debug_params('params_kw', params_kw)
if params_kw.off_peak:
	off_peak = params_kw.off_peak
	if off_peak in [True, "Y" "y", 1, "1"]:
		off_peak = True
	else:
		off_peak = False
	params_kw.off_peak = off_peak
userid = await get_user()
userorgid = await get_userorgid()
if userid is None:
	return UiError(title='llm inference', message='Please login first')
f = await checkCustomerBalance(params_kw.llmid, userid, userorgid)
if not f:
	return UiError(title='llm inference', message='余额不足或模型未配置定价')
kdbids = params_kw.kdbids
if kdbids:
	data = {
		'prompt': params_kw.prompt
	}
	params = {
		'query': params_kw.prompt,
		'fiids': params_kw.kdbids,
		'limit': params_kw.recall_cnt
	}
	tmpl = params_kw.prompt_tmpl
	ret = await rfexe('fusedsearch', request, params)
	data.update(ret)
	params_kw.prompt = await tmpl_engine.renders(tmpl, data)
	debug_params('rag', {'query': params.get('query',''), 'prompt_len': len(str(params_kw.prompt))})

env = DictObject(**globals())
return await inference(request, env=env)