xpark.dataset.TextGenerate#

class xpark.dataset.TextGenerate(*, base_url: str, model: str, api_key: str = 'NOT_SET', max_qps: int | None = None, max_concurrency: int | None = None, max_retries: int = 0, fallback_response: str | None = None, fallback_on_request_error: bool = True, hint: str | list[str] | None = None, **kwargs: dict[str, Any])#

TextGenerate processor generates content based on the input parameters. generate by llm model.

Parameters:

  • base_url – The base URL of the LLM server.

  • model – The request model name.

  • api_key – The request API key.

  • max_qps – The maximum query-per-second rate for remote LLM requests.

  • max_concurrency – The maximum number of in-flight remote LLM requests allowed concurrently.

  • max_retries – The maximum number of retries per request in the event of failures. We retry with exponential backoff upto this specific maximum retries.

  • fallback_response – The response value to return when the LLM returns an empty response. If set to None, a ValueError will be raised instead.

  • fallback_on_request_error – If True and fallback_response is set, request-time errors (e.g. content-safety rejection, single-record 4xx) raised by the underlying LLM client are also translated into fallback_response instead of aborting the whole batch. Defaults to True; set it to False to let infrastructure failures surface to the caller instead.

  • hint – Optional extra instructions or constraints to guide the model (e.g. target length, tone, audience, style). Accepts either a single string or a list of strings, where each item is one hint written in plain text. Passing a list is recommended — use one string per hint.

  • **kwargs – Keyword arguments to pass to the openai.AsyncClient.chat.completions.create API.

Examples

from xpark.dataset.expressions import col
from xpark.dataset import TextGenerate, from_items

ds = from_items([""])
ds = ds.with_column(
    "generated_text",
    TextGenerate(
        model="deepseek-v3-0324",
        base_url=os.getenv("LLM_ENDPOINT"),
        api_key=os.getenv("LLM_API_KEY"),
    )
    .options(num_workers={"IO": 1}, batch_size=1)
    .with_column(col("item")),
)

print(ds.take_all())

Methods

__call__(texts)

Call self as a function.

options(**kwargs)

with_column(texts)

__call__(texts: pa.ChunkedArray) → pa.Array#

Call self as a function.

options(**kwargs: Unpack[ExprUDFOptions]) → Self#
with_column(texts: pa.ChunkedArray) → pa.Array#