MockGPT

KIT · class · Runtime value

Configurable mock model instance. Create one per test or per mock server.

Import

TypeScript

import { MockGPT } from '@jangts/agent-core/mock';

JavaScript

import { MockGPT } from '@jangts/agent-core/mock';

Other public entry points:@jangts/agent-core/mock.

Definition

class MockGPT

Constructor

new MockGPT(options: MockGPTOptions): MockGPT

Validate options and snapshot every rule.

Each rule's id, prompt, model, predicate and output are validated here, so configuration mistakes surface at construction rather than on the first request.

Parameter	Type	Passing	Description
options	MockGPTOptions	Required	—

Fields of options

Field	Type	Required	Description
rules	readonly Rule[]	Required	Ordered rules; the first rule whose filters all pass wins, so order is significant.
fallback	Generator | undefined	Optional	Generator used when no rule matches; absent makes an unmatched call a MOCK_RULE_NOT_FOUND error.
selectPrompt	((request: ModelRequest) => string) | undefined	Optional	Overrides prompt extraction; must return a string synchronously; default is the last user text.
stream	Pick<OutputOptions, "chunkSize" | "intervalMs" | "wireChunkSize"> | undefined	Optional	Defaults merged under every generated output; explicit output fields take precedence.
seed	number | undefined	Optional	Seed for the deterministic PRNG (0-0xffffffff); default 1, so identical call sequences repeat.
clock	(() => number) | undefined	Optional	Millisecond timestamp source for date generators and response created; must be finite.
recordLimit	number | undefined	Optional	Ring-buffer size of records; 0 disables recording; max 10 000; default 100.
maxBodyBytes	number | undefined	Optional	Maximum accepted request body in bytes for fetch; default 1 MiB.
maxOutputChars	number | undefined	Optional	Maximum generated content in code points plus serialized tool calls; default 1 000 000.

Returns: MockGPT.

Instance properties

Field	Type	Required	Description
records	readonly CallRecord[]	Required	Read-only. Chronological records, oldest first, as detached copies. Rotates the ring buffer back into insertion order before returning, and deep-copies each record so callers cannot mutate runtime state through the result.
counts	Readonly<Record<string, number>>	Required	Read-only. Accepted-call count per rule plus the fallback bucket, keyed by rule id.
activeRequests	number	Required	Read-only. Number of calls or response streams still in flight.

Member index

Member	Purpose
reset	Clear history, hit counts and the PRNG sequence.
fetch	Bound callback suitable for injection into an SDK's fetch option. Performs the HTTP side of the contract: route and method checks, byte-limited body reading, request validation, then either a JSON/SSE Response or an OpenAI-shaped error response. Transport faults (responses.reject) are rethrown as MOCK_NETWORK so SDKs observe a network failure rather than an HTTP error.
generate	Evaluate rules and return data. No Request, Response, delay or streaming is created. Transport-only options in a generated descriptor (delayMs, intervalMs, interruptAfter) are validated but ignored here, so SDK tests can exercise generators and matching without simulating a network.

Member details

reset

Instance method.

Clear history, hit counts and the PRNG sequence.

reset(): void

Clear history, hit counts and the PRNG sequence.

No call parameters.

Returns: void.

fetch

Instance method.

Bound callback suitable for injection into an SDK's fetch option.

Performs the HTTP side of the contract: route and method checks, byte-limited body reading, request validation, then either a JSON/SSE Response or an OpenAI-shaped error response. Transport faults (responses.reject) are rethrown as MOCK_NETWORK so SDKs observe a network failure rather than an HTTP error.

fetch(input: RequestInfo | URL, init?: RequestInit): Promise<Response>

MDN Reference

Parameter	Type	Passing	Description
input	RequestInfo | URL	Required	—
init	RequestInit | undefined	Optional	—

Returns: Promise<Response>.

Awaited value: Response.

generate

Instance method.

Evaluate rules and return data. No Request, Response, delay or streaming is created.

Transport-only options in a generated descriptor (delayMs, intervalMs, interruptAfter) are validated but ignored here, so SDK tests can exercise generators and matching without simulating a network.

generate(input: GenerationInput, { signal }?: GenerationOptions): Promise<Generated>

Evaluate rules and return data. No Request, Response, delay or streaming is created.

Transport-only options in a generated descriptor (delayMs, intervalMs, interruptAfter) are validated but ignored here, so SDK tests can exercise generators and matching without simulating a network.

Parameter	Type	Passing	Description
input	GenerationInput	Required	—
__1	GenerationOptions	Optional	—

Fields of input

Field	Type	Required	Description
model	string	Required	Read-only. Model identifier used for Rule.model filtering.
messages	readonly Readonly<{ [key: string]: unknown; role: string; content?: unknown; tool_calls?: unknown; tool_call_id?: string; }>[]	Required	Read-only. Conversation whose last user message becomes the prompt.

Fields of __1

Field	Type	Required	Description
signal	AbortSignal | undefined	Optional	Cancels the call before or during generation; rejection reason is preserved as-is.

Returns: Promise<Generated>.

Awaited value: Generated.

Examples

• MockGPT: direct generation and Provider injection: Get a string or JSON directly; inject the same instance into a Provider when you need to exercise model communication.

Generate native JSON directly

Beginner · Node.js

TypeScript

import { MockGPT } from '@jangts/agent-core/mock';
const mock = new MockGPT({ rules: [{ prompt: '服务状态', generator: { online: true, source: 'fixture' } }] });
const value = await mock.generate({ model: 'demo', messages: [{ role: 'user', content: '服务状态' }] });
console.log(value); // { online: true, source: 'fixture' }
console.log(mock.counts, mock.records);

JavaScript

import { MockGPT } from '@jangts/agent-core/mock';
const mock = new MockGPT({ rules: [{ prompt: '服务状态', generator: { online: true, source: 'fixture' } }] });
const value = await mock.generate({ model: 'demo', messages: [{ role: 'user', content: '服务状态' }] });
console.log(value); // { online: true, source: 'fixture' }
console.log(mock.counts, mock.records);

Inject the same instance into the real Core model boundary

Configuration · Node.js

TypeScript

import { ModelRunner, Support } from '@jangts/agent-core';
import { MockGPT, responses } from '@jangts/agent-core/mock';
const mock = new MockGPT({ rules: [{ prompt: '服务状态', generator: responses.text('服务在线。', { chunkSize: 2, wireChunkSize: 3 }) }] });
const provider = new Support.CompatibleModelProvider({
  config: { baseUrl: 'https://mock.local/v1', apiKey: 'offline-example', model: 'demo' },
  ports: { request: mock.fetch },
});
const result = await new ModelRunner({ provider, maxCalls: 1 }).invoke('服务状态');
console.log(result.value, result.invocation.status);

JavaScript

import { ModelRunner, Support } from '@jangts/agent-core';
import { MockGPT, responses } from '@jangts/agent-core/mock';
const mock = new MockGPT({ rules: [{ prompt: '服务状态', generator: responses.text('服务在线。', { chunkSize: 2, wireChunkSize: 3 }) }] });
const provider = new Support.CompatibleModelProvider({
    config: { baseUrl: 'https://mock.local/v1', apiKey: 'offline-example', model: 'demo' },
    ports: { request: mock.fetch },
});
const result = await new ModelRunner({ provider, maxCalls: 1 }).invoke('服务状态');
console.log(result.value, result.invocation.status);

Related definitions

• CallRecord
• Generated
• GenerationInput
• GenerationOptions
• JsonValue
• MockGPTOptions
• ResponseDescription