SemanticSlotDefinition
KIT · interface · Type contract
Declares the {{namespace:key}} slots a model may use, and how often.
Import
ts
import type { SemanticSlotDefinition } from '@jangts/agent-core/kits/presentation';js
/** @typedef {import('@jangts/agent-core/kits/presentation').SemanticSlotDefinition} SemanticSlotDefinition */Other public entry points:@jangts/agent-core, @jangts/agent-core/types, @jangts/agent-core/structs, @jangts/agent-core/kits/presentation, @jangts/agent-core/kits.
Definition
ts
interface SemanticSlotDefinitionField structure
| Field | Type | Required | Description |
|---|---|---|---|
namespace | string | Required | Read-only. Slot namespace; must match ^[a-z][a-z0-9-]*$. |
keys | readonly string[] | Required | Read-only. Allowed keys after the namespace; any other key makes the text invalid. |
maxOccurrences | number | Required | Read-only. Maximum total occurrences across the text; must be a non-negative integer. |
Examples
- Presentation: field ownership and final output: The model fills only the fields it is assigned; fixed text, bound values, and semantic keys are owned by the program.
- Templates, semantic slots, and expression selection: The program owns templates and preference order, inserted values are replaced only once, and the App supplies the random source for variants.
First compose one piece of generated text
Beginner · Node.js
ts
import { Presentation } from '@jangts/agent-core/kits/presentation';
const outputs = Presentation.defineOutputCatalog([
{ id: 'reply', kind: 'text', fields: [{ id: 'text', owner: 'generated' }] },
]);
console.log(outputs.toModelSchema('reply'));
const output = outputs.compose({ outputId: 'reply', input: { generated: { text: '普通对象即可定义节点。' } } }, { app: {}, signal: new AbortController().signal });
console.log(output.value);js
import { Presentation } from '@jangts/agent-core/kits/presentation';
const outputs = Presentation.defineOutputCatalog([
{ id: 'reply', kind: 'text', fields: [{ id: 'text', owner: 'generated' }] },
]);
console.log(outputs.toModelSchema('reply'));
const output = outputs.compose({ outputId: 'reply', input: { generated: { text: '普通对象即可定义节点。' } } }, { app: {}, signal: new AbortController().signal });
console.log(output.value);Let the program own the explanatory prefix in the same compose
Configuration · Node.js
ts
import { Presentation } from '@jangts/agent-core/kits/presentation';
const outputs = Presentation.defineOutputCatalog([
{ id: 'reply', kind: 'master', master: '{{prefix}}{{text}}', fields: [
{ id: 'prefix', owner: 'constant', constantValue: '接口说明:' },
{ id: 'text', owner: 'generated' },
] },
]);
console.log(outputs.toModelSchema('reply')); // 只有 text 由模型填写
console.log(outputs.compose({ outputId: 'reply', input: { generated: { text: '普通对象即可定义节点。' } } }, { app: {}, signal: new AbortController().signal }).value);js
import { Presentation } from '@jangts/agent-core/kits/presentation';
const outputs = Presentation.defineOutputCatalog([
{ id: 'reply', kind: 'master', master: '{{prefix}}{{text}}', fields: [
{ id: 'prefix', owner: 'constant', constantValue: '接口说明:' },
{ id: 'text', owner: 'generated' },
] },
]);
console.log(outputs.toModelSchema('reply')); // 只有 text 由模型填写
console.log(outputs.compose({ outputId: 'reply', input: { generated: { text: '普通对象即可定义节点。' } } }, { app: {}, signal: new AbortController().signal }).value);