Files
2026-05-23 16:35:27 -07:00

56 lines
1.9 KiB
Python

"""Request builder for Cerebras Inference (OpenAI-compatible chat completions).
Docs: https://inference-docs.cerebras.ai/resources/openai — use ``max_completion_tokens``
in API examples; non-standard fields via ``extra_body`` with the OpenAI client.
"""
from __future__ import annotations
from typing import Any
from loguru import logger
from core.anthropic import ReasoningReplayMode, build_base_request_body
from core.anthropic.conversion import OpenAIConversionError
from providers.exceptions import InvalidRequestError
def _normalize_max_completion_tokens(body: dict[str, Any]) -> None:
if "max_completion_tokens" in body:
body.pop("max_tokens", None)
return
if "max_tokens" in body and body["max_tokens"] is not None:
body["max_completion_tokens"] = body.pop("max_tokens")
def build_request_body(request_data: Any, *, thinking_enabled: bool) -> dict:
"""Build OpenAI-format request body from an Anthropic request for Cerebras."""
logger.debug(
"CEREBRAS_REQUEST: conversion start model={} msgs={}",
getattr(request_data, "model", "?"),
len(getattr(request_data, "messages", [])),
)
try:
body = build_base_request_body(
request_data,
reasoning_replay=ReasoningReplayMode.REASONING_CONTENT
if thinking_enabled
else ReasoningReplayMode.DISABLED,
)
except OpenAIConversionError as exc:
raise InvalidRequestError(str(exc)) from exc
request_extra = getattr(request_data, "extra_body", None)
if isinstance(request_extra, dict) and request_extra:
body["extra_body"] = dict(request_extra)
_normalize_max_completion_tokens(body)
logger.debug(
"CEREBRAS_REQUEST: conversion done model={} msgs={} tools={}",
body.get("model"),
len(body.get("messages", [])),
len(body.get("tools", [])),
)
return body