DB-GBP功能使用探索

1.DB-GPT支持的功能
2.配置StarRocks数据库
[3. ChatBI](#3. ChatBI)
[4. APP创建使用](#4. APP创建使用)
[5. 元数据库配置](#5. 元数据库配置)
[6. API 调用探索](#6. API 调用探索)
- [6.1 app](#6.1 app)
- [6.2 datasources](#6.2 datasources)
- [6.3 Knowledge](#6.3 Knowledge)
[7. 问题总结](#7. 问题总结)
[8. 贡献的PR](#8. 贡献的PR)

1.DB-GPT支持的功能

2.配置StarRocks数据库

支持多种数据库链接
但是只支持一个库访问查询

3. ChatBI

数据准确性不够精准，如果用于生产需要做相关优化
查询可能会做大SQL查询，拖死web进程,无法响应

4. APP创建使用

dbgpt app install awel-flow-web-info-search
dbgpt app install awel-flow-example-chat
dbgpt app install awel-flow-simple-streaming-chat

安装可能因为网络问题报错，多执行几次就好了

5. 元数据库配置

 LOCAL_DB_TYPE=mysql
 LOCAL_DB_USER=
 LOCAL_DB_PASSWORD=
 LOCAL_DB_HOST=127.0.0.1
 LOCAL_DB_PORT=3306
 LOCAL_DB_NAME=dbgpt

6. API 调用探索

6.1 app

app chat api 模型替换

sk-xxx
https://dashscope.aliyuncs.com/compatible-mode/v1
llm model: qwen-long

6.2 datasources

1.list datasources api error

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "D:\software\PyCharm 2023.3.5\plugins\python\helpers-pro\pydevd_asyncio\pydevd_nest_asyncio.py", line 120, in run
    return loop.run_until_complete(task)
  File "D:\software\PyCharm 2023.3.5\plugins\python\helpers-pro\pydevd_asyncio\pydevd_nest_asyncio.py", line 211, in run_until_complete
    return f.result()
  File "C:\Users\jack\AppData\Local\Programs\Python\Python310\lib\asyncio\futures.py", line 201, in result
    raise self._exception.with_traceback(self._exception_tb)
  File "C:\Users\jack\AppData\Local\Programs\Python\Python310\lib\asyncio\tasks.py", line 232, in __step
    result = coro.send(None)
  File "C:\Users\jack\PycharmProjects\dbgpt-test\api\datasource.py", line 12, in list_ds
    res = await list_datasource(client=client)
  File "C:\Users\jack\PycharmProjects\dbgpt-test\.venv\lib\site-packages\dbgpt\client\datasource.py", line 120, in list_datasource
    raise ClientException(f"Failed to list datasource: {e}")
dbgpt.client.client.ClientException: (Failed to list datasource: (E0003)
Reason: {'success': False, 'err_code': 'E0003', 'err_msg': "1 validation errors:\n  {'type': 'model_attributes_type', 'loc': ('response', 'data'), 'msg': 'Input should be a valid dictionary or object to extract fields from', 'input': [DatasourceServeResponse(id=1, db_type='mysql', db_name='dbgpt_test', db_path='', db_host='', db_port=3306, db_user='', db_pwd='', comment='')]}\n", 'data': None}
)
Reason: None

6.3 Knowledge

通过调用create_space接口失败但是创建了多个空间

调用delete_space 删除不了空间

Traceback (most recent call last):
File "C:\Users\jack\PycharmProjects\dbgpt-test\api\knowledge.py", line 39, in <module>
asyncio.run(delete())
File "C:\Users\jack\AppData\Local\Programs\Python\Python310\lib\asyncio\runners.py", line 44, in run
return loop.run_until_complete(main)
File "C:\Users\jack\AppData\Local\Programs\Python\Python310\lib\asyncio\base_events.py", line 649, in run_until_complete
return future.result()
File "C:\Users\jack\PycharmProjects\dbgpt-test\api\knowledge.py", line 24, in delete
res = await delete_space(client=client, space_id='2')
File "C:\Users\jack\PycharmProjects\dbgpt-test.venv\lib\site-packages\dbgpt\client\knowledge.py", line 75, in delete_space
raise ClientException(f"Failed to delete space: {e}")
dbgpt.client.client.ClientException: (Failed to delete space: (E0003)
Reason: {'success': False, 'err_code': 'E0003', 'err_msg': "1 validation errors:\n {'type': 'none_required', 'loc': ('response', 'data'), 'msg': 'Input should be None', 'input': SpaceServeResponse(id=2, name='test_space', vector_type='Chroma', desc='for client space', context=None, owner='dbgpt', sys_code=None, domain_type=None)}\n", 'data': None}
)
Reason: None

space delete error invalid space name:test_space_abc

7. 问题总结

可能会生成大查询SQL，导致查询卡死，响应非常慢
gbt任务会影响界面请求，gbt任务出问题会导致请求响应超时
目前Chat data 、DB等响应还不够准确，可能跟模型有关？
知识库创建的elsticsearch 卡片删除不掉
无法强制杀死正在运行中的对话任务

8. 贡献的PR

[1] https://github.com/eosphoros-ai/DB-GPT/pull/1794

[2] https://github.com/eosphoros-ai/DB-GPT/pull/1784

[3] https://github.com/eosphoros-ai/DB-GPT/pull/1789