44 lines
808 B
Markdown
44 lines
808 B
Markdown
---
|
|
layout: default
|
|
---
|
|
|
|
# Structured Output
|
|
|
|
<div class="text-sm mt-6">
|
|
|
|
<v-click>
|
|
|
|
<div class="mb-4">
|
|
|
|
**Problema:** Las respuestas de LLMs son texto plano, difíciles de procesar
|
|
|
|
**Solución:** Forzar al modelo a responder con JSON estructurado
|
|
|
|
</div>
|
|
|
|
</v-click>
|
|
|
|
<v-click>
|
|
|
|
```typescript
|
|
import { ChatOpenAI } from "@langchain/openai";
|
|
import { z } from "zod";
|
|
|
|
// Definir el esquema con Zod
|
|
const schema = z.object({
|
|
sentiment: z.enum(["positive", "negative", "neutral"]),
|
|
score: z.number().min(0).max(1),
|
|
keywords: z.array(z.string()),
|
|
});
|
|
|
|
const model = new ChatOpenAI({ model: "gpt-4.1-mini" })
|
|
.withStructuredOutput(schema);
|
|
|
|
const result = await model.invoke("Me encanta LangChain!");
|
|
// { sentiment: "positive", score: 0.95, keywords: ["encanta", "langchain"] }
|
|
```
|
|
|
|
</v-click>
|
|
|
|
</div>
|