feat(tools): 增加 OpenAI API 多轮调用功能
- 在 call_openai_api 函数中添加 rounds 参数，支持多次调用 - 累加每次调用的耗时和 token 使用情况 - 将多次调用的结果存储在 LLMRequest 对象的 response 列表中 - 更新函数返回类型，返回包含多次调用信息的 LLMRequest 对象 - 优化错误处理，记录每轮调用的错误信息
This commit is contained in:
carry
parent
5fc90903fb
commit
90fde639ff
@ -7,56 +7,81 @@ from typing import Optional
|
|||||||
sys.path.append(str(Path(__file__).resolve().parent.parent))
|
sys.path.append(str(Path(__file__).resolve().parent.parent))
|
||||||
from schema import APIProvider, LLMRequest, LLMResponse, TokensUsage, LLMParameters
|
from schema import APIProvider, LLMRequest, LLMResponse, TokensUsage, LLMParameters
|
||||||
|
|
||||||
async def call_openai_api(llm_request: LLMRequest, llm_parameters: Optional[LLMParameters] = None) -> LLMResponse:
|
async def call_openai_api(llm_request: LLMRequest, rounds: int = 1, llm_parameters: Optional[LLMParameters] = None) -> LLMRequest:
|
||||||
start_time = datetime.now(timezone.utc)
|
start_time = datetime.now(timezone.utc)
|
||||||
client = openai.AsyncOpenAI(
|
client = openai.AsyncOpenAI(
|
||||||
api_key=llm_request.api_provider.api_key,
|
api_key=llm_request.api_provider.api_key,
|
||||||
base_url=llm_request.api_provider.base_url
|
base_url=llm_request.api_provider.base_url
|
||||||
)
|
)
|
||||||
|
|
||||||
try:
|
total_duration = 0.0
|
||||||
messages = [{"role": "user", "content": llm_request.prompt}]
|
total_tokens = TokensUsage()
|
||||||
response = await client.chat.completions.create(
|
|
||||||
model=llm_request.api_provider.model_id,
|
for i in range(rounds):
|
||||||
messages=messages,
|
try:
|
||||||
temperature=llm_parameters.temperature if llm_parameters else None,
|
round_start = datetime.now(timezone.utc)
|
||||||
max_tokens=llm_parameters.max_tokens if llm_parameters else None,
|
messages = [{"role": "user", "content": llm_request.prompt}]
|
||||||
top_p=llm_parameters.top_p if llm_parameters else None,
|
response = await client.chat.completions.create(
|
||||||
frequency_penalty=llm_parameters.frequency_penalty if llm_parameters else None,
|
model=llm_request.api_provider.model_id,
|
||||||
presence_penalty=llm_parameters.presence_penalty if llm_parameters else None,
|
messages=messages,
|
||||||
seed=llm_parameters.seed if llm_parameters else None
|
temperature=llm_parameters.temperature if llm_parameters else None,
|
||||||
)
|
max_tokens=llm_parameters.max_tokens if llm_parameters else None,
|
||||||
|
top_p=llm_parameters.top_p if llm_parameters else None,
|
||||||
end_time = datetime.now(timezone.utc)
|
frequency_penalty=llm_parameters.frequency_penalty if llm_parameters else None,
|
||||||
duration = (end_time - start_time).total_seconds()
|
presence_penalty=llm_parameters.presence_penalty if llm_parameters else None,
|
||||||
|
seed=llm_parameters.seed if llm_parameters else None
|
||||||
# 处理可能不存在的缓存token字段
|
)
|
||||||
usage = response.usage
|
|
||||||
cache_hit = getattr(usage, 'prompt_cache_hit_tokens', None)
|
round_end = datetime.now(timezone.utc)
|
||||||
cache_miss = getattr(usage, 'prompt_cache_miss_tokens', None)
|
duration = (round_end - round_start).total_seconds()
|
||||||
|
total_duration += duration
|
||||||
tokens_usage = TokensUsage(
|
|
||||||
prompt_tokens=usage.prompt_tokens,
|
# 处理可能不存在的缓存token字段
|
||||||
completion_tokens=usage.completion_tokens,
|
usage = response.usage
|
||||||
prompt_cache_hit_tokens=cache_hit,
|
cache_hit = getattr(usage, 'prompt_cache_hit_tokens', None)
|
||||||
prompt_cache_miss_tokens=cache_miss if cache_miss is not None else usage.prompt_tokens
|
cache_miss = getattr(usage, 'prompt_cache_miss_tokens', None)
|
||||||
)
|
|
||||||
|
tokens_usage = TokensUsage(
|
||||||
return LLMResponse(
|
prompt_tokens=usage.prompt_tokens,
|
||||||
response_id=response.id,
|
completion_tokens=usage.completion_tokens,
|
||||||
tokens_usage=tokens_usage,
|
prompt_cache_hit_tokens=cache_hit,
|
||||||
response_content={"content": response.choices[0].message.content},
|
prompt_cache_miss_tokens=cache_miss if cache_miss is not None else usage.prompt_tokens
|
||||||
total_duration=duration,
|
)
|
||||||
llm_parameters=llm_parameters
|
|
||||||
)
|
# 累加总token使用量
|
||||||
except Exception as e:
|
total_tokens.prompt_tokens += tokens_usage.prompt_tokens
|
||||||
end_time = datetime.now(timezone.utc)
|
total_tokens.completion_tokens += tokens_usage.completion_tokens
|
||||||
duration = (end_time - start_time).total_seconds()
|
if tokens_usage.prompt_cache_hit_tokens:
|
||||||
return LLMResponse(
|
total_tokens.prompt_cache_hit_tokens = (total_tokens.prompt_cache_hit_tokens or 0) + tokens_usage.prompt_cache_hit_tokens
|
||||||
response_id="error",
|
if tokens_usage.prompt_cache_miss_tokens:
|
||||||
response_content={"error": str(e)},
|
total_tokens.prompt_cache_miss_tokens = (total_tokens.prompt_cache_miss_tokens or 0) + tokens_usage.prompt_cache_miss_tokens
|
||||||
total_duration=duration
|
|
||||||
)
|
llm_request.response.append(LLMResponse(
|
||||||
|
response_id=response.id,
|
||||||
|
tokens_usage=tokens_usage,
|
||||||
|
response_content={"content": response.choices[0].message.content},
|
||||||
|
total_duration=duration,
|
||||||
|
llm_parameters=llm_parameters
|
||||||
|
))
|
||||||
|
except Exception as e:
|
||||||
|
round_end = datetime.now(timezone.utc)
|
||||||
|
duration = (round_end - round_start).total_seconds()
|
||||||
|
total_duration += duration
|
||||||
|
|
||||||
|
llm_request.response.append(LLMResponse(
|
||||||
|
response_id=f"error-round-{i+1}",
|
||||||
|
response_content={"error": str(e)},
|
||||||
|
total_duration=duration
|
||||||
|
))
|
||||||
|
if llm_request.error is None:
|
||||||
|
llm_request.error = []
|
||||||
|
llm_request.error.append(str(e))
|
||||||
|
|
||||||
|
# 更新总耗时和总token使用量
|
||||||
|
llm_request.total_duration = total_duration
|
||||||
|
llm_request.total_tokens_usage = total_tokens
|
||||||
|
|
||||||
|
return llm_request
|
||||||
|
|
||||||
if __name__ == "__main__":
|
if __name__ == "__main__":
|
||||||
from sqlmodel import Session, select
|
from sqlmodel import Session, select
|
||||||
@ -69,14 +94,17 @@ if __name__ == "__main__":
|
|||||||
prompt="你好，世界！",
|
prompt="你好，世界！",
|
||||||
api_provider=api_provider
|
api_provider=api_provider
|
||||||
)
|
)
|
||||||
# 不使用LLM参数调用
|
|
||||||
result = asyncio.run(call_openai_api(llm_request))
|
|
||||||
print(f"\n不使用LLM参数调用结果: {result}")
|
|
||||||
|
|
||||||
# 使用LLM参数调用
|
# # 单次调用示例
|
||||||
params = LLMParameters(
|
# result = asyncio.run(call_openai_api(llm_request))
|
||||||
temperature=0.7,
|
# print(f"\n单次调用结果 - 响应数量: {len(result.response)}")
|
||||||
max_tokens=100
|
# for i, resp in enumerate(result.response, 1):
|
||||||
)
|
# print(f"响应{i}: {resp.response_content}")
|
||||||
result = asyncio.run(call_openai_api(llm_request, params))
|
|
||||||
print(f"\nOpenAI API 响应: {result}")
|
# 多次调用示例
|
||||||
|
params = LLMParameters(temperature=0.7, max_tokens=100)
|
||||||
|
result = asyncio.run(call_openai_api(llm_request, 3,params))
|
||||||
|
print(f"\n3次调用结果 - 总耗时: {result.total_duration:.2f}s")
|
||||||
|
print(f"总token使用: prompt={result.total_tokens_usage.prompt_tokens}, completion={result.total_tokens_usage.completion_tokens}")
|
||||||
|
for i, resp in enumerate(result.response, 1):
|
||||||
|
print(f"响应{i}: {resp.response_content}")
|
||||||
Loading…
x
Reference in New Issue
Block a user