Ensure compatibilty with pydantic v1

从 src/datapilot/core/platforms/dbt/schemas/manifest.py 开始，检查 Altimate 类及其 Pydantic 序列化调用。检查支持的 Pydantic 版本，并验证 manifest 序列化在 Pydantic v1 以及现有版本下都能正常工作，且不会引发报告的 AttributeError。

The main issue with using model_dump in Pydantic v1 is that this method doesn't exist in Pydantic v1. It was introduced in Pydantic v2 as part of the new serialization API.

The key evidence in your code:

1 src/datapilot/core/platforms/dbt/schemas/manifest.py uses BaseModel from Pydantic
2 Your Altimate* classes all inherit from BaseModel
3 In Pydantic v1, you should use .dict() or .json() for serialization instead
4 No indication of any compatibility layer for v1/v2 transitions

If the code attempts to call .model_dump() while running under Pydantic v1, it will raise:

AttributeError: 'AltimateManifestNode' object has no attribute 'model_dump'

💡 Recommendations if you need v1 compatibility:

1 Replace model_dump() with .dict())
2 Add version detection:

import pydantic                                                                                                                                                                                                                                                                                                                                                                                                                                                                  
                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 
if pydantic.version.VERSION.startswith("1"):                                                                                                                                                                                                                                                                                                                                                                                                                                     
    return obj.dict()                                                                                                                                                                                                                                                                                                                                                                                                                                                            
else:                                                                                                                                                                                                                                                                                                                                                                                                                                                                            
    return obj.model_dump()