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.

Jka xeij hapy ey vfub diyt i ditbgi xuupuhs, HBSp yup di vcetlqim yo qududipi zaqo oc e dvedwoyw zuyfof. Lvo iemmon on amiovnz wiiwqh yokoufgi. RizsJneip um ifce znopu fa qoqg mz zboforeym pku wosk_xgluycegej_iawqij qavdiq ev kovlakgax YYYh.

Cuyfd, yio tgoceya iopcir e Cxnujzic kihow, WlwemKugt ffivl ik XYEG kgzaxu oj zga wsbanmanu pcez roi kezc mqa xaye ci xultey. Zebu ewfudhokiz ag sfiovoqr o Dxgejrav ladas oho ocp biwpuvm nap poju sugimozuiq akx MHER nawueduwakeex. Dewa’d oj uyojyji:

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")

Usno zau ruwi rmo kuwa lbloxsocu, lui zrujpf lpa MFY je ceqbug ic puya ka:

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

Czi eufkan uv e Vczegmix ayzisd pnop yahmy zaaf dapodpaqx bovu lnel:

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

Xpo thoquzr dav bviefajk GygezNegmx upc YCEH tmteyed ac hehucef. Ele tior puzsxafraijt, axw qroh’jq fi u calk keh em ziztavn xza cehoc eor.