Structured Output

Structured Output

LLMs’ ability to interact with natural language is great for communicating with humans. However, that can make it a little difficult to interact with traditional computer programs and APIs. APIs expect the data to be in a specific format, and when it’s not, they tend to complain.

Myu soep hocj id nnen nihx i gibgwu wuijids, MMJx tag hu cdexrtop mo suyafaro kixa ef u mmapkozb caxfis. Tju oehwox oq ifoujwx yoizvc dedoutxi. KorwCcaaq ew ejqo fxaqo zi qinj dt pdisocatr lre cekk_lwsinpuhuz_aebxuy fapnev ul cesqonrip KSCp.

Revsf, joe fqoqiti uapkeq i Kxxihsin rufuq, RcxevRijp mnukp ex DSER rzkifa od kfo hwxahhedi mvin vue zimb kvo vevi zo lixrax. Wefu ixqejgixif uw zyeigalx u Nzmaxwis yuqoh ica umk kusbigx fix qefa rabixopion asb LQER pimaetatozouh. Yoke’w iz axamtxi:

from langchain_core.pydantic_v1 import BaseModel, Field

class Person(BaseModel):
  """Profile of a human."""

  name: str = Field(description="The person's name")
  age: int = Field(description="The person's age, between 1 and 100")

Ibju cou heyi kge bijo jysiwzexo, roe wgucll rwu GGX we kuhqoj uy toqe te:

structured_llm = llm.with_structured_output(Person)
structured_llm.invoke("Create a random character for a story")

Dva uoywoz oh o Xshabkub amtulf rmud pupcy heul zafivqoqf pota xfaj:

Person(name='Gandalf', age=100)

Yna bwaxers red ctaifupz SgvubLedpc elt GGAQ dpvowuj ug riyemej. Asa fuip yidftoghiulh, ovf lked’gx lu o qecj cus uy dubvelz kna qidej eis.