@semiont/inference 0.5.23 → 0.5.25

package/README.md

@@ -90,10 +90,16 @@ interface InferenceClient {
   readonly type: string;     // 'anthropic' | 'ollama' | 'mock'
   readonly modelId: string;  // configured model name
 
+  limits(): Promise<InferenceLimits>;
   generateText(prompt, maxTokens, temperature, options?): Promise<string>;
   generateTextWithMetadata(prompt, maxTokens, temperature, options?): Promise<InferenceResponse>;
 }
 
+interface InferenceLimits {
+  contextTokens: number;     // context window (Anthropic: max input; Ollama: shared input+output)
+  maxOutputTokens: number;   // max output per generation (Ollama mirrors the shared window here)
+}
+
 interface InferenceResponse {
   text: string;
   stopReason: 'end_turn' | 'max_tokens' | 'stop_sequence' | string;
@@ -115,9 +121,22 @@ Each implementation honors the contract with its provider's mechanism:
 
 Current callers all expect arrays (entity extraction, motivation detection). If an object-emitting caller appears, the option grows a `root: 'array' | 'object'` field — see the notes in [src/interface.ts](src/interface.ts).
 
+### Provider limits
+
+`limits()` publishes the provider's **actual** context/output ceilings for the configured model — discovered from the provider itself, never hand-maintained constants:
+
+- **Anthropic**: the Models API (`models.retrieve`) — `max_input_tokens` / `max_tokens`.
+- **Ollama**: `POST /api/show` — the model's context window. Input and output share that window, so it is published as both fields (`maxOutputTokens === contextTokens` signals a shared window).
+
+Discovery is lazy and cached per client; a failed discovery is **not** cached — the next call retries. When ceilings cannot be determined (unknown model, endpoint unreachable), `limits()` **throws**: fail-loud, never a guessed floor.
+
+Two request-time behaviors ride on the limits:
+- **Ollama sets `num_ctx` explicitly** on every generate request — sized to the prompt estimate + output budget, capped at the model window. Without it, Ollama's model-*default* window silently clips large prompts. A request that genuinely cannot fit **throws** instead of being clipped.
+- **Anthropic streams internally** above the SDK's non-streaming output ceiling (≈21K tokens) — same interface, same response shape.
+
 ### `MockInferenceClient`
 
-A scripted test double ([src/implementations/mock.ts](src/implementations/mock.ts)): construct it with a list of canned responses, then inspect `calls` (recorded prompt/maxTokens/temperature/options per invocation). `reset()` and `setResponses()` helpers included.
+A scripted test double ([src/implementations/mock.ts](src/implementations/mock.ts)): construct it with a list of canned responses, then inspect `calls` (recorded prompt/maxTokens/temperature/options per invocation). `reset()` and `setResponses()` helpers included. An optional third constructor argument injects `InferenceLimits` for chunking/budget tests; the default is generous (1M/1M) so ordinary tests never trip window guards.
 
 ```typescript
 import { MockInferenceClient } from '@semiont/inference';

package/dist/index.d.ts

@@ -27,11 +27,36 @@ interface InferenceOptions {
      */
     format?: 'json';
 }
+/**
+ * A provider's actual ceilings for the configured model, discovered from the
+ * provider itself (Anthropic Models API; Ollama `/api/show`) — never
+ * hand-maintained constants. Detection budget arithmetic derives from these.
+ */
+interface InferenceLimits {
+    /**
+     * The context window in tokens. Semantics differ by provider shape:
+     * Anthropic reports maximum *input* tokens (output has its own ceiling);
+     * Ollama reports the *shared* input+output window and mirrors it in
+     * `maxOutputTokens` (there is no separate output ceiling), so
+     * `maxOutputTokens === contextTokens` signals a shared window.
+     */
+    contextTokens: number;
+    /** Maximum output tokens per generation. */
+    maxOutputTokens: number;
+}
 interface InferenceClient {
     /** Provider type identifier (e.g. 'anthropic', 'ollama') */
     readonly type: string;
     /** Model identifier used for generation (e.g. 'claude-opus-4-6', 'llama3') */
     readonly modelId: string;
+    /**
+     * The provider's actual context/output ceilings for `modelId`. Discovered
+     * lazily on first call and cached for the client's lifetime; a failed
+     * discovery is NOT cached — the next call retries. Throws when the ceilings
+     * cannot be determined (unknown model, discovery endpoint unreachable):
+     * fail-loud, never a guessed floor.
+     */
+    limits(): Promise<InferenceLimits>;
     /**
      * Generate text from a prompt (simple interface)
      */
@@ -57,7 +82,11 @@ declare class AnthropicInferenceClient implements InferenceClient {
     readonly modelId: string;
     private client;
     private logger?;
+    private limitsPromise?;
     constructor(apiKey: string, model: string, baseURL?: string, logger?: Logger);
+    limits(): Promise<InferenceLimits>;
+    private discoverLimits;
+    private requestMessage;
     generateText(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<string>;
     generateTextWithMetadata(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<InferenceResponse>;
 }
@@ -67,7 +96,10 @@ declare class OllamaInferenceClient implements InferenceClient {
     readonly modelId: string;
     private baseURL;
     private logger?;
+    private limitsPromise?;
     constructor(model: string, baseURL?: string, logger?: Logger);
+    limits(): Promise<InferenceLimits>;
+    private discoverLimits;
     generateText(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<string>;
     generateTextWithMetadata(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<InferenceResponse>;
 }
@@ -78,13 +110,15 @@ declare class MockInferenceClient implements InferenceClient {
     private responses;
     private responseIndex;
     private stopReasons;
+    private injectedLimits;
     calls: Array<{
         prompt: string;
         maxTokens: number;
         temperature: number;
         options?: InferenceOptions;
     }>;
-    constructor(responses?: string[], stopReasons?: string[]);
+    constructor(responses?: string[], stopReasons?: string[], limits?: InferenceLimits);
+    limits(): Promise<InferenceLimits>;
     generateText(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<string>;
     generateTextWithMetadata(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<InferenceResponse>;
     reset(): void;
@@ -92,4 +126,4 @@ declare class MockInferenceClient implements InferenceClient {
 }
 
 export { AnthropicInferenceClient, MockInferenceClient, OllamaInferenceClient, createInferenceClient };
-export type { InferenceClient, InferenceClientConfig, InferenceClientType, InferenceOptions, InferenceResponse };
+export type { InferenceClient, InferenceClientConfig, InferenceClientType, InferenceLimits, InferenceOptions, InferenceResponse };

package/dist/index.js

@@ -1,6 +1,7 @@
 // src/implementations/anthropic.ts
 import Anthropic from "@anthropic-ai/sdk";
 import { recordInferenceUsage } from "@semiont/observability";
+var NONSTREAMING_MAX_OUTPUT_TOKENS = Math.floor(128e3 / 6);
 var JSON_ARRAY_TOOL = {
   name: "emit_json_array",
   description: 'Return your entire answer by calling this tool. Put the JSON array of results under the "items" property, and emit no prose.',
@@ -19,6 +20,7 @@ var AnthropicInferenceClient = class {
   modelId;
   client;
   logger;
+  limitsPromise;
   constructor(apiKey, model, baseURL, logger) {
     this.client = new Anthropic({
       apiKey,
@@ -27,6 +29,33 @@ var AnthropicInferenceClient = class {
     this.modelId = model;
     this.logger = logger;
   }
+  limits() {
+    if (!this.limitsPromise) {
+      this.limitsPromise = this.discoverLimits().catch((err) => {
+        this.limitsPromise = void 0;
+        throw err;
+      });
+    }
+    return this.limitsPromise;
+  }
+  async discoverLimits() {
+    const info = await this.client.models.retrieve(this.modelId).catch((err) => {
+      throw new Error(
+        `Failed to discover model limits for '${this.modelId}' from the Models API`,
+        { cause: err }
+      );
+    });
+    if (info.max_input_tokens == null || info.max_tokens == null) {
+      throw new Error(`Models API reports no context/output ceilings for '${this.modelId}'`);
+    }
+    return { contextTokens: info.max_input_tokens, maxOutputTokens: info.max_tokens };
+  }
+  requestMessage(params) {
+    if (params.max_tokens > NONSTREAMING_MAX_OUTPUT_TOKENS) {
+      return this.client.messages.stream(params).finalMessage();
+    }
+    return this.client.messages.create(params);
+  }
   async generateText(prompt, maxTokens, temperature, options) {
     const response = await this.generateTextWithMetadata(prompt, maxTokens, temperature, options);
     return response.text;
@@ -40,18 +69,19 @@ var AnthropicInferenceClient = class {
       temperature,
       format: options?.format
     });
+    const params = {
+      model: this.modelId,
+      max_tokens: maxTokens,
+      temperature,
+      messages: [{ role: "user", content: prompt }],
+      // JSON mode → force the structured-output tool. No prefill assistant
+      // turn: the constraint now lives in the tool call, not in free text.
+      ...jsonMode ? { tools: [JSON_ARRAY_TOOL], tool_choice: { type: "tool", name: JSON_ARRAY_TOOL.name } } : {}
+    };
     const start = performance.now();
     let response;
     try {
-      response = await this.client.messages.create({
-        model: this.modelId,
-        max_tokens: maxTokens,
-        temperature,
-        messages: [{ role: "user", content: prompt }],
-        // JSON mode → force the structured-output tool. No prefill assistant
-        // turn: the constraint now lives in the tool call, not in free text.
-        ...jsonMode ? { tools: [JSON_ARRAY_TOOL], tool_choice: { type: "tool", name: JSON_ARRAY_TOOL.name } } : {}
-      });
+      response = await this.requestMessage(params);
     } catch (err) {
       recordInferenceUsage({
         provider: this.type,
@@ -127,17 +157,49 @@ var AnthropicInferenceClient = class {
 };
 
 // src/implementations/ollama.ts
+import { estimateTokens, isNumber, isObject } from "@semiont/core";
 import { recordInferenceUsage as recordInferenceUsage2 } from "@semiont/observability";
+var NUM_CTX_ESTIMATE_SLACK = 0.2;
+var NUM_CTX_TEMPLATE_ALLOWANCE = 64;
 var OllamaInferenceClient = class {
   type = "ollama";
   modelId;
   baseURL;
   logger;
+  limitsPromise;
   constructor(model, baseURL, logger) {
     this.baseURL = (baseURL || "http://localhost:11434").replace(/\/+$/, "");
     this.modelId = model;
     this.logger = logger;
   }
+  limits() {
+    if (!this.limitsPromise) {
+      this.limitsPromise = this.discoverLimits().catch((err) => {
+        this.limitsPromise = void 0;
+        throw err;
+      });
+    }
+    return this.limitsPromise;
+  }
+  async discoverLimits() {
+    const res = await fetch(`${this.baseURL}/api/show`, {
+      method: "POST",
+      headers: { "Content-Type": "application/json" },
+      body: JSON.stringify({ model: this.modelId })
+    });
+    if (!res.ok) {
+      throw new Error(
+        `Failed to discover model limits: /api/show returned ${res.status} for '${this.modelId}'`
+      );
+    }
+    const data = await res.json();
+    const modelInfo = isObject(data) && isObject(data["model_info"]) ? data["model_info"] : void 0;
+    const contextTokens = readContextLength(modelInfo);
+    if (contextTokens === void 0) {
+      throw new Error(`/api/show reports no context length for '${this.modelId}'`);
+    }
+    return { contextTokens, maxOutputTokens: contextTokens };
+  }
   async generateText(prompt, maxTokens, temperature, options) {
     const response = await this.generateTextWithMetadata(prompt, maxTokens, temperature, options);
     return response.text;
@@ -150,6 +212,17 @@ var OllamaInferenceClient = class {
       temperature,
       format: options?.format
     });
+    const limits = await this.limits();
+    const promptTokens = estimateTokens(prompt);
+    if (promptTokens + maxTokens > limits.contextTokens) {
+      throw new Error(
+        `Prompt (~${promptTokens} tokens) + output budget (${maxTokens}) exceed the '${this.modelId}' context window (${limits.contextTokens} tokens)`
+      );
+    }
+    const numCtx = Math.min(
+      limits.contextTokens,
+      promptTokens + maxTokens + Math.ceil(promptTokens * NUM_CTX_ESTIMATE_SLACK) + NUM_CTX_TEMPLATE_ALLOWANCE
+    );
     const url = `${this.baseURL}/api/generate`;
     const start = performance.now();
     const body = {
@@ -159,6 +232,7 @@ var OllamaInferenceClient = class {
       think: false,
       options: {
         num_predict: maxTokens,
+        num_ctx: numCtx,
         temperature
       }
     };
@@ -229,6 +303,20 @@ var OllamaInferenceClient = class {
     };
   }
 };
+function readContextLength(modelInfo) {
+  if (!modelInfo) return void 0;
+  const arch = modelInfo["general.architecture"];
+  if (typeof arch === "string") {
+    const direct = modelInfo[`${arch}.context_length`];
+    if (isNumber(direct) && direct > 0) return direct;
+  }
+  const fallbackKey = Object.keys(modelInfo).find((k) => k.endsWith(".context_length"));
+  if (fallbackKey !== void 0) {
+    const fallback = modelInfo[fallbackKey];
+    if (isNumber(fallback) && fallback > 0) return fallback;
+  }
+  return void 0;
+}
 function mapStopReason(doneReason) {
   switch (doneReason) {
     case "stop":
@@ -267,16 +355,25 @@ function createInferenceClient(config, logger) {
 }
 
 // src/implementations/mock.ts
+var GENEROUS_LIMITS = {
+  contextTokens: 1e6,
+  maxOutputTokens: 1e6
+};
 var MockInferenceClient = class {
   type = "mock";
   modelId = "mock-model";
   responses = [];
   responseIndex = 0;
   stopReasons = [];
+  injectedLimits;
   calls = [];
-  constructor(responses = ["Mock response"], stopReasons) {
+  constructor(responses = ["Mock response"], stopReasons, limits) {
     this.responses = responses;
     this.stopReasons = stopReasons || responses.map(() => "end_turn");
+    this.injectedLimits = limits ?? GENEROUS_LIMITS;
+  }
+  async limits() {
+    return this.injectedLimits;
   }
   async generateText(prompt, maxTokens, temperature, options) {
     const response = await this.generateTextWithMetadata(prompt, maxTokens, temperature, options);

package/dist/index.js.map

@@ -1 +1 @@
-{"version":3,"sources":["../src/implementations/anthropic.ts","../src/implementations/ollama.ts","../src/factory.ts","../src/implementations/mock.ts"],"sourcesContent":["// Anthropic Claude implementation of InferenceClient interface\n\nimport Anthropic from '@anthropic-ai/sdk';\nimport type { Logger } from '@semiont/core';\nimport { recordInferenceUsage } from '@semiont/observability';\nimport { InferenceClient, InferenceOptions, InferenceResponse } from '../interface.js';\n\n// Forced-tool channel for JSON mode. Anthropic has no grammar-constrained\n// sampling like Ollama's `format`; the equivalent hard guarantee is a *tool\n// call*. We offer exactly one tool and force it via `tool_choice`, so the model\n// must answer by filling the tool's input — which the API serializes as\n// properly-escaped JSON. That kills both free-text failure modes at the source:\n// trailing prose after the `]` (variant 1) and an unescaped `\"` inside a string\n// (variant 2), neither of which a prefill could prevent.\n//\n// A tool's input must be an *object*, so the array is carried under `items`\n// and unwrapped on return (see generateTextWithMetadata) — the caller still\n// receives a top-level JSON array in `text`, exactly as on Ollama.\nconst JSON_ARRAY_TOOL: Anthropic.Tool = {\n  name: 'emit_json_array',\n  description:\n    'Return your entire answer by calling this tool. Put the JSON array of results under the \"items\" property, and emit no prose.',\n  input_schema: {\n    type: 'object',\n    properties: {\n      // Element shape is unconstrained here — the prompt carries the per-element\n      // schema; the tool only enforces that the top-level result is an array.\n      items: { type: 'array', items: {} },\n    },\n    required: ['items'],\n  },\n};\n\nexport class AnthropicInferenceClient implements InferenceClient {\n  readonly type = 'anthropic' as const;\n  readonly modelId: string;\n  private client: Anthropic;\n  private logger?: Logger;\n\n  constructor(apiKey: string, model: string, baseURL?: string, logger?: Logger) {\n    this.client = new Anthropic({\n      apiKey,\n      baseURL: baseURL || 'https://api.anthropic.com',\n    });\n    this.modelId = model;\n    this.logger = logger;\n  }\n\n  async generateText(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<string> {\n    const response = await this.generateTextWithMetadata(prompt, maxTokens, temperature, options);\n    return response.text;\n  }\n\n  async generateTextWithMetadata(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<InferenceResponse> {\n    const jsonMode = options?.format === 'json';\n\n    this.logger?.debug('Generating text with inference client', {\n      model: this.modelId,\n      promptLength: prompt.length,\n      maxTokens,\n      temperature,\n      format: options?.format,\n    });\n\n    const start = performance.now();\n    let response: Awaited<ReturnType<typeof this.client.messages.create>>;\n    try {\n      response = await this.client.messages.create({\n        model: this.modelId,\n        max_tokens: maxTokens,\n        temperature,\n        messages: [{ role: 'user', content: prompt }],\n        // JSON mode → force the structured-output tool. No prefill assistant\n        // turn: the constraint now lives in the tool call, not in free text.\n        ...(jsonMode\n          ? { tools: [JSON_ARRAY_TOOL], tool_choice: { type: 'tool' as const, name: JSON_ARRAY_TOOL.name } }\n          : {}),\n      });\n    } catch (err) {\n      recordInferenceUsage({\n        provider: this.type,\n        model: this.modelId,\n        durationMs: performance.now() - start,\n        outcome: 'error',\n      });\n      throw err;\n    }\n\n    this.logger?.debug('Inference response received', {\n      model: this.modelId,\n      contentBlocks: response.content.length,\n      stopReason: response.stop_reason\n    });\n\n    let text: string;\n    if (jsonMode) {\n      // The answer arrives as a tool_use block, not text. Unwrap the `items`\n      // array and re-serialize it so `text` is a complete, parseable top-level\n      // JSON array — the cross-provider contract every consumer reads.\n      const toolUse = response.content.find(c => c.type === 'tool_use');\n      if (!toolUse || toolUse.type !== 'tool_use') {\n        recordInferenceUsage({\n          provider: this.type,\n          model: this.modelId,\n          durationMs: performance.now() - start,\n          outcome: 'error',\n          inputTokens: response.usage?.input_tokens,\n          outputTokens: response.usage?.output_tokens,\n        });\n        this.logger?.error('No tool_use content in inference response', {\n          model: this.modelId,\n          contentTypes: response.content.map(c => c.type)\n        });\n        throw new Error('No tool_use content in inference response');\n      }\n      // `input` is typed `unknown` by the SDK. A truncated (`max_tokens`)\n      // response may carry partial or absent `items` — fall back to the partial\n      // array, or `[]` if absent; the consumer flags truncation via stopReason.\n      const input = toolUse.input as { items?: unknown };\n      const items = Array.isArray(input.items) ? input.items : [];\n      text = JSON.stringify(items);\n    } else {\n      const textContent = response.content.find(c => c.type === 'text');\n      if (!textContent || textContent.type !== 'text') {\n        recordInferenceUsage({\n          provider: this.type,\n          model: this.modelId,\n          durationMs: performance.now() - start,\n          outcome: 'error',\n          inputTokens: response.usage?.input_tokens,\n          outputTokens: response.usage?.output_tokens,\n        });\n        this.logger?.error('No text content in inference response', {\n          model: this.modelId,\n          contentTypes: response.content.map(c => c.type)\n        });\n        throw new Error('No text content in inference response');\n      }\n      text = textContent.text;\n    }\n\n    recordInferenceUsage({\n      provider: this.type,\n      model: this.modelId,\n      durationMs: performance.now() - start,\n      outcome: 'success',\n      inputTokens: response.usage?.input_tokens,\n      outputTokens: response.usage?.output_tokens,\n    });\n\n    this.logger?.info('Text generation completed', {\n      model: this.modelId,\n      textLength: text.length,\n      stopReason: response.stop_reason\n    });\n\n    return {\n      text,\n      stopReason: response.stop_reason || 'unknown'\n    };\n  }\n}\n","// Ollama implementation of InferenceClient interface\n// Uses native Ollama HTTP API (no SDK dependency)\n\nimport type { Logger } from '@semiont/core';\nimport { recordInferenceUsage } from '@semiont/observability';\nimport { InferenceClient, InferenceOptions, InferenceResponse } from '../interface.js';\n\ninterface OllamaGenerateResponse {\n  response: string;\n  done: boolean;\n  done_reason?: string;\n  /** Number of prompt tokens evaluated. Available on most Ollama versions. */\n  prompt_eval_count?: number;\n  /** Number of tokens generated. */\n  eval_count?: number;\n}\n\nexport class OllamaInferenceClient implements InferenceClient {\n  readonly type = 'ollama' as const;\n  readonly modelId: string;\n  private baseURL: string;\n  private logger?: Logger;\n\n  constructor(model: string, baseURL?: string, logger?: Logger) {\n    this.baseURL = (baseURL || 'http://localhost:11434').replace(/\\/+$/, '');\n    this.modelId = model;\n    this.logger = logger;\n  }\n\n  async generateText(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<string> {\n    const response = await this.generateTextWithMetadata(prompt, maxTokens, temperature, options);\n    return response.text;\n  }\n\n  async generateTextWithMetadata(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<InferenceResponse> {\n    this.logger?.debug('Generating text with Ollama', {\n      model: this.modelId,\n      promptLength: prompt.length,\n      maxTokens,\n      temperature,\n      format: options?.format,\n    });\n\n    const url = `${this.baseURL}/api/generate`;\n    const start = performance.now();\n\n    // Ollama's `format` parameter accepts either the literal string\n    // `\"json\"` (any valid JSON, including objects, numbers, etc.) or a\n    // JSON schema (constrains the top-level shape). The contract on the\n    // inference side is \"parseable JSON array,\" so we pass a minimal\n    // array schema rather than the bare `\"json\"` string — without it,\n    // the model can satisfy \"valid JSON\" by emitting `{\"entities\": [...]}`\n    // and break every consumer that expects to call `.map` on the\n    // top-level value. The schema's `items: {}` keeps element shape\n    // unconstrained — the prompt still carries the per-element schema;\n    // we only enforce the outer array.\n    const body: Record<string, unknown> = {\n      model: this.modelId,\n      prompt,\n      stream: false,\n      think: false,\n      options: {\n        num_predict: maxTokens,\n        temperature,\n      },\n    };\n    if (options?.format === 'json') {\n      body['format'] = { type: 'array', items: {} };\n    }\n\n    let res: Response;\n    try {\n      res = await fetch(url, {\n        method: 'POST',\n        headers: { 'Content-Type': 'application/json' },\n        body: JSON.stringify(body),\n      });\n    } catch (err) {\n      recordInferenceUsage({\n        provider: this.type,\n        model: this.modelId,\n        durationMs: performance.now() - start,\n        outcome: 'error',\n      });\n      throw err;\n    }\n\n    if (!res.ok) {\n      recordInferenceUsage({\n        provider: this.type,\n        model: this.modelId,\n        durationMs: performance.now() - start,\n        outcome: 'error',\n      });\n      const body = await res.text();\n      this.logger?.error('Ollama API error', {\n        model: this.modelId,\n        status: res.status,\n        body,\n      });\n      throw new Error(`Ollama API error (${res.status}): ${body}`);\n    }\n\n    const data = await res.json() as OllamaGenerateResponse;\n\n    if (!data.response) {\n      recordInferenceUsage({\n        provider: this.type,\n        model: this.modelId,\n        durationMs: performance.now() - start,\n        outcome: 'error',\n        inputTokens: data.prompt_eval_count,\n        outputTokens: data.eval_count,\n      });\n      this.logger?.error('Empty response from Ollama', { model: this.modelId });\n      throw new Error('Empty response from Ollama');\n    }\n\n    recordInferenceUsage({\n      provider: this.type,\n      model: this.modelId,\n      durationMs: performance.now() - start,\n      outcome: 'success',\n      inputTokens: data.prompt_eval_count,\n      outputTokens: data.eval_count,\n    });\n\n    const stopReason = mapStopReason(data.done_reason);\n\n    this.logger?.info('Text generation completed', {\n      model: this.modelId,\n      textLength: data.response.length,\n      stopReason,\n    });\n\n    return {\n      text: data.response,\n      stopReason,\n    };\n  }\n}\n\nfunction mapStopReason(doneReason: string | undefined): string {\n  switch (doneReason) {\n    case 'stop': return 'end_turn';\n    case 'length': return 'max_tokens';\n    default: return doneReason || 'unknown';\n  }\n}\n","// Factory for creating inference client instances based on configuration\n\nimport type { Logger } from '@semiont/core';\nimport { InferenceClient } from './interface.js';\nimport { AnthropicInferenceClient } from './implementations/anthropic.js';\nimport { OllamaInferenceClient } from './implementations/ollama.js';\n\nexport type InferenceClientType = 'anthropic' | 'ollama';\n\nexport interface InferenceClientConfig {\n  type: InferenceClientType;\n  apiKey?: string;\n  model: string;\n  endpoint?: string;\n  baseURL?: string;\n}\n\nexport function createInferenceClient(config: InferenceClientConfig, logger?: Logger): InferenceClient {\n  switch (config.type) {\n    case 'anthropic': {\n      if (!config.apiKey || config.apiKey.trim() === '') {\n        throw new Error('apiKey is required for Anthropic inference client');\n      }\n      return new AnthropicInferenceClient(\n        config.apiKey,\n        config.model,\n        config.endpoint || config.baseURL,\n        logger\n      );\n    }\n\n    case 'ollama': {\n      return new OllamaInferenceClient(\n        config.model,\n        config.endpoint || config.baseURL,\n        logger\n      );\n    }\n\n    default:\n      throw new Error(`Unsupported inference client type: ${config.type}`);\n  }\n}\n","// Mock implementation of InferenceClient for testing\n\nimport { InferenceClient, InferenceOptions, InferenceResponse } from '../interface.js';\n\nexport class MockInferenceClient implements InferenceClient {\n  readonly type = 'mock' as const;\n  readonly modelId = 'mock-model' as const;\n  private responses: string[] = [];\n  private responseIndex: number = 0;\n  private stopReasons: string[] = [];\n  public calls: Array<{ prompt: string; maxTokens: number; temperature: number; options?: InferenceOptions }> = [];\n\n  constructor(responses: string[] = ['Mock response'], stopReasons?: string[]) {\n    this.responses = responses;\n    this.stopReasons = stopReasons || responses.map(() => 'end_turn');\n  }\n\n  async generateText(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<string> {\n    const response = await this.generateTextWithMetadata(prompt, maxTokens, temperature, options);\n    return response.text;\n  }\n\n  async generateTextWithMetadata(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<InferenceResponse> {\n    this.calls.push({ prompt, maxTokens, temperature, ...(options ? { options } : {}) });\n\n    const text = this.responses[this.responseIndex];\n    const stopReason = this.stopReasons[this.responseIndex] || 'end_turn';\n\n    if (this.responseIndex < this.responses.length - 1) {\n      this.responseIndex++;\n    }\n\n    return { text, stopReason };\n  }\n\n  // Test helper methods\n  reset(): void {\n    this.calls = [];\n    this.responseIndex = 0;\n  }\n\n  setResponses(responses: string[], stopReasons?: string[]): void {\n    this.responses = responses;\n    this.stopReasons = stopReasons || responses.map(() => 'end_turn');\n    this.responseIndex = 0;\n  }\n}\n"],"mappings":";AAEA,OAAO,eAAe;AAEtB,SAAS,4BAA4B;AAcrC,IAAM,kBAAkC;AAAA,EACtC,MAAM;AAAA,EACN,aACE;AAAA,EACF,cAAc;AAAA,IACZ,MAAM;AAAA,IACN,YAAY;AAAA;AAAA;AAAA,MAGV,OAAO,EAAE,MAAM,SAAS,OAAO,CAAC,EAAE;AAAA,IACpC;AAAA,IACA,UAAU,CAAC,OAAO;AAAA,EACpB;AACF;AAEO,IAAM,2BAAN,MAA0D;AAAA,EACtD,OAAO;AAAA,EACP;AAAA,EACD;AAAA,EACA;AAAA,EAER,YAAY,QAAgB,OAAe,SAAkB,QAAiB;AAC5E,SAAK,SAAS,IAAI,UAAU;AAAA,MAC1B;AAAA,MACA,SAAS,WAAW;AAAA,IACtB,CAAC;AACD,SAAK,UAAU;AACf,SAAK,SAAS;AAAA,EAChB;AAAA,EAEA,MAAM,aAAa,QAAgB,WAAmB,aAAqB,SAA6C;AACtH,UAAM,WAAW,MAAM,KAAK,yBAAyB,QAAQ,WAAW,aAAa,OAAO;AAC5F,WAAO,SAAS;AAAA,EAClB;AAAA,EAEA,MAAM,yBAAyB,QAAgB,WAAmB,aAAqB,SAAwD;AAC7I,UAAM,WAAW,SAAS,WAAW;AAErC,SAAK,QAAQ,MAAM,yCAAyC;AAAA,MAC1D,OAAO,KAAK;AAAA,MACZ,cAAc,OAAO;AAAA,MACrB;AAAA,MACA;AAAA,MACA,QAAQ,SAAS;AAAA,IACnB,CAAC;AAED,UAAM,QAAQ,YAAY,IAAI;AAC9B,QAAI;AACJ,QAAI;AACF,iBAAW,MAAM,KAAK,OAAO,SAAS,OAAO;AAAA,QAC3C,OAAO,KAAK;AAAA,QACZ,YAAY;AAAA,QACZ;AAAA,QACA,UAAU,CAAC,EAAE,MAAM,QAAQ,SAAS,OAAO,CAAC;AAAA;AAAA;AAAA,QAG5C,GAAI,WACA,EAAE,OAAO,CAAC,eAAe,GAAG,aAAa,EAAE,MAAM,QAAiB,MAAM,gBAAgB,KAAK,EAAE,IAC/F,CAAC;AAAA,MACP,CAAC;AAAA,IACH,SAAS,KAAK;AACZ,2BAAqB;AAAA,QACnB,UAAU,KAAK;AAAA,QACf,OAAO,KAAK;AAAA,QACZ,YAAY,YAAY,IAAI,IAAI;AAAA,QAChC,SAAS;AAAA,MACX,CAAC;AACD,YAAM;AAAA,IACR;AAEA,SAAK,QAAQ,MAAM,+BAA+B;AAAA,MAChD,OAAO,KAAK;AAAA,MACZ,eAAe,SAAS,QAAQ;AAAA,MAChC,YAAY,SAAS;AAAA,IACvB,CAAC;AAED,QAAI;AACJ,QAAI,UAAU;AAIZ,YAAM,UAAU,SAAS,QAAQ,KAAK,OAAK,EAAE,SAAS,UAAU;AAChE,UAAI,CAAC,WAAW,QAAQ,SAAS,YAAY;AAC3C,6BAAqB;AAAA,UACnB,UAAU,KAAK;AAAA,UACf,OAAO,KAAK;AAAA,UACZ,YAAY,YAAY,IAAI,IAAI;AAAA,UAChC,SAAS;AAAA,UACT,aAAa,SAAS,OAAO;AAAA,UAC7B,cAAc,SAAS,OAAO;AAAA,QAChC,CAAC;AACD,aAAK,QAAQ,MAAM,6CAA6C;AAAA,UAC9D,OAAO,KAAK;AAAA,UACZ,cAAc,SAAS,QAAQ,IAAI,OAAK,EAAE,IAAI;AAAA,QAChD,CAAC;AACD,cAAM,IAAI,MAAM,2CAA2C;AAAA,MAC7D;AAIA,YAAM,QAAQ,QAAQ;AACtB,YAAM,QAAQ,MAAM,QAAQ,MAAM,KAAK,IAAI,MAAM,QAAQ,CAAC;AAC1D,aAAO,KAAK,UAAU,KAAK;AAAA,IAC7B,OAAO;AACL,YAAM,cAAc,SAAS,QAAQ,KAAK,OAAK,EAAE,SAAS,MAAM;AAChE,UAAI,CAAC,eAAe,YAAY,SAAS,QAAQ;AAC/C,6BAAqB;AAAA,UACnB,UAAU,KAAK;AAAA,UACf,OAAO,KAAK;AAAA,UACZ,YAAY,YAAY,IAAI,IAAI;AAAA,UAChC,SAAS;AAAA,UACT,aAAa,SAAS,OAAO;AAAA,UAC7B,cAAc,SAAS,OAAO;AAAA,QAChC,CAAC;AACD,aAAK,QAAQ,MAAM,yCAAyC;AAAA,UAC1D,OAAO,KAAK;AAAA,UACZ,cAAc,SAAS,QAAQ,IAAI,OAAK,EAAE,IAAI;AAAA,QAChD,CAAC;AACD,cAAM,IAAI,MAAM,uCAAuC;AAAA,MACzD;AACA,aAAO,YAAY;AAAA,IACrB;AAEA,yBAAqB;AAAA,MACnB,UAAU,KAAK;AAAA,MACf,OAAO,KAAK;AAAA,MACZ,YAAY,YAAY,IAAI,IAAI;AAAA,MAChC,SAAS;AAAA,MACT,aAAa,SAAS,OAAO;AAAA,MAC7B,cAAc,SAAS,OAAO;AAAA,IAChC,CAAC;AAED,SAAK,QAAQ,KAAK,6BAA6B;AAAA,MAC7C,OAAO,KAAK;AAAA,MACZ,YAAY,KAAK;AAAA,MACjB,YAAY,SAAS;AAAA,IACvB,CAAC;AAED,WAAO;AAAA,MACL;AAAA,MACA,YAAY,SAAS,eAAe;AAAA,IACtC;AAAA,EACF;AACF;;;AC7JA,SAAS,wBAAAA,6BAA4B;AAa9B,IAAM,wBAAN,MAAuD;AAAA,EACnD,OAAO;AAAA,EACP;AAAA,EACD;AAAA,EACA;AAAA,EAER,YAAY,OAAe,SAAkB,QAAiB;AAC5D,SAAK,WAAW,WAAW,0BAA0B,QAAQ,QAAQ,EAAE;AACvE,SAAK,UAAU;AACf,SAAK,SAAS;AAAA,EAChB;AAAA,EAEA,MAAM,aAAa,QAAgB,WAAmB,aAAqB,SAA6C;AACtH,UAAM,WAAW,MAAM,KAAK,yBAAyB,QAAQ,WAAW,aAAa,OAAO;AAC5F,WAAO,SAAS;AAAA,EAClB;AAAA,EAEA,MAAM,yBAAyB,QAAgB,WAAmB,aAAqB,SAAwD;AAC7I,SAAK,QAAQ,MAAM,+BAA+B;AAAA,MAChD,OAAO,KAAK;AAAA,MACZ,cAAc,OAAO;AAAA,MACrB;AAAA,MACA;AAAA,MACA,QAAQ,SAAS;AAAA,IACnB,CAAC;AAED,UAAM,MAAM,GAAG,KAAK,OAAO;AAC3B,UAAM,QAAQ,YAAY,IAAI;AAY9B,UAAM,OAAgC;AAAA,MACpC,OAAO,KAAK;AAAA,MACZ;AAAA,MACA,QAAQ;AAAA,MACR,OAAO;AAAA,MACP,SAAS;AAAA,QACP,aAAa;AAAA,QACb;AAAA,MACF;AAAA,IACF;AACA,QAAI,SAAS,WAAW,QAAQ;AAC9B,WAAK,QAAQ,IAAI,EAAE,MAAM,SAAS,OAAO,CAAC,EAAE;AAAA,IAC9C;AAEA,QAAI;AACJ,QAAI;AACF,YAAM,MAAM,MAAM,KAAK;AAAA,QACrB,QAAQ;AAAA,QACR,SAAS,EAAE,gBAAgB,mBAAmB;AAAA,QAC9C,MAAM,KAAK,UAAU,IAAI;AAAA,MAC3B,CAAC;AAAA,IACH,SAAS,KAAK;AACZ,MAAAA,sBAAqB;AAAA,QACnB,UAAU,KAAK;AAAA,QACf,OAAO,KAAK;AAAA,QACZ,YAAY,YAAY,IAAI,IAAI;AAAA,QAChC,SAAS;AAAA,MACX,CAAC;AACD,YAAM;AAAA,IACR;AAEA,QAAI,CAAC,IAAI,IAAI;AACX,MAAAA,sBAAqB;AAAA,QACnB,UAAU,KAAK;AAAA,QACf,OAAO,KAAK;AAAA,QACZ,YAAY,YAAY,IAAI,IAAI;AAAA,QAChC,SAAS;AAAA,MACX,CAAC;AACD,YAAMC,QAAO,MAAM,IAAI,KAAK;AAC5B,WAAK,QAAQ,MAAM,oBAAoB;AAAA,QACrC,OAAO,KAAK;AAAA,QACZ,QAAQ,IAAI;AAAA,QACZ,MAAAA;AAAA,MACF,CAAC;AACD,YAAM,IAAI,MAAM,qBAAqB,IAAI,MAAM,MAAMA,KAAI,EAAE;AAAA,IAC7D;AAEA,UAAM,OAAO,MAAM,IAAI,KAAK;AAE5B,QAAI,CAAC,KAAK,UAAU;AAClB,MAAAD,sBAAqB;AAAA,QACnB,UAAU,KAAK;AAAA,QACf,OAAO,KAAK;AAAA,QACZ,YAAY,YAAY,IAAI,IAAI;AAAA,QAChC,SAAS;AAAA,QACT,aAAa,KAAK;AAAA,QAClB,cAAc,KAAK;AAAA,MACrB,CAAC;AACD,WAAK,QAAQ,MAAM,8BAA8B,EAAE,OAAO,KAAK,QAAQ,CAAC;AACxE,YAAM,IAAI,MAAM,4BAA4B;AAAA,IAC9C;AAEA,IAAAA,sBAAqB;AAAA,MACnB,UAAU,KAAK;AAAA,MACf,OAAO,KAAK;AAAA,MACZ,YAAY,YAAY,IAAI,IAAI;AAAA,MAChC,SAAS;AAAA,MACT,aAAa,KAAK;AAAA,MAClB,cAAc,KAAK;AAAA,IACrB,CAAC;AAED,UAAM,aAAa,cAAc,KAAK,WAAW;AAEjD,SAAK,QAAQ,KAAK,6BAA6B;AAAA,MAC7C,OAAO,KAAK;AAAA,MACZ,YAAY,KAAK,SAAS;AAAA,MAC1B;AAAA,IACF,CAAC;AAED,WAAO;AAAA,MACL,MAAM,KAAK;AAAA,MACX;AAAA,IACF;AAAA,EACF;AACF;AAEA,SAAS,cAAc,YAAwC;AAC7D,UAAQ,YAAY;AAAA,IAClB,KAAK;AAAQ,aAAO;AAAA,IACpB,KAAK;AAAU,aAAO;AAAA,IACtB;AAAS,aAAO,cAAc;AAAA,EAChC;AACF;;;ACnIO,SAAS,sBAAsB,QAA+B,QAAkC;AACrG,UAAQ,OAAO,MAAM;AAAA,IACnB,KAAK,aAAa;AAChB,UAAI,CAAC,OAAO,UAAU,OAAO,OAAO,KAAK,MAAM,IAAI;AACjD,cAAM,IAAI,MAAM,mDAAmD;AAAA,MACrE;AACA,aAAO,IAAI;AAAA,QACT,OAAO;AAAA,QACP,OAAO;AAAA,QACP,OAAO,YAAY,OAAO;AAAA,QAC1B;AAAA,MACF;AAAA,IACF;AAAA,IAEA,KAAK,UAAU;AACb,aAAO,IAAI;AAAA,QACT,OAAO;AAAA,QACP,OAAO,YAAY,OAAO;AAAA,QAC1B;AAAA,MACF;AAAA,IACF;AAAA,IAEA;AACE,YAAM,IAAI,MAAM,sCAAsC,OAAO,IAAI,EAAE;AAAA,EACvE;AACF;;;ACtCO,IAAM,sBAAN,MAAqD;AAAA,EACjD,OAAO;AAAA,EACP,UAAU;AAAA,EACX,YAAsB,CAAC;AAAA,EACvB,gBAAwB;AAAA,EACxB,cAAwB,CAAC;AAAA,EAC1B,QAAuG,CAAC;AAAA,EAE/G,YAAY,YAAsB,CAAC,eAAe,GAAG,aAAwB;AAC3E,SAAK,YAAY;AACjB,SAAK,cAAc,eAAe,UAAU,IAAI,MAAM,UAAU;AAAA,EAClE;AAAA,EAEA,MAAM,aAAa,QAAgB,WAAmB,aAAqB,SAA6C;AACtH,UAAM,WAAW,MAAM,KAAK,yBAAyB,QAAQ,WAAW,aAAa,OAAO;AAC5F,WAAO,SAAS;AAAA,EAClB;AAAA,EAEA,MAAM,yBAAyB,QAAgB,WAAmB,aAAqB,SAAwD;AAC7I,SAAK,MAAM,KAAK,EAAE,QAAQ,WAAW,aAAa,GAAI,UAAU,EAAE,QAAQ,IAAI,CAAC,EAAG,CAAC;AAEnF,UAAM,OAAO,KAAK,UAAU,KAAK,aAAa;AAC9C,UAAM,aAAa,KAAK,YAAY,KAAK,aAAa,KAAK;AAE3D,QAAI,KAAK,gBAAgB,KAAK,UAAU,SAAS,GAAG;AAClD,WAAK;AAAA,IACP;AAEA,WAAO,EAAE,MAAM,WAAW;AAAA,EAC5B;AAAA;AAAA,EAGA,QAAc;AACZ,SAAK,QAAQ,CAAC;AACd,SAAK,gBAAgB;AAAA,EACvB;AAAA,EAEA,aAAa,WAAqB,aAA8B;AAC9D,SAAK,YAAY;AACjB,SAAK,cAAc,eAAe,UAAU,IAAI,MAAM,UAAU;AAChE,SAAK,gBAAgB;AAAA,EACvB;AACF;","names":["recordInferenceUsage","body"]}
+{"version":3,"sources":["../src/implementations/anthropic.ts","../src/implementations/ollama.ts","../src/factory.ts","../src/implementations/mock.ts"],"sourcesContent":["// Anthropic Claude implementation of InferenceClient interface\n\nimport Anthropic from '@anthropic-ai/sdk';\nimport type { Logger } from '@semiont/core';\nimport { recordInferenceUsage } from '@semiont/observability';\nimport { InferenceClient, InferenceLimits, InferenceOptions, InferenceResponse } from '../interface.js';\n\n// The SDK refuses non-streaming create() calls whose projected duration\n// exceeds its 10-minute timeout: it throws when\n// (60min × max_tokens) / 128_000 > 10min, i.e. above 128_000/6 ≈ 21,333\n// output tokens (client.js, calculateNonstreamingTimeout). Above that we\n// stream internally and assemble the final message — same request shape,\n// same response handling, same interface.\nconst NONSTREAMING_MAX_OUTPUT_TOKENS = Math.floor(128_000 / 6);\n\n// Forced-tool channel for JSON mode. Anthropic has no grammar-constrained\n// sampling like Ollama's `format`; the equivalent hard guarantee is a *tool\n// call*. We offer exactly one tool and force it via `tool_choice`, so the model\n// must answer by filling the tool's input — which the API serializes as\n// properly-escaped JSON. That kills both free-text failure modes at the source:\n// trailing prose after the `]` (variant 1) and an unescaped `\"` inside a string\n// (variant 2), neither of which a prefill could prevent.\n//\n// A tool's input must be an *object*, so the array is carried under `items`\n// and unwrapped on return (see generateTextWithMetadata) — the caller still\n// receives a top-level JSON array in `text`, exactly as on Ollama.\nconst JSON_ARRAY_TOOL: Anthropic.Tool = {\n  name: 'emit_json_array',\n  description:\n    'Return your entire answer by calling this tool. Put the JSON array of results under the \"items\" property, and emit no prose.',\n  input_schema: {\n    type: 'object',\n    properties: {\n      // Element shape is unconstrained here — the prompt carries the per-element\n      // schema; the tool only enforces that the top-level result is an array.\n      items: { type: 'array', items: {} },\n    },\n    required: ['items'],\n  },\n};\n\nexport class AnthropicInferenceClient implements InferenceClient {\n  readonly type = 'anthropic' as const;\n  readonly modelId: string;\n  private client: Anthropic;\n  private logger?: Logger;\n  private limitsPromise?: Promise<InferenceLimits>;\n\n  constructor(apiKey: string, model: string, baseURL?: string, logger?: Logger) {\n    this.client = new Anthropic({\n      apiKey,\n      baseURL: baseURL || 'https://api.anthropic.com',\n    });\n    this.modelId = model;\n    this.logger = logger;\n  }\n\n  limits(): Promise<InferenceLimits> {\n    if (!this.limitsPromise) {\n      this.limitsPromise = this.discoverLimits().catch((err: unknown) => {\n        // Never cache a failed discovery — a transient outage would otherwise\n        // pin every future call to the same rejection.\n        this.limitsPromise = undefined;\n        throw err;\n      });\n    }\n    return this.limitsPromise;\n  }\n\n  private async discoverLimits(): Promise<InferenceLimits> {\n    // The Models API publishes the actual ceilings per model — no\n    // hand-maintained table to go stale when a new model ships.\n    const info = await this.client.models.retrieve(this.modelId).catch((err: unknown) => {\n      throw new Error(\n        `Failed to discover model limits for '${this.modelId}' from the Models API`,\n        { cause: err },\n      );\n    });\n    if (info.max_input_tokens == null || info.max_tokens == null) {\n      throw new Error(`Models API reports no context/output ceilings for '${this.modelId}'`);\n    }\n    return { contextTokens: info.max_input_tokens, maxOutputTokens: info.max_tokens };\n  }\n\n  private requestMessage(params: Anthropic.MessageCreateParamsNonStreaming): Promise<Anthropic.Message> {\n    if (params.max_tokens > NONSTREAMING_MAX_OUTPUT_TOKENS) {\n      return this.client.messages.stream(params).finalMessage();\n    }\n    return this.client.messages.create(params);\n  }\n\n  async generateText(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<string> {\n    const response = await this.generateTextWithMetadata(prompt, maxTokens, temperature, options);\n    return response.text;\n  }\n\n  async generateTextWithMetadata(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<InferenceResponse> {\n    const jsonMode = options?.format === 'json';\n\n    this.logger?.debug('Generating text with inference client', {\n      model: this.modelId,\n      promptLength: prompt.length,\n      maxTokens,\n      temperature,\n      format: options?.format,\n    });\n\n    const params: Anthropic.MessageCreateParamsNonStreaming = {\n      model: this.modelId,\n      max_tokens: maxTokens,\n      temperature,\n      messages: [{ role: 'user', content: prompt }],\n      // JSON mode → force the structured-output tool. No prefill assistant\n      // turn: the constraint now lives in the tool call, not in free text.\n      ...(jsonMode\n        ? { tools: [JSON_ARRAY_TOOL], tool_choice: { type: 'tool' as const, name: JSON_ARRAY_TOOL.name } }\n        : {}),\n    };\n\n    const start = performance.now();\n    let response: Anthropic.Message;\n    try {\n      response = await this.requestMessage(params);\n    } catch (err) {\n      recordInferenceUsage({\n        provider: this.type,\n        model: this.modelId,\n        durationMs: performance.now() - start,\n        outcome: 'error',\n      });\n      throw err;\n    }\n\n    this.logger?.debug('Inference response received', {\n      model: this.modelId,\n      contentBlocks: response.content.length,\n      stopReason: response.stop_reason\n    });\n\n    let text: string;\n    if (jsonMode) {\n      // The answer arrives as a tool_use block, not text. Unwrap the `items`\n      // array and re-serialize it so `text` is a complete, parseable top-level\n      // JSON array — the cross-provider contract every consumer reads.\n      const toolUse = response.content.find(c => c.type === 'tool_use');\n      if (!toolUse || toolUse.type !== 'tool_use') {\n        recordInferenceUsage({\n          provider: this.type,\n          model: this.modelId,\n          durationMs: performance.now() - start,\n          outcome: 'error',\n          inputTokens: response.usage?.input_tokens,\n          outputTokens: response.usage?.output_tokens,\n        });\n        this.logger?.error('No tool_use content in inference response', {\n          model: this.modelId,\n          contentTypes: response.content.map(c => c.type)\n        });\n        throw new Error('No tool_use content in inference response');\n      }\n      // `input` is typed `unknown` by the SDK. A truncated (`max_tokens`)\n      // response may carry partial or absent `items` — fall back to the partial\n      // array, or `[]` if absent; the consumer flags truncation via stopReason.\n      const input = toolUse.input as { items?: unknown };\n      const items = Array.isArray(input.items) ? input.items : [];\n      text = JSON.stringify(items);\n    } else {\n      const textContent = response.content.find(c => c.type === 'text');\n      if (!textContent || textContent.type !== 'text') {\n        recordInferenceUsage({\n          provider: this.type,\n          model: this.modelId,\n          durationMs: performance.now() - start,\n          outcome: 'error',\n          inputTokens: response.usage?.input_tokens,\n          outputTokens: response.usage?.output_tokens,\n        });\n        this.logger?.error('No text content in inference response', {\n          model: this.modelId,\n          contentTypes: response.content.map(c => c.type)\n        });\n        throw new Error('No text content in inference response');\n      }\n      text = textContent.text;\n    }\n\n    recordInferenceUsage({\n      provider: this.type,\n      model: this.modelId,\n      durationMs: performance.now() - start,\n      outcome: 'success',\n      inputTokens: response.usage?.input_tokens,\n      outputTokens: response.usage?.output_tokens,\n    });\n\n    this.logger?.info('Text generation completed', {\n      model: this.modelId,\n      textLength: text.length,\n      stopReason: response.stop_reason\n    });\n\n    return {\n      text,\n      stopReason: response.stop_reason || 'unknown'\n    };\n  }\n}\n","// Ollama implementation of InferenceClient interface\n// Uses native Ollama HTTP API (no SDK dependency)\n\nimport { estimateTokens, isNumber, isObject } from '@semiont/core';\nimport type { Logger } from '@semiont/core';\nimport { recordInferenceUsage } from '@semiont/observability';\nimport { InferenceClient, InferenceLimits, InferenceOptions, InferenceResponse } from '../interface.js';\n\n// Slack added to the chars/4 prompt estimate when sizing `num_ctx`:\n// proportional to the estimate (the heuristic's error grows with prompt size)\n// plus a small fixed allowance for the model's chat template. The risk profile\n// is asymmetric — an undersized window silently clips input (the exact hole\n// managed num_ctx exists to close) while an oversized one only costs memory —\n// so the slack leans generous. Always capped at the model's real window.\nconst NUM_CTX_ESTIMATE_SLACK = 0.2;\nconst NUM_CTX_TEMPLATE_ALLOWANCE = 64;\n\ninterface OllamaGenerateResponse {\n  response: string;\n  done: boolean;\n  done_reason?: string;\n  /** Number of prompt tokens evaluated. Available on most Ollama versions. */\n  prompt_eval_count?: number;\n  /** Number of tokens generated. */\n  eval_count?: number;\n}\n\nexport class OllamaInferenceClient implements InferenceClient {\n  readonly type = 'ollama' as const;\n  readonly modelId: string;\n  private baseURL: string;\n  private logger?: Logger;\n\n  private limitsPromise?: Promise<InferenceLimits>;\n\n  constructor(model: string, baseURL?: string, logger?: Logger) {\n    this.baseURL = (baseURL || 'http://localhost:11434').replace(/\\/+$/, '');\n    this.modelId = model;\n    this.logger = logger;\n  }\n\n  limits(): Promise<InferenceLimits> {\n    if (!this.limitsPromise) {\n      this.limitsPromise = this.discoverLimits().catch((err: unknown) => {\n        // Never cache a failed discovery — a transient outage would otherwise\n        // pin every future call to the same rejection.\n        this.limitsPromise = undefined;\n        throw err;\n      });\n    }\n    return this.limitsPromise;\n  }\n\n  private async discoverLimits(): Promise<InferenceLimits> {\n    const res = await fetch(`${this.baseURL}/api/show`, {\n      method: 'POST',\n      headers: { 'Content-Type': 'application/json' },\n      body: JSON.stringify({ model: this.modelId }),\n    });\n    if (!res.ok) {\n      throw new Error(\n        `Failed to discover model limits: /api/show returned ${res.status} for '${this.modelId}'`,\n      );\n    }\n    const data: unknown = await res.json();\n    const modelInfo = isObject(data) && isObject(data['model_info']) ? data['model_info'] : undefined;\n    const contextTokens = readContextLength(modelInfo);\n    if (contextTokens === undefined) {\n      throw new Error(`/api/show reports no context length for '${this.modelId}'`);\n    }\n    // Shared window: input and output draw from the same context — there is\n    // no separate output ceiling, so the window is published as both (the\n    // `maxOutputTokens === contextTokens` shape consumers key the split on).\n    return { contextTokens, maxOutputTokens: contextTokens };\n  }\n\n  async generateText(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<string> {\n    const response = await this.generateTextWithMetadata(prompt, maxTokens, temperature, options);\n    return response.text;\n  }\n\n  async generateTextWithMetadata(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<InferenceResponse> {\n    this.logger?.debug('Generating text with Ollama', {\n      model: this.modelId,\n      promptLength: prompt.length,\n      maxTokens,\n      temperature,\n      format: options?.format,\n    });\n\n    // Managed context window: size num_ctx to cover this request, capped at\n    // the model's discovered window. Without an explicit num_ctx Ollama uses\n    // the model's *default* window and SILENTLY CLIPS any prompt beyond it —\n    // input loss with no error (found 2026-07-30).\n    const limits = await this.limits();\n    const promptTokens = estimateTokens(prompt);\n    if (promptTokens + maxTokens > limits.contextTokens) {\n      throw new Error(\n        `Prompt (~${promptTokens} tokens) + output budget (${maxTokens}) exceed the ` +\n        `'${this.modelId}' context window (${limits.contextTokens} tokens)`,\n      );\n    }\n    const numCtx = Math.min(\n      limits.contextTokens,\n      promptTokens + maxTokens\n        + Math.ceil(promptTokens * NUM_CTX_ESTIMATE_SLACK) + NUM_CTX_TEMPLATE_ALLOWANCE,\n    );\n\n    const url = `${this.baseURL}/api/generate`;\n    const start = performance.now();\n\n    // Ollama's `format` parameter accepts either the literal string\n    // `\"json\"` (any valid JSON, including objects, numbers, etc.) or a\n    // JSON schema (constrains the top-level shape). The contract on the\n    // inference side is \"parseable JSON array,\" so we pass a minimal\n    // array schema rather than the bare `\"json\"` string — without it,\n    // the model can satisfy \"valid JSON\" by emitting `{\"entities\": [...]}`\n    // and break every consumer that expects to call `.map` on the\n    // top-level value. The schema's `items: {}` keeps element shape\n    // unconstrained — the prompt still carries the per-element schema;\n    // we only enforce the outer array.\n    const body: Record<string, unknown> = {\n      model: this.modelId,\n      prompt,\n      stream: false,\n      think: false,\n      options: {\n        num_predict: maxTokens,\n        num_ctx: numCtx,\n        temperature,\n      },\n    };\n    if (options?.format === 'json') {\n      body['format'] = { type: 'array', items: {} };\n    }\n\n    let res: Response;\n    try {\n      res = await fetch(url, {\n        method: 'POST',\n        headers: { 'Content-Type': 'application/json' },\n        body: JSON.stringify(body),\n      });\n    } catch (err) {\n      recordInferenceUsage({\n        provider: this.type,\n        model: this.modelId,\n        durationMs: performance.now() - start,\n        outcome: 'error',\n      });\n      throw err;\n    }\n\n    if (!res.ok) {\n      recordInferenceUsage({\n        provider: this.type,\n        model: this.modelId,\n        durationMs: performance.now() - start,\n        outcome: 'error',\n      });\n      const body = await res.text();\n      this.logger?.error('Ollama API error', {\n        model: this.modelId,\n        status: res.status,\n        body,\n      });\n      throw new Error(`Ollama API error (${res.status}): ${body}`);\n    }\n\n    const data = await res.json() as OllamaGenerateResponse;\n\n    if (!data.response) {\n      recordInferenceUsage({\n        provider: this.type,\n        model: this.modelId,\n        durationMs: performance.now() - start,\n        outcome: 'error',\n        inputTokens: data.prompt_eval_count,\n        outputTokens: data.eval_count,\n      });\n      this.logger?.error('Empty response from Ollama', { model: this.modelId });\n      throw new Error('Empty response from Ollama');\n    }\n\n    recordInferenceUsage({\n      provider: this.type,\n      model: this.modelId,\n      durationMs: performance.now() - start,\n      outcome: 'success',\n      inputTokens: data.prompt_eval_count,\n      outputTokens: data.eval_count,\n    });\n\n    const stopReason = mapStopReason(data.done_reason);\n\n    this.logger?.info('Text generation completed', {\n      model: this.modelId,\n      textLength: data.response.length,\n      stopReason,\n    });\n\n    return {\n      text: data.response,\n      stopReason,\n    };\n  }\n}\n\n/**\n * The context length lives in `model_info` under an architecture-prefixed key\n * (e.g. `llama.context_length`); `general.architecture` names the prefix.\n * Falls back to any `*.context_length` key for models whose metadata omits\n * the architecture field.\n */\nfunction readContextLength(modelInfo: Record<string, unknown> | undefined): number | undefined {\n  if (!modelInfo) return undefined;\n  const arch = modelInfo['general.architecture'];\n  if (typeof arch === 'string') {\n    const direct = modelInfo[`${arch}.context_length`];\n    if (isNumber(direct) && direct > 0) return direct;\n  }\n  const fallbackKey = Object.keys(modelInfo).find(k => k.endsWith('.context_length'));\n  if (fallbackKey !== undefined) {\n    const fallback = modelInfo[fallbackKey];\n    if (isNumber(fallback) && fallback > 0) return fallback;\n  }\n  return undefined;\n}\n\nfunction mapStopReason(doneReason: string | undefined): string {\n  switch (doneReason) {\n    case 'stop': return 'end_turn';\n    case 'length': return 'max_tokens';\n    default: return doneReason || 'unknown';\n  }\n}\n","// Factory for creating inference client instances based on configuration\n\nimport type { Logger } from '@semiont/core';\nimport { InferenceClient } from './interface.js';\nimport { AnthropicInferenceClient } from './implementations/anthropic.js';\nimport { OllamaInferenceClient } from './implementations/ollama.js';\n\nexport type InferenceClientType = 'anthropic' | 'ollama';\n\nexport interface InferenceClientConfig {\n  type: InferenceClientType;\n  apiKey?: string;\n  model: string;\n  endpoint?: string;\n  baseURL?: string;\n}\n\nexport function createInferenceClient(config: InferenceClientConfig, logger?: Logger): InferenceClient {\n  switch (config.type) {\n    case 'anthropic': {\n      if (!config.apiKey || config.apiKey.trim() === '') {\n        throw new Error('apiKey is required for Anthropic inference client');\n      }\n      return new AnthropicInferenceClient(\n        config.apiKey,\n        config.model,\n        config.endpoint || config.baseURL,\n        logger\n      );\n    }\n\n    case 'ollama': {\n      return new OllamaInferenceClient(\n        config.model,\n        config.endpoint || config.baseURL,\n        logger\n      );\n    }\n\n    default:\n      throw new Error(`Unsupported inference client type: ${config.type}`);\n  }\n}\n","// Mock implementation of InferenceClient for testing\n\nimport { InferenceClient, InferenceLimits, InferenceOptions, InferenceResponse } from '../interface.js';\n\n// Generous defaults so existing consumers never trip chunking or window\n// guards unless a test injects tighter limits deliberately.\nconst GENEROUS_LIMITS: InferenceLimits = {\n  contextTokens: 1_000_000,\n  maxOutputTokens: 1_000_000,\n};\n\nexport class MockInferenceClient implements InferenceClient {\n  readonly type = 'mock' as const;\n  readonly modelId = 'mock-model' as const;\n  private responses: string[] = [];\n  private responseIndex: number = 0;\n  private stopReasons: string[] = [];\n  private injectedLimits: InferenceLimits;\n  public calls: Array<{ prompt: string; maxTokens: number; temperature: number; options?: InferenceOptions }> = [];\n\n  constructor(responses: string[] = ['Mock response'], stopReasons?: string[], limits?: InferenceLimits) {\n    this.responses = responses;\n    this.stopReasons = stopReasons || responses.map(() => 'end_turn');\n    this.injectedLimits = limits ?? GENEROUS_LIMITS;\n  }\n\n  async limits(): Promise<InferenceLimits> {\n    return this.injectedLimits;\n  }\n\n  async generateText(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<string> {\n    const response = await this.generateTextWithMetadata(prompt, maxTokens, temperature, options);\n    return response.text;\n  }\n\n  async generateTextWithMetadata(prompt: string, maxTokens: number, temperature: number, options?: InferenceOptions): Promise<InferenceResponse> {\n    this.calls.push({ prompt, maxTokens, temperature, ...(options ? { options } : {}) });\n\n    const text = this.responses[this.responseIndex];\n    const stopReason = this.stopReasons[this.responseIndex] || 'end_turn';\n\n    if (this.responseIndex < this.responses.length - 1) {\n      this.responseIndex++;\n    }\n\n    return { text, stopReason };\n  }\n\n  // Test helper methods\n  reset(): void {\n    this.calls = [];\n    this.responseIndex = 0;\n  }\n\n  setResponses(responses: string[], stopReasons?: string[]): void {\n    this.responses = responses;\n    this.stopReasons = stopReasons || responses.map(() => 'end_turn');\n    this.responseIndex = 0;\n  }\n}\n"],"mappings":";AAEA,OAAO,eAAe;AAEtB,SAAS,4BAA4B;AASrC,IAAM,iCAAiC,KAAK,MAAM,QAAU,CAAC;AAa7D,IAAM,kBAAkC;AAAA,EACtC,MAAM;AAAA,EACN,aACE;AAAA,EACF,cAAc;AAAA,IACZ,MAAM;AAAA,IACN,YAAY;AAAA;AAAA;AAAA,MAGV,OAAO,EAAE,MAAM,SAAS,OAAO,CAAC,EAAE;AAAA,IACpC;AAAA,IACA,UAAU,CAAC,OAAO;AAAA,EACpB;AACF;AAEO,IAAM,2BAAN,MAA0D;AAAA,EACtD,OAAO;AAAA,EACP;AAAA,EACD;AAAA,EACA;AAAA,EACA;AAAA,EAER,YAAY,QAAgB,OAAe,SAAkB,QAAiB;AAC5E,SAAK,SAAS,IAAI,UAAU;AAAA,MAC1B;AAAA,MACA,SAAS,WAAW;AAAA,IACtB,CAAC;AACD,SAAK,UAAU;AACf,SAAK,SAAS;AAAA,EAChB;AAAA,EAEA,SAAmC;AACjC,QAAI,CAAC,KAAK,eAAe;AACvB,WAAK,gBAAgB,KAAK,eAAe,EAAE,MAAM,CAAC,QAAiB;AAGjE,aAAK,gBAAgB;AACrB,cAAM;AAAA,MACR,CAAC;AAAA,IACH;AACA,WAAO,KAAK;AAAA,EACd;AAAA,EAEA,MAAc,iBAA2C;AAGvD,UAAM,OAAO,MAAM,KAAK,OAAO,OAAO,SAAS,KAAK,OAAO,EAAE,MAAM,CAAC,QAAiB;AACnF,YAAM,IAAI;AAAA,QACR,wCAAwC,KAAK,OAAO;AAAA,QACpD,EAAE,OAAO,IAAI;AAAA,MACf;AAAA,IACF,CAAC;AACD,QAAI,KAAK,oBAAoB,QAAQ,KAAK,cAAc,MAAM;AAC5D,YAAM,IAAI,MAAM,sDAAsD,KAAK,OAAO,GAAG;AAAA,IACvF;AACA,WAAO,EAAE,eAAe,KAAK,kBAAkB,iBAAiB,KAAK,WAAW;AAAA,EAClF;AAAA,EAEQ,eAAe,QAA+E;AACpG,QAAI,OAAO,aAAa,gCAAgC;AACtD,aAAO,KAAK,OAAO,SAAS,OAAO,MAAM,EAAE,aAAa;AAAA,IAC1D;AACA,WAAO,KAAK,OAAO,SAAS,OAAO,MAAM;AAAA,EAC3C;AAAA,EAEA,MAAM,aAAa,QAAgB,WAAmB,aAAqB,SAA6C;AACtH,UAAM,WAAW,MAAM,KAAK,yBAAyB,QAAQ,WAAW,aAAa,OAAO;AAC5F,WAAO,SAAS;AAAA,EAClB;AAAA,EAEA,MAAM,yBAAyB,QAAgB,WAAmB,aAAqB,SAAwD;AAC7I,UAAM,WAAW,SAAS,WAAW;AAErC,SAAK,QAAQ,MAAM,yCAAyC;AAAA,MAC1D,OAAO,KAAK;AAAA,MACZ,cAAc,OAAO;AAAA,MACrB;AAAA,MACA;AAAA,MACA,QAAQ,SAAS;AAAA,IACnB,CAAC;AAED,UAAM,SAAoD;AAAA,MACxD,OAAO,KAAK;AAAA,MACZ,YAAY;AAAA,MACZ;AAAA,MACA,UAAU,CAAC,EAAE,MAAM,QAAQ,SAAS,OAAO,CAAC;AAAA;AAAA;AAAA,MAG5C,GAAI,WACA,EAAE,OAAO,CAAC,eAAe,GAAG,aAAa,EAAE,MAAM,QAAiB,MAAM,gBAAgB,KAAK,EAAE,IAC/F,CAAC;AAAA,IACP;AAEA,UAAM,QAAQ,YAAY,IAAI;AAC9B,QAAI;AACJ,QAAI;AACF,iBAAW,MAAM,KAAK,eAAe,MAAM;AAAA,IAC7C,SAAS,KAAK;AACZ,2BAAqB;AAAA,QACnB,UAAU,KAAK;AAAA,QACf,OAAO,KAAK;AAAA,QACZ,YAAY,YAAY,IAAI,IAAI;AAAA,QAChC,SAAS;AAAA,MACX,CAAC;AACD,YAAM;AAAA,IACR;AAEA,SAAK,QAAQ,MAAM,+BAA+B;AAAA,MAChD,OAAO,KAAK;AAAA,MACZ,eAAe,SAAS,QAAQ;AAAA,MAChC,YAAY,SAAS;AAAA,IACvB,CAAC;AAED,QAAI;AACJ,QAAI,UAAU;AAIZ,YAAM,UAAU,SAAS,QAAQ,KAAK,OAAK,EAAE,SAAS,UAAU;AAChE,UAAI,CAAC,WAAW,QAAQ,SAAS,YAAY;AAC3C,6BAAqB;AAAA,UACnB,UAAU,KAAK;AAAA,UACf,OAAO,KAAK;AAAA,UACZ,YAAY,YAAY,IAAI,IAAI;AAAA,UAChC,SAAS;AAAA,UACT,aAAa,SAAS,OAAO;AAAA,UAC7B,cAAc,SAAS,OAAO;AAAA,QAChC,CAAC;AACD,aAAK,QAAQ,MAAM,6CAA6C;AAAA,UAC9D,OAAO,KAAK;AAAA,UACZ,cAAc,SAAS,QAAQ,IAAI,OAAK,EAAE,IAAI;AAAA,QAChD,CAAC;AACD,cAAM,IAAI,MAAM,2CAA2C;AAAA,MAC7D;AAIA,YAAM,QAAQ,QAAQ;AACtB,YAAM,QAAQ,MAAM,QAAQ,MAAM,KAAK,IAAI,MAAM,QAAQ,CAAC;AAC1D,aAAO,KAAK,UAAU,KAAK;AAAA,IAC7B,OAAO;AACL,YAAM,cAAc,SAAS,QAAQ,KAAK,OAAK,EAAE,SAAS,MAAM;AAChE,UAAI,CAAC,eAAe,YAAY,SAAS,QAAQ;AAC/C,6BAAqB;AAAA,UACnB,UAAU,KAAK;AAAA,UACf,OAAO,KAAK;AAAA,UACZ,YAAY,YAAY,IAAI,IAAI;AAAA,UAChC,SAAS;AAAA,UACT,aAAa,SAAS,OAAO;AAAA,UAC7B,cAAc,SAAS,OAAO;AAAA,QAChC,CAAC;AACD,aAAK,QAAQ,MAAM,yCAAyC;AAAA,UAC1D,OAAO,KAAK;AAAA,UACZ,cAAc,SAAS,QAAQ,IAAI,OAAK,EAAE,IAAI;AAAA,QAChD,CAAC;AACD,cAAM,IAAI,MAAM,uCAAuC;AAAA,MACzD;AACA,aAAO,YAAY;AAAA,IACrB;AAEA,yBAAqB;AAAA,MACnB,UAAU,KAAK;AAAA,MACf,OAAO,KAAK;AAAA,MACZ,YAAY,YAAY,IAAI,IAAI;AAAA,MAChC,SAAS;AAAA,MACT,aAAa,SAAS,OAAO;AAAA,MAC7B,cAAc,SAAS,OAAO;AAAA,IAChC,CAAC;AAED,SAAK,QAAQ,KAAK,6BAA6B;AAAA,MAC7C,OAAO,KAAK;AAAA,MACZ,YAAY,KAAK;AAAA,MACjB,YAAY,SAAS;AAAA,IACvB,CAAC;AAED,WAAO;AAAA,MACL;AAAA,MACA,YAAY,SAAS,eAAe;AAAA,IACtC;AAAA,EACF;AACF;;;AC3MA,SAAS,gBAAgB,UAAU,gBAAgB;AAEnD,SAAS,wBAAAA,6BAA4B;AASrC,IAAM,yBAAyB;AAC/B,IAAM,6BAA6B;AAY5B,IAAM,wBAAN,MAAuD;AAAA,EACnD,OAAO;AAAA,EACP;AAAA,EACD;AAAA,EACA;AAAA,EAEA;AAAA,EAER,YAAY,OAAe,SAAkB,QAAiB;AAC5D,SAAK,WAAW,WAAW,0BAA0B,QAAQ,QAAQ,EAAE;AACvE,SAAK,UAAU;AACf,SAAK,SAAS;AAAA,EAChB;AAAA,EAEA,SAAmC;AACjC,QAAI,CAAC,KAAK,eAAe;AACvB,WAAK,gBAAgB,KAAK,eAAe,EAAE,MAAM,CAAC,QAAiB;AAGjE,aAAK,gBAAgB;AACrB,cAAM;AAAA,MACR,CAAC;AAAA,IACH;AACA,WAAO,KAAK;AAAA,EACd;AAAA,EAEA,MAAc,iBAA2C;AACvD,UAAM,MAAM,MAAM,MAAM,GAAG,KAAK,OAAO,aAAa;AAAA,MAClD,QAAQ;AAAA,MACR,SAAS,EAAE,gBAAgB,mBAAmB;AAAA,MAC9C,MAAM,KAAK,UAAU,EAAE,OAAO,KAAK,QAAQ,CAAC;AAAA,IAC9C,CAAC;AACD,QAAI,CAAC,IAAI,IAAI;AACX,YAAM,IAAI;AAAA,QACR,uDAAuD,IAAI,MAAM,SAAS,KAAK,OAAO;AAAA,MACxF;AAAA,IACF;AACA,UAAM,OAAgB,MAAM,IAAI,KAAK;AACrC,UAAM,YAAY,SAAS,IAAI,KAAK,SAAS,KAAK,YAAY,CAAC,IAAI,KAAK,YAAY,IAAI;AACxF,UAAM,gBAAgB,kBAAkB,SAAS;AACjD,QAAI,kBAAkB,QAAW;AAC/B,YAAM,IAAI,MAAM,4CAA4C,KAAK,OAAO,GAAG;AAAA,IAC7E;AAIA,WAAO,EAAE,eAAe,iBAAiB,cAAc;AAAA,EACzD;AAAA,EAEA,MAAM,aAAa,QAAgB,WAAmB,aAAqB,SAA6C;AACtH,UAAM,WAAW,MAAM,KAAK,yBAAyB,QAAQ,WAAW,aAAa,OAAO;AAC5F,WAAO,SAAS;AAAA,EAClB;AAAA,EAEA,MAAM,yBAAyB,QAAgB,WAAmB,aAAqB,SAAwD;AAC7I,SAAK,QAAQ,MAAM,+BAA+B;AAAA,MAChD,OAAO,KAAK;AAAA,MACZ,cAAc,OAAO;AAAA,MACrB;AAAA,MACA;AAAA,MACA,QAAQ,SAAS;AAAA,IACnB,CAAC;AAMD,UAAM,SAAS,MAAM,KAAK,OAAO;AACjC,UAAM,eAAe,eAAe,MAAM;AAC1C,QAAI,eAAe,YAAY,OAAO,eAAe;AACnD,YAAM,IAAI;AAAA,QACR,YAAY,YAAY,6BAA6B,SAAS,iBAC1D,KAAK,OAAO,qBAAqB,OAAO,aAAa;AAAA,MAC3D;AAAA,IACF;AACA,UAAM,SAAS,KAAK;AAAA,MAClB,OAAO;AAAA,MACP,eAAe,YACX,KAAK,KAAK,eAAe,sBAAsB,IAAI;AAAA,IACzD;AAEA,UAAM,MAAM,GAAG,KAAK,OAAO;AAC3B,UAAM,QAAQ,YAAY,IAAI;AAY9B,UAAM,OAAgC;AAAA,MACpC,OAAO,KAAK;AAAA,MACZ;AAAA,MACA,QAAQ;AAAA,MACR,OAAO;AAAA,MACP,SAAS;AAAA,QACP,aAAa;AAAA,QACb,SAAS;AAAA,QACT;AAAA,MACF;AAAA,IACF;AACA,QAAI,SAAS,WAAW,QAAQ;AAC9B,WAAK,QAAQ,IAAI,EAAE,MAAM,SAAS,OAAO,CAAC,EAAE;AAAA,IAC9C;AAEA,QAAI;AACJ,QAAI;AACF,YAAM,MAAM,MAAM,KAAK;AAAA,QACrB,QAAQ;AAAA,QACR,SAAS,EAAE,gBAAgB,mBAAmB;AAAA,QAC9C,MAAM,KAAK,UAAU,IAAI;AAAA,MAC3B,CAAC;AAAA,IACH,SAAS,KAAK;AACZ,MAAAA,sBAAqB;AAAA,QACnB,UAAU,KAAK;AAAA,QACf,OAAO,KAAK;AAAA,QACZ,YAAY,YAAY,IAAI,IAAI;AAAA,QAChC,SAAS;AAAA,MACX,CAAC;AACD,YAAM;AAAA,IACR;AAEA,QAAI,CAAC,IAAI,IAAI;AACX,MAAAA,sBAAqB;AAAA,QACnB,UAAU,KAAK;AAAA,QACf,OAAO,KAAK;AAAA,QACZ,YAAY,YAAY,IAAI,IAAI;AAAA,QAChC,SAAS;AAAA,MACX,CAAC;AACD,YAAMC,QAAO,MAAM,IAAI,KAAK;AAC5B,WAAK,QAAQ,MAAM,oBAAoB;AAAA,QACrC,OAAO,KAAK;AAAA,QACZ,QAAQ,IAAI;AAAA,QACZ,MAAAA;AAAA,MACF,CAAC;AACD,YAAM,IAAI,MAAM,qBAAqB,IAAI,MAAM,MAAMA,KAAI,EAAE;AAAA,IAC7D;AAEA,UAAM,OAAO,MAAM,IAAI,KAAK;AAE5B,QAAI,CAAC,KAAK,UAAU;AAClB,MAAAD,sBAAqB;AAAA,QACnB,UAAU,KAAK;AAAA,QACf,OAAO,KAAK;AAAA,QACZ,YAAY,YAAY,IAAI,IAAI;AAAA,QAChC,SAAS;AAAA,QACT,aAAa,KAAK;AAAA,QAClB,cAAc,KAAK;AAAA,MACrB,CAAC;AACD,WAAK,QAAQ,MAAM,8BAA8B,EAAE,OAAO,KAAK,QAAQ,CAAC;AACxE,YAAM,IAAI,MAAM,4BAA4B;AAAA,IAC9C;AAEA,IAAAA,sBAAqB;AAAA,MACnB,UAAU,KAAK;AAAA,MACf,OAAO,KAAK;AAAA,MACZ,YAAY,YAAY,IAAI,IAAI;AAAA,MAChC,SAAS;AAAA,MACT,aAAa,KAAK;AAAA,MAClB,cAAc,KAAK;AAAA,IACrB,CAAC;AAED,UAAM,aAAa,cAAc,KAAK,WAAW;AAEjD,SAAK,QAAQ,KAAK,6BAA6B;AAAA,MAC7C,OAAO,KAAK;AAAA,MACZ,YAAY,KAAK,SAAS;AAAA,MAC1B;AAAA,IACF,CAAC;AAED,WAAO;AAAA,MACL,MAAM,KAAK;AAAA,MACX;AAAA,IACF;AAAA,EACF;AACF;AAQA,SAAS,kBAAkB,WAAoE;AAC7F,MAAI,CAAC,UAAW,QAAO;AACvB,QAAM,OAAO,UAAU,sBAAsB;AAC7C,MAAI,OAAO,SAAS,UAAU;AAC5B,UAAM,SAAS,UAAU,GAAG,IAAI,iBAAiB;AACjD,QAAI,SAAS,MAAM,KAAK,SAAS,EAAG,QAAO;AAAA,EAC7C;AACA,QAAM,cAAc,OAAO,KAAK,SAAS,EAAE,KAAK,OAAK,EAAE,SAAS,iBAAiB,CAAC;AAClF,MAAI,gBAAgB,QAAW;AAC7B,UAAM,WAAW,UAAU,WAAW;AACtC,QAAI,SAAS,QAAQ,KAAK,WAAW,EAAG,QAAO;AAAA,EACjD;AACA,SAAO;AACT;AAEA,SAAS,cAAc,YAAwC;AAC7D,UAAQ,YAAY;AAAA,IAClB,KAAK;AAAQ,aAAO;AAAA,IACpB,KAAK;AAAU,aAAO;AAAA,IACtB;AAAS,aAAO,cAAc;AAAA,EAChC;AACF;;;AC1NO,SAAS,sBAAsB,QAA+B,QAAkC;AACrG,UAAQ,OAAO,MAAM;AAAA,IACnB,KAAK,aAAa;AAChB,UAAI,CAAC,OAAO,UAAU,OAAO,OAAO,KAAK,MAAM,IAAI;AACjD,cAAM,IAAI,MAAM,mDAAmD;AAAA,MACrE;AACA,aAAO,IAAI;AAAA,QACT,OAAO;AAAA,QACP,OAAO;AAAA,QACP,OAAO,YAAY,OAAO;AAAA,QAC1B;AAAA,MACF;AAAA,IACF;AAAA,IAEA,KAAK,UAAU;AACb,aAAO,IAAI;AAAA,QACT,OAAO;AAAA,QACP,OAAO,YAAY,OAAO;AAAA,QAC1B;AAAA,MACF;AAAA,IACF;AAAA,IAEA;AACE,YAAM,IAAI,MAAM,sCAAsC,OAAO,IAAI,EAAE;AAAA,EACvE;AACF;;;ACpCA,IAAM,kBAAmC;AAAA,EACvC,eAAe;AAAA,EACf,iBAAiB;AACnB;AAEO,IAAM,sBAAN,MAAqD;AAAA,EACjD,OAAO;AAAA,EACP,UAAU;AAAA,EACX,YAAsB,CAAC;AAAA,EACvB,gBAAwB;AAAA,EACxB,cAAwB,CAAC;AAAA,EACzB;AAAA,EACD,QAAuG,CAAC;AAAA,EAE/G,YAAY,YAAsB,CAAC,eAAe,GAAG,aAAwB,QAA0B;AACrG,SAAK,YAAY;AACjB,SAAK,cAAc,eAAe,UAAU,IAAI,MAAM,UAAU;AAChE,SAAK,iBAAiB,UAAU;AAAA,EAClC;AAAA,EAEA,MAAM,SAAmC;AACvC,WAAO,KAAK;AAAA,EACd;AAAA,EAEA,MAAM,aAAa,QAAgB,WAAmB,aAAqB,SAA6C;AACtH,UAAM,WAAW,MAAM,KAAK,yBAAyB,QAAQ,WAAW,aAAa,OAAO;AAC5F,WAAO,SAAS;AAAA,EAClB;AAAA,EAEA,MAAM,yBAAyB,QAAgB,WAAmB,aAAqB,SAAwD;AAC7I,SAAK,MAAM,KAAK,EAAE,QAAQ,WAAW,aAAa,GAAI,UAAU,EAAE,QAAQ,IAAI,CAAC,EAAG,CAAC;AAEnF,UAAM,OAAO,KAAK,UAAU,KAAK,aAAa;AAC9C,UAAM,aAAa,KAAK,YAAY,KAAK,aAAa,KAAK;AAE3D,QAAI,KAAK,gBAAgB,KAAK,UAAU,SAAS,GAAG;AAClD,WAAK;AAAA,IACP;AAEA,WAAO,EAAE,MAAM,WAAW;AAAA,EAC5B;AAAA;AAAA,EAGA,QAAc;AACZ,SAAK,QAAQ,CAAC;AACd,SAAK,gBAAgB;AAAA,EACvB;AAAA,EAEA,aAAa,WAAqB,aAA8B;AAC9D,SAAK,YAAY;AACjB,SAAK,cAAc,eAAe,UAAU,IAAI,MAAM,UAAU;AAChE,SAAK,gBAAgB;AAAA,EACvB;AACF;","names":["recordInferenceUsage","body"]}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@semiont/inference",
-  "version": "0.5.23",
+  "version": "0.5.25",
   "engines": {
     "node": ">=24.0.0"
   },
@@ -27,13 +27,13 @@
     "test:coverage": "vitest run --coverage"
   },
   "dependencies": {
-    "@anthropic-ai/sdk": "^0.112.5",
-    "@semiont/core": "0.5.23",
-    "@semiont/observability": "0.5.23"
+    "@anthropic-ai/sdk": "^0.115.0",
+    "@semiont/core": "0.5.25",
+    "@semiont/observability": "0.5.25"
   },
   "devDependencies": {
     "@vitest/coverage-v8": "^4.1.8",
-    "rollup": "^4.61.0",
+    "rollup": "^4.62.3",
     "rollup-plugin-dts": "^6.4.1",
     "tsup": "^8.0.1",
     "typescript": "^6.0.2",