){throw new Error("Not implemented")}addResponseSendingFinalHook(e){throw new Error("Not implemented")}addPreRequestHandler(e,t){let n;try{n=u.getContextExtensions(this)}catch{n=u.initialize(this,new d(this.#t??new Request("https://api.example.com")))}n.preRequestPhase??="active",p(this,n,e,t)}get preRequestHandlers(){let e;try{e=u.getContextExtensions(this)}catch{return[]}return(e.preRequestState?.entries??[]).map(t=>({policyName:t.policyName,name:t.name}))}addEventListener(e,t,n){let r=o(s=>{try{typeof t=="function"?t(s):t.handleEvent(s)}catch(m){throw this.log.error(`Error invoking event ${e}. See following logs for details.`),m}},"wrapped");super.addEventListener(e,r,n)}};export{a as MockZuploContext,w as createMockContext};
 //# sourceMappingURL=index.js.map

package/out/types/index.d.ts

@@ -540,6 +540,61 @@ export declare const AIGatewayInternalOnlyInboundPolicy: typeof InternalOnlyInbo
 export declare type AIGatewayInternalOnlyInboundPolicyOptions =
   InternalOnlyInboundPolicyOptions;
 
+/**
+ * Spreads AI requests across a list of models. The policy cycles through the
+ * list, or keeps requests that share a key, such as a user or a conversation,
+ * on the same model. Most requests also get a retry model from the list. A
+ * request moves there when the chosen model stays rate limited. Chat
+ * completions and embeddings requests also move there when the model returns a
+ * server error, can't be reached, or times out.
+ *
+ * The policy picks the model for every request it handles and ignores the
+ * request's own `model`, so the list also limits which models clients can
+ * reach. An earlier policy's model selection is never overwritten.
+ *
+ * @title Model Load Balancing
+ * @product ai-gateway
+ * @requiresAI
+ * @param request - The ZuploRequest
+ * @param context - The ZuploContext
+ * @param options - The policy options set in policies.json
+ * @param policyName - The name of the policy as set in policies.json
+ * @returns The request with model routing stored in context, or a refusal
+ * @example
+ * ```json
+ * {
+ *   "strategy": "round-robin",
+ *   "models": {
+ *     "completions": [
+ *       "azure-eastus/gpt-5",
+ *       "azure-westus/gpt-5",
+ *       "openai/gpt-5"
+ *     ]
+ *   }
+ * }
+ * ```
+ */
+export declare class AIGatewayLoadBalancingInboundPolicy extends InboundPolicy<AIGatewayLoadBalancingInboundPolicyOptions> {
+  #private;
+  static readonly policyType = "ai-gateway-load-balancing";
+  handler(
+    request: ZuploRequest,
+    context: ZuploContext
+  ): Promise<ZuploRequest | Response>;
+}
+
+/**
+ * Options for spreading AI Gateway requests across a list of models.
+ * @public
+ */
+export declare interface AIGatewayLoadBalancingInboundPolicyOptions {
+  strategy?: Strategy;
+  stickyBy?: StickyBy;
+  expression?: Expression;
+  models: Models;
+  fallbackTimeoutSeconds?: FallbackTimeoutSeconds_2;
+}
+
 /* Excluded from this release type: AIGatewayMeterIncrements */
 
 /**
@@ -721,7 +776,7 @@ export { AIGatewayModelFilteringInboundPolicy as AIGatewayModelFilteringV2Inboun
  * @public
  */
 declare interface AIGatewayModelFilteringInboundPolicyOptions {
-  models: Models;
+  models: Models_2;
 }
 export { AIGatewayModelFilteringInboundPolicyOptions };
 export { AIGatewayModelFilteringInboundPolicyOptions as AIGatewayModelFilteringV2InboundPolicyOptions };
@@ -799,7 +854,7 @@ export declare class AIGatewayModelOverrideInboundPolicy extends InboundPolicy<A
  * @public
  */
 export declare interface AIGatewayModelOverrideInboundPolicyOptions {
-  models: Models_2;
+  models: Models_3;
 }
 
 /**
@@ -1608,13 +1663,13 @@ export declare interface AkamaiFirewallForAiOutboundPolicyOptions {
  * @minItems 1
  * @public
  */
-declare type AllowedModels = [ProviderAndModel, ...ProviderAndModel[]];
+declare type AllowedModels = [ProviderAndModel_2, ...ProviderAndModel_2[]];
 
 /**
  * @minItems 1
  * @public
  */
-declare type AllowedModels1 = [ProviderAndModel, ...ProviderAndModel[]];
+declare type AllowedModels1 = [ProviderAndModel_2, ...ProviderAndModel_2[]];
 
 /**
  * Amberflo is a usage metering and billing service. This policy allows
@@ -3203,13 +3258,13 @@ declare interface BatchDispatcherOptions {
  * @minItems 1
  * @public
  */
-declare type BlockedModels = [ProviderAndModel, ...ProviderAndModel[]];
+declare type BlockedModels = [ProviderAndModel_2, ...ProviderAndModel_2[]];
 
 /**
  * @minItems 1
  * @public
  */
-declare type BlockedModels1 = [ProviderAndModel, ...ProviderAndModel[]];
+declare type BlockedModels1 = [ProviderAndModel_2, ...ProviderAndModel_2[]];
 
 /**
  * The brownout policy allows performing scheduled downtime on your API
@@ -4179,11 +4234,20 @@ declare interface CompletionsFallbacks {
   quotaFallback?: QuotaFallbackModel;
 }
 
+/**
+ * Models for chat completions, Responses, Messages, and System One requests, as `providerName/model`. From 1 through 32 models, each listed once. Provider names aren't case-sensitive, so `openai/gpt-5` and `OpenAI/gpt-5` are the same model. Write each model ID exactly as your provider spells it.
+ *
+ * @minItems 1
+ * @maxItems 32
+ * @public
+ */
+declare type CompletionsModels = [ProviderAndModel, ...ProviderAndModel[]];
+
 /**
  * Rules for chat completions, Responses, and Anthropic Messages requests.
  * @public
  */
-declare type CompletionsModels =
+declare type CompletionsModels_2 =
   | {
       allowList: AllowedModels;
     }
@@ -5321,11 +5385,20 @@ declare interface EmbeddingsFallbacks {
   quotaFallback?: QuotaFallbackModel;
 }
 
+/**
+ * Models for embeddings requests, as `providerName/model`. From 1 through 32 models, each listed once. Provider names aren't case-sensitive. Write each model ID exactly as your provider spells it. Every entry must be the same embedding model, such as one model on several accounts.
+ *
+ * @minItems 1
+ * @maxItems 32
+ * @public
+ */
+declare type EmbeddingsModels = [ProviderAndModel, ...ProviderAndModel[]];
+
 /**
  * Rules for embedding requests.
  * @public
  */
-declare type EmbeddingsModels =
+declare type EmbeddingsModels_2 =
   | {
       allowList: AllowedModels1;
     }
@@ -5429,6 +5502,12 @@ declare type EventType = (typeof EventType)[keyof typeof EventType];
  */
 declare type Execution = "sequential" | "parallel";
 
+/**
+ * The value to key on, written like a {@link https://zuplo.com/docs/policies/ai-gateway-metering-inbound#budget-expressions|Metering budget expression}, such as `request.headers.get("x-end-user-id")`. Required when `stickyBy` is `expression`, and not allowed otherwise.
+ * @public
+ */
+declare type Expression = string;
+
 /**
  * Best-effort flatten of ALL human/assistant text in a request OR response body,
  * across every shape — the one sanctioned cross-shape helper. Intended for
@@ -5492,6 +5571,12 @@ declare interface FallbackModels {
  */
 declare type FallbackTimeoutSeconds = number;
 
+/**
+ * How long to wait for a model to start responding before retrying the request on another model. Applies to chat completions and embeddings. The timeout ends when the response starts, so it doesn't cut off a long streaming response. A response that doesn't stream starts only when it's complete, so the timeout covers the whole response. From 1 through 300.
+ * @public
+ */
+declare type FallbackTimeoutSeconds_2 = number;
+
 declare interface FetchOptions<T> {
   method?: HttpMethod_2;
   data?: T;
@@ -10681,7 +10766,7 @@ declare interface ModelConfiguration {
 }
 
 /**
- * Model filtering rules grouped by AI Gateway capability.
+ * The models to spread requests across, by request type. Set at least one list.
  * @public
  */
 declare interface Models {
@@ -10690,10 +10775,19 @@ declare interface Models {
 }
 
 /**
- * Override rules grouped by AI Gateway capability.
+ * Model filtering rules grouped by AI Gateway capability.
  * @public
  */
 declare interface Models_2 {
+  completions?: CompletionsModels_2;
+  embeddings?: EmbeddingsModels_2;
+}
+
+/**
+ * Override rules grouped by AI Gateway capability.
+ * @public
+ */
+declare interface Models_3 {
   completions?: CompletionsOverride;
   embeddings?: EmbeddingsOverride;
 }
@@ -12686,6 +12780,12 @@ export declare interface PropelAuthJwtInboundPolicyOptions {
  */
 declare type ProviderAndModel = string;
 
+/**
+ * A model reference in providerName/model form, split on the first slash.
+ * @public
+ */
+declare type ProviderAndModel_2 = string;
+
 /**
  * Extracts a query parameter and sets it as a header in the request.
  *
@@ -14445,6 +14545,18 @@ declare interface StartsWithRule {
   startsWith: string;
 }
 
+/**
+ * The key that `sticky` uses. `user` uses the authenticated caller, `request.user.sub`. `conversation` uses the conversation's system prompt and its first user message with text. `expression` uses the value that `expression` selects. Required when `strategy` is `sticky`, and not allowed with `round-robin`.
+ * @public
+ */
+declare type StickyBy = "user" | "conversation" | "expression";
+
+/**
+ * How the policy picks a model for each request. `round-robin` cycles through the list in order. `sticky` sends every request with the same key to the same model.
+ * @public
+ */
+declare type Strategy = "round-robin" | "sticky";
+
 /**
  * Configuration for accumulating and validating streaming responses.
  * @public

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@zuplo/runtime",
   "type": "module",
-  "version": "7.9.17",
+  "version": "7.9.18",
   "repository": "https://github.com/zuplo/zuplo",
   "author": "Zuplo, Inc.",
   "exports": {