gpt-apt接口调用并行处理：python 多线程处理ThreadPoolExecutor

说在前面

近日因使用open ai的api调用，因数据量大，故不能一一申请调用gpt api，现来学习一下pyhton并行处理的功能ThreadPoolExecutor。它可以开n个线程，给定任务量m，这n个线程会同时处理这m个任务，直至任务量处理完毕。

代码

python 复制代码

def request_openai(system_prompt):
    response = client.chat.completions.create(
        model=model,
        messages=[{"role": "user", "content": system_prompt}]
    )
    # 获取token信息
    prompt_tokens = response.usage.prompt_tokens
    completion_tokens = response.usage.completion_tokens
    total_tokens = response.usage.total_tokens
    return response.choices[0].message.content, total_tokens



# 使用ThreadPoolExecutor并发处理多个请求
def handle_multiple_requests(prompts):

    # Create a ThreadPoolExecutor to process multiple tasks concurrently
    with concurrent.futures.ThreadPoolExecutor(max_workers=3) as executor:

        # Submit each task for processing
        futures = [executor.submit(request_openai, prompt) for prompt in prompts]

        # Process the results as they complete
        for future in concurrent.futures.as_completed(futures):
            try:
                response, tokens = future.result()

            except Exception as e:
                print(f"Error processing: {e}")

讲解：开通线程数量 max_workers=3，任务函数 request_openai，任务数据 prompts。

解释：假设1000条prompt，该代码功能就是这3个线程并行调用这1000条prompts，这三个线程会轮流处理这些任务，当一个线程完成一个任务后，它会去处理下一个任务，直到所有任务都被处理完毕。

llm_output：llm的输出数据的格式是由request_openai函数中规定的，而future.result()可以获取每次调用LLM的输出结果。一个future代表着一次调用gpt api。

改进版本代码

python 复制代码

def request_openai(item):
    prompt = get_prompts(item)
    response = client.chat.completions.create(
        model=model,
        messages=[{"role": "user", "content": prompt}]
    )
    total_tokens = response.usage.total_tokens
    return response.choices[0].message.content, total_tokens


open_path = './data/chinese1/testData.json'
with open(open_path, 'r', encoding='utf-8') as file:
    data = json.load(file)

with concurrent.futures.ThreadPoolExecutor(max_workers=3) as executor:
    # 使用字典 futures 来存储 future 对象，键是 future 对象，值是对应的 item
    futures = {executor.submit(request_openai, item) : item for item in data[:2]}
        with tqdm(total=total_tasks, desc="Processing") as pbar:
            for idx, future in enumerate(concurrent.futures.as_completed(futures)):
                item = futures[future]
                try:
                    response_dict, total_tokens = future.result()
                    # 更新进度条
                    pbar.update(1)

                except Exception as e:
                    print(f"Error processing id {item['id']}: {e}")
                    # 更新进度条即便发生错误
                    pbar.update(1)

不同之处：在这里为了能让 for future里面也能使用data数据，把每个data对象保存到字典futures中。最后再加上tqdm显示进度条。