需求背景：目前调用方较多，费用异常时很难统计
需求实现：
1、在GPT接口中增加一个参数，用于区分业务类型，各业务调用时根据我们分配的ID进行传参（服务内部调用的业务也按照刊登，重生标题进行拆分）
2、在调用时每次请求增加调用统计，按照业务id进行归类，统计模型版本，调用方式（实时或异步），调用次数，请求token数，返回token数，总token数（一个小时汇总一次到天表）
3、新增一个mysql的表，按天统计各业务id的调用量，并根据模型版本的单价预估费用
价格附录（左边为请求单价，右边是返回单价，都是1000token的单价，注意计算费用的时候转换）
"ChatGpt4o":[0.005*7.11,0.015*7.11],
"doubao-seed-1-6-250615":[0.0008,0.002],
"chatgpt4o":[0.005*7.11,0.015*7.11],
"gpt-4.1-2025-04-14": [0.002*7.11, 0.0088*7.11],
"gpt-4.1-mini-2025-04-14": [0.0004*7.11, 0.0016*7.11],
"4omini":[0.00015*7.11,0.0006*7.11],
"chatgpt4o-mini":[0.00015*7.11,0.0006*7.11],
"gpt-4o-mini-2024-07-18":[0.00015*7.11,0.0006*7.11],
"ChatGpt35":[0.002*7.11,0.002*7.11],
"chatgpt3.5":[0.002*7.11,0.002*7.11],
"deepseek-v3":[0.002,0.008],
"deepseek-r1":[0.002,0.016],
"qwen-max":[0.0024,0.0096],
"gpt-5":[0.00125*7.11,0.01*7.11],
"gpt-5-2025-08-07":[0.00125*7.11,0.01*7.11],
"qwen-vl-max-latest":[0.003,0.009],
"qwen-plus":[0.0008,0.002],
"qwen-max-2025-01-25":[0.0024,0.0096],